1 DEPARTMENT OF HEALTH AND HUMAN SERVICES FOOD AND DRUG ADMINISTRATION CENTER FOR DRUG EVALUATION AND RESEARCH PEDIATRIC SUBCOMMITTEE OF THE ONCOLOGIC DRUGS ADVISORY COMMITTEE Thursday, June 28, 2001 8:25 a.m. Advisors and Consultants Conference Room 5630 Fishers Lane Rockville, Maryland 2 PARTICIPANTS Victor M. Santana, M.D., Chairman Karen M. Templeton-Somers, Ph.D., Executive Secretary MEMBER: Donna Przepiorka, M.D., Ph.D. AD HOC MEMBERS: Susan L. Cohn, M.D. Alice Ettinger, MSN, RN, CPON, CPNP Jerry Z. Finklestein, M.D. Henry S. Friedman, M.D. C. Patrick Reynolds, M.D., Ph.D. PATIENT ADVOCATES: Nancy Keen Susan L. Weiner, Ph.D. CONSULTANTS: Larry Kun, M.D. David M. Parham, M.D. GUESTS AND GUEST SPEAKERS: Robert S. Benjamin, Peter Burger, M.D. Anthony Elias, M.D. Howard A. Fine, M.D. Amar Gajjar, M.D. Stuart A. Grossman, M.D. Frederic Kaye, M.D. Victor A. Levin, M.D. Michael P. Link, M.D. Paul A. Meyers, M.D. Roger Packer, M.D. Elizabeth J. Perlman, M.D. Scott L. Pomeroy, M.D. David Poplack, M.D. Malcolm Smith, M.D., Ph.D. Susan M. Staugaitis, M.D., Ph.D. FDA: Richard Pazdur, M.D. Steven Hirschfeld, M.D., Ph.D. Joseph Gootenberg, M.D. 3 C O N T E N T S Call to Order 4 Welcome, Richard Pazdur, M.D. 4 Introduction of Committee 5 Conflict of Interest, Karen Templeton-Somers 8 Charge to the Committee, Steven Hirschfeld, M.D., Ph.D. 11 Challenges and Considerations in Linking Adult and Pediatric Solid Tumors, Victor Santana, M.D. 18 Challenges and Considerations in Linking Adult and Pediatric CNS Malignancies, Henry S. Friedman, M.D. 28 Discussion 36 Perspective on Sarcomas, Michael Link, M.D. 77 Perspectives and Background, Robert S. Benjamin, M.D. 98 Discussion 114 Questions 135 Perspectives on Lung Tumors and Neuroblastoma, Frederic Kaye, M.D. 155 Neuroblastoma and Small Cell Carcinoma of the Lung: Differences and Similarities, C. Patrick Reynolds, M.D., Ph.D. 171 Discussion 188 Questions to the Committee 197 Perspectives on CNS Malignancies, Susan M. Staugaitis, M.D., Ph.D. 213 Perspectives on CNS Malignancies: Clinical Aspects, Howard A. Fine, M.D. 228 Discussion 246 Questions to the Committee 270 Summary Comments: Paul A. Meyers, M.D. 275 Victor A. Levin, M.D. 279 4 1 P R O C E E D I N G S 2 Call to Order 3 DR. SANTANA: Good morning. We are 4 meeting this morning as part of the Pediatric 5 Subcommittee of the Oncology Drugs Advisory 6 Committee. This meeting was called by the agency 7 to give them advice and guidance on issues related 8 to pediatric development and, in particular, 9 extrapolation of information from adult studies 10 that could be relevant to pediatric studies as it 11 applies to the agency's regulatory role and the 12 Pediatric Rule. 13 We are going to go ahead and get started. 14 The first item is to have Dr. Pazdur address the 15 committee. Richard? 16 Welcome 17 DR. PAZDUR: Thank you very much. This is 18 one of three meetings that we are having to look at 19 the 1998 Pediatric Rule which, as Victor alluded 20 to, allows for the extrapolation of adult data to 21 the pediatric population. The first meeting looked 22 at leukemia and lymphomas and, obviously, the 23 nature of this meeting is looking at other 24 malignancies, particularly sarcoma, lung and CNS 25 malignancies and other solid tumors. Our third 5 1 meeting, which I believe is going to be held in 2 September, or to be announced -- some of you may be 3 asked to come back so we will get back to you with 4 specific dates and your calendars -- will look at 5 clinical trial design issues in pediatrics to 6 address issues of extrapolation of data, etc. So, 7 on behalf of the FDA, our Division of Oncology Drug 8 Products and our colleagues at CBER who handle 9 biologics, we would like to welcome you to this 10 committee meeting and look forward to an ongoing 11 dialogue with you. Thanks. 12 DR. SANTANA: Thanks, Richard. I want to 13 go ahead and introduce the committee members. 14 There are some people that are new to the meeting 15 and, for the purposes of record-keeping, we need to 16 state our name and affiliation. So, Stuart, can 17 you get started from that side of the table please? 18 Introduction of the Committee 19 DR. GROSSMAN: Stuart Grossman, from Johns 20 Hopkins University. 21 DR. LINK: Michael Link, from Stanford. 22 DR. MEYERS: Paul Meyers from Memorial 23 Sloan-Kettering. 24 DR. PACKER: Roger Packer, Children's 25 National Medical Center, Washington, D.C. 6 1 DR. POMEROY: Scott Pomeroy, Harvard 2 Medical School. 3 DR. PAZDUR: Richard Pazdur, Oncology 4 Division, FDA. 5 DR. HIRSCHFELD: Steven Hirschfeld, 6 Oncology Division, CDER, FDA. 7 DR. GOOTENBERG: Joe Gootenberg, with 8 Oncology at Biologics, CBER. 9 DR. PARHAM: David Parham, Arkansas 10 Children's Hospital. 11 DR. KUN: Larry Kun, St. Jude Children's 12 Research Hospital. 13 DR. COHN: Susan Cohn, Children's Memorial 14 Hospital in Chicago. 15 DR. ETTINGER: Alice Ettinger, St. Peter's 16 University Hospital, New Brunswick, New Jersey. 17 DR. FRIEDMAN: Henry Friedman, Duke. 18 DR. TEMPLETON-SOMERS: Karen Somers, 19 Executive Secretary to the ODAC, FDA. 20 DR. SANTANA: Victor Santana, St. Jude 21 Children's Research Hospital. 22 DR. FINKLESTEIN: Jerry Finklestein, Long 23 Beach Memorial, UCLA. 24 DR. PRZEPIORKA: Donna Przepiorka, Baylor, 25 Houston. 7 1 DR. REYNOLDS: Patrick Reynolds, 2 Children's Hospital, Los Angeles. 3 DR. WEINER: I am Susan Weiner. I am the 4 patient advocate from The Children's Cause. 5 DR. LEVIN: Victor Levin, Department of 6 Neuro-Oncology, M.D. Anderson Cancer Center. 7 DR. ELIAS: Anthony Elias, University of 8 Colorado. 9 DR. BENJAMIN: Bob Benjamin, M.D. 10 Anderson. 11 DR. GAJJAR: Amar Gajjar, St. Jude 12 Children's Research Hospital. 13 DR. PERLMAN: Elizabeth Perlman, Johns 14 Hopkins University. 15 DR. POPLACK: David Poplack, Baylor 16 College of Medicine. 17 DR. SMITH: Malcolm Smith, National Cancer 18 Institute. 19 DR. STAUGAITIS: Susan Staugaitis, 20 Cleveland Clinic Foundation. 21 DR. FINE: Howard Fine, Neuro-Oncology 22 Branch, NIH. 23 DR. SANTANA: That is it. Thank you so 24 much. We have to read a conflict of interest 25 statement. So, Karen, can you please proceed with 8 1 that? 2 Conflict of Interest 3 DR. TEMPLETON-SOMERS: The following 4 announcement addresses the issue of conflict of 5 interest with regard to this meeting and is made a 6 part of the record to preclude even the appearance 7 of such at this meeting. 8 Since the issues to be discussed by the 9 subcommittee at this meeting will not have a unique 10 impact on any particular firm or product but, 11 rather, may have widespread implications with 12 respect to an entire class of products, in 13 accordance with 18 U.S.C. Section 208(b), waivers 14 have been granted to all members and consultants 15 who have reported interests in any pharmaceutical 16 companies. 17 A copy of these waiver statements may be 18 obtained by submitting a written request to the 19 FDA's Freedom of Information Office, Room 12A-30 of 20 the Parklawn Building. 21 With respect to FDA's invited guests, 22 there are reported affiliations which we believe 23 should be made public to allow the participants to 24 objectively evaluate their comments. 25 Victor Levin, M.D., would like to disclose 9 1 that his retirement fund holds stock in Amgen, 2 Bristol Myers Squibb, Merck, Alza, Pfizer and 3 Pharmacia Corporation. Dr. Levin is also the 4 Program Director of an NIH, NCI National 5 Cooperative Drug Discovery Group grant, 6 "Development of Drug Inhibitors of Src" and he is 7 the Program Director of an NIH, NCI grant "Gliomas: 8 Biologic, Molecular and Genetic Studies." He is 9 also on the scientific advisory boards of Direct 10 Therapeutics, Signase and Oncology Services 11 Corporation. None of the companies he consults 12 with have anticancer drugs in clinical trials 13 except Direct Therapeutics, Inc. Dr. Levin is also 14 the founder and current member of the Board of 15 Directors of Signase, Inc. Lastly, his son is 16 employed by Alza Pharmaceuticals. 17 Susan Staugaitis, M.D. would like to 18 disclose that she owns stock in American Home 19 Products, Bristol Myers Squibb and various mutual 20 funds that may have investments in pharmaceutical 21 firms. 22 Paul Meyers, M.D. is the principal 23 investigator on a Bristol Myers Squibb sponsored 24 Phase I study of Irinotecan in children with 25 recurrent solid tumor. Dr. Meyers is also a 10 1 co-investigator for an Ortho-Biotech sponsored 2 study of erythropoietin in children with solid 3 tumors. Lastly, he is the principal investigator 4 on a Genentech sponsored study of Trastuzumab for 5 recurrent osteosarcoma. 6 Amar Gajjar, M.D. has a grant from 7 Schering Plough. 8 Anthony Elias, M.D. would like to disclose 9 that he is a researcher on clinical trials 10 sponsored by Eli Lilly, Pharmacia and Ribozyme 11 Pharmaceuticals. 12 Robert Benjamin, M.D. has received 13 consulting fees from Bristol Myers Squibb, Nexstar 14 and Sequus. He has also received speaker fees from 15 Bristol Myers Squibb. 16 Lastly, David Poplack, M.D. would like to 17 disclose that he has previously received speaker 18 fees from Chiron and he is an unpaid scientific 19 advisor to ASTA Corporation. 20 In the event that the discussions involve 21 any other products or firms not already on the 22 agenda for which an FDA participant has a financial 23 interest, the participants are aware of the need to 24 exclude themselves from such involvement and their 25 exclusion will be noted for the record. 11 1 With respect to all other participants, we 2 ask in the interest of fairness that they address 3 any current or pervious involvement with any firm 4 whose products they may wish to comment upon. 5 Thank you. 6 DR. SANTANA: Thanks, Karen. Any other 7 committee members that want to make any comments 8 regarding their conflict of interest? 9 [No response] 10 Thank you. We have some time now 11 allocated for an open public hearing. Anybody in 12 the audience that wishes to address the committee, 13 this is the time to do so. If you want to address 14 the committee, please come to the podium and state 15 your name and your affiliation. Nobody from the 16 audience wants to talk to us. Okay, thank you. 17 We are going to go ahead and start the 18 meeting. The first item on the agenda is Steven 19 Hirschfeld who will present the charge to the 20 committee. Steven has been a major force at the 21 FDA in trying to understand the issues of the 22 Pediatric Rule as it relates to oncology. So, I 23 want to thank Steven for all his efforts on behalf 24 of the pediatric oncology community. Steven? 25 Charge to the Committee 12 1 DR. HIRSCHFELD: Thank you, and I want to 2 thank and commend Dr. Santana for being the 3 initial, first and unprecedented chair for this 4 committee and for guiding it through its first 5 year. 6 DR. SANTANA: And hopefully not the last! 7 DR. HIRSCHFELD: Right! 8 [Slide] 9 Pediatrics has been a driving force for 10 changes in healthcare and particularly in clinical 11 investigations. The major regulatory initiatives 12 of this century were in reaction to 13 pediatric-driven events. It was the morphine 14 poisonings in the turn of the 19th to the 20th 15 century. It was the alfa-nilomide-tainting scandal 16 which led to the Food, Drug and Cosmetic Act, and 17 then the amendments to the Food, Drug and Cosmetic 18 Act which resulted in establishing the three 19 principles that we use for regulatory science which 20 is labeling, safety and efficacy which occurred in 21 1962 as a reaction to the malformations that were 22 caused by thalidomide. 23 In addition, children have had a key role 24 in the development of clinical investigations, and 25 most particularly in oncology. The first 13 1 chemotherapy studies were done at first in 2 uncontrolled studies in children and then in 3 controlled studies. The formation of the National 4 Cancer Institute and its clinical branches 5 initially had studies which examined the roles of 6 chemotherapy and also of statistics and of 7 randomized controlled study design in children with 8 leukemia. The advent of adjuvant therapy was first 9 done in children. 10 Yet, despite all the contributions toward 11 the development of clinical research and regulatory 12 efforts, there has never been a robust therapeutic 13 development program in children. So, there are 14 some efforts that were initiated over the course of 15 the last century but most explicitly in the last 16 decade to try to remedy what many felt was an 17 unjust situation. 18 We recognize that there are therapies that 19 were administered to children without adequate 20 study, both in general and in specific instances 21 which relate to oncology. We recognize the 22 extraordinary efforts of the cooperative groups in 23 developing clinical protocols, and the 24 extraordinary track record of both enrollment and 25 of scientific progress. Nevertheless, many of the 14 1 treatments that are used have been difficult to 2 come by, and many of the supportive care measures 3 have never been studied in the types of 4 environments which we would consider to be ideal, 5 and we would strive for this ideal. We also note 6 that many therapies are not made available for 7 pediatric study until adult marketing studies or at 8 least the adult program is well under way. 9 [Slide] 10 So, we have here a paradigm where the 11 conventional and historical method is that 12 preclinical studies with a new drug or biological 13 lead to clinical trials in adults, and then 14 following the adult development sometimes 15 unintended, sometimes intended, sometimes as an 16 afterthought comes pediatric development. What we 17 would like to engender is a new paradigm where 18 preclinical or non-clinical studies could lead to 19 either simultaneous adult and pediatric 20 development, or for those particular instances 21 where there is an unmet medical need and there is a 22 scientific basis for proceeding where studies can 23 lead to therapeutic development in children and 24 then, if applicable, for adults. 25 These inter-relationships is what we are 15 1 trying to explore in this committee over the course 2 of the past year, looking at where we can form a 3 matrix rather than a linear development plan. 4 [Slide] 5 The FDA, in the 1990's, attempted to 6 facilitate the availability of drugs for study in 7 children, and by drugs I mean drugs and 8 biologicals. With the Rule in 1994 that attempted 9 to ease the burden of clinical studies by allowing 10 extrapolation of efficacy data from adult 11 populations to pediatric populations certain 12 conditions were met. 13 The conditions were, in brief, that the 14 indication, which means the disease or condition, 15 but that the indication is similar in adults and 16 children and that the mode of action of the 17 intended therapy is considered similar in adults 18 and children. Therefore, the burden for scientific 19 studies would rely on study designs which could 20 establish appropriate dosing and appropriate safety 21 information but would not necessarily have to 22 recapitulate efficacy data. 23 This program was not the success it was 24 intended to be. So, two other programs were 25 initiated to replace it. The first was an 16 1 incentive program, which was part of the 1997 Food 2 and Drug Administration Modernization Act, which 3 offered a financial incentive to companies that 4 were willing to pursue pediatric studies in 5 response to a written request from the FDA. We 6 recognize the FDA does not have the resources nor 7 necessarily the wisdom to know which types of 8 studies to request so a mechanism was developed to 9 allow companies or interested third parties to 10 propose to the FDA pediatric studies, which then 11 the FDA would evaluate and then amend or issue a 12 written request on the basis of that proposal. 13 This program has been highly successful. 14 More pediatric studies have been initiated in the 15 past five years than ever in the history of 16 clinical investigations. This program has also 17 resulted in the issuance of twenty written requests 18 for pediatric oncology. 19 [Slide] 20 The other regulatory initiative is a 21 mandate, and the mandate states that if the 22 indication for an application under review can be 23 found in children -- and the operative words here 24 are "indication" and "under review" -- then the FDA 25 can mandate -- and again the operative word is 17 1 "can" -- mandate pediatric studies. It applies to 2 drugs and biologicals. If the indication does not 3 apply to children or there are other compelling 4 reasons not to pursue studies in children, then a 5 waiver can be granted. 6 This rule does not specifically address 7 the issue of extrapolation of efficacy. What this 8 rule asks and what I ask this committee to bear in 9 mind today is are studies warranted. Is there a 10 scientific basis for considering pediatric studies? 11 I should also note that this rule is not 12 intended nor has it ever, and we hope ever a 13 situation would arise where a question comes, 14 should it delay development for an adult indication 15 because pediatric studies can always be deferred 16 and there is no intent to ever delay the 17 availability or marketing of a new therapy for 18 adults. 19 [Slide] 20 So, the specific question we would like to 21 ask the committee this morning and this afternoon 22 is how should this rule be applied for solid tumors 23 and central nervous system malignancies. 24 [Slide] 25 What we would hope is that by the end of 18 1 the day we could have some recommendations for 2 adult indications that should trigger the Pediatric 3 Rule; some specific recommendations for adult 4 indications that should be waived from compliance 5 with the Pediatric Rule; and when this rule was 6 written we anticipated the situation, and there are 7 circumstances such as breast cancer where the 8 disease does not occur in children or occur in 9 sufficient numbers that an examination is warranted 10 every time an application is under review, there is 11 an automatic waiver. So, our question is should 12 there be other such conditions? 13 We would like, lastly, recommendations for 14 general principles that may be used to apply the 15 Pediatric Rule. We recognize that classification 16 schema are always changing, are fluid, as they 17 should be, and rather than convene a committee on a 18 regular basis to generate lists to update, it would 19 be helpful and preferable if we could have some 20 principles articulated to help us apply and 21 interpret the rule. Thank you. 22 Challenges and Considerations 23 in Linking Adult and Pediatric Solid Tumors 24 DR. SANTANA: We will go ahead and do the 25 presentations and we will have plenty of time for 19 1 questions and discussion to kind of keep it moving. 2 I am going to go ahead and take the podium. 3 [Slide] 4 What I want to do in the next ten minutes 5 or so is not to review all the challenges and 6 indications that may relate to pediatric solid 7 tumors but actually when I was thinking about doing 8 this what I decided to do were two things. One is 9 to kind of give a general overview consensus of 10 what I have taken out of the past couple of 11 discussions of this committee and my understanding 12 of where pediatric research and FDA regulatory 13 issues converge. Then, lastly, I would like to 14 bring forth the two points that to me are critical 15 as we move forward in considering extrapolation of 16 data, the two questions that we should always ask 17 when we are faced with that challenge. So, 18 hopefully, in the next ten minutes I will be able 19 to cover all that. 20 [Slide] 21 Clearly, there are two major issues here. 22 One is the research implications and the other one 23 is the regulatory implications, and by regulatory 24 implications I am only focusing on the FDA 25 perspective as it relates to the Pediatric Rule. 20 1 [Slide] 2 I think these are really a continuum, and 3 I think in pediatrics, and particularly in 4 pediatric oncology, we have a major advantage in 5 that pediatric oncology practice really occurs 6 almost exclusively within the research setting and 7 research trials are really the standard of care for 8 children in the United States who have cancer. 9 This is in real contrast to what happens in adult 10 oncology in which this is not the case or what may 11 happen in other pediatric diseases that are not 12 oncology in which research trials are not the 13 primary driving force of how patients are taken 14 care of. 15 From the regulatory perspective, once 16 again just focusing on the comment of how it 17 relates to the FDA and the Pediatric Rule, I think 18 we have to remember that the FDA is always looking 19 and the sponsors are always presenting data to the 20 agency in support of indications. I mean, that is 21 the ultimate goal of why they come to the agency. 22 In support of indications, obviously, they are 23 interested in looking at issues of efficacy as an 24 important endpoint but, as Steven addressed a 25 little bit earlier, a major component relates to 21 1 issues of safety and most of the mishaps that have 2 occurred in pediatric regulatory issues have 3 actually been issues related to safety and I am 4 going to talk a little bit about that later in 5 regards to some of the oncology drugs and how we 6 may address those. 7 I think whatever sponsors and the FDA do 8 with indications ultimately influences medical 9 practice not only in adults but also to a certain 10 degree in pediatrics, although in pediatric 11 oncology the ongoing theme is always that it is 12 done in the setting of research. 13 [Slide] 14 Now, I think we have to recognize that 15 there are some major limitations in pediatrics. 16 One is that we have a limited patient population. 17 So, many of the questions that we would like to 18 address many times cannot be addressed because 19 there is a limiting factor in terms of the number 20 of patients. A corollary to that is that many of 21 the diseases and solid tumors, for example, that we 22 treat are very heterogeneous in nature and there 23 are not large populations of patients within one 24 tumor category in which we can ask many different 25 questions. So, this is very different if you look 22 1 at it from the adult perspective because from the 2 adult perspective, in terms of drug development, 3 there are many agents that can be tested in a Phase 4 I setting because there are many adults in terms of 5 the numbers that can help us address those 6 questions. 7 Secondly, there are even fewer new agents 8 that can be evaluated in Phase II trials in 9 children because of the historical notion that many 10 trials first had to be conducted in adults before 11 any studies could be conducted in children. As 12 Malcolm Smith has reminded us many times, for many 13 of the pediatric solid tumors we can realistically 14 only do a Phase III study every four or five years 15 because of the issues of number of patients and the 16 issues of which are the real important questions 17 that have to be answered. I think the example 18 there is what has happened with Ewing's sarcoma and 19 osteosarcoma in the last decade in which 20 realistically, at the national level, Phase III 21 studies in those tumor types could only be carried 22 on in the context of every four to five years. I 23 think that is important as, from the research 24 perspective, we try to address what are the real 25 questions that we should be asking. 23 1 So, from the research perspective there 2 need to be mechanisms by which we can prioritize 3 what we can do in pediatric oncology with our 4 trials, and I think these three points that Malcolm 5 Smith has expressed before are that these 6 prioritizations have to be based on some idea of a 7 successful approach in adults because of the issue 8 of the limitation of patient numbers; that there be 9 compelling preclinical rationales for why these 10 questions with these agents should be asked in 11 children; and then paying some close attention to 12 the patient population at hand because there may be 13 specific patient populations in pediatric oncology 14 in which this may be more reasonable. For example, 15 patients at high risk for recurrence provide a 16 unique mechanism for us to be able to ask some of 17 these research questions. 18 [Slide] 19 However, as Steven addressed this a little 20 bit earlier, one of the primary concerns always in 21 pediatric research is this issue that we have to 22 obtain useful data. It is going to be limited 23 data, and a central issue is always the issue of 24 safety in children. None of us wants to be 25 involved with issues in which an agent, even in a 24 1 research setting or a regulatory setting, has had 2 children involved and major mishaps occur. I think 3 it not only presents issues of our relationship 4 with the community but also from an ethical point. 5 We want to make sure that what we do with children 6 is always safe. 7 So, I think we have to recognize that 8 there always have to be studies done in children 9 with new agents to help us understand whether the 10 MTD, the pharmacokinetics and the pharmacodynamics 11 are truly different so that when these agents then 12 become publicly available we don't have issues with 13 safety. 14 The two that I have outlined here are good 15 examples. As you know, Taxol is not a drug that we 16 use a lot in solid tumors or in pediatric oncology, 17 but the schedules of administration of Taxol are 18 really very different in adults versus children, 19 and that relates primarily to the vehicles in which 20 this drug was originally formulated and the 21 toxicity that the vehicle may present when it is 22 given to children in very short infusions. 23 Similarly, teniposide, where the vehicle 24 preparation has a lot of alcohol in it, one has to 25 be very careful with high doses of teniposide in 25 1 children because potentially issues of alcohol 2 toxicity may be related to the safety in use of 3 this drug. 4 So, the point here is just to present to 5 you two very brief examples of how we cannot 6 technically extrapolate all the adult data in terms 7 of pharmacokinetics and dynamics to children 8 because there may be particular issues with 9 children that have to be addressed in the safety 10 issue. 11 Then, lastly -- I don't want to beleaguer 12 this point of safety but we have to recognize that 13 there are different populations and even babies are 14 different from ten-year olds and fifteen-year olds 15 as relates to the metabolism of drugs. 16 [Slide] 17 So, the question that we have for us today 18 that Steven presented, under the auspices of this 19 Pediatric Rule, how do we consider whether solid 20 tumors in adults and children are either similar or 21 different, and why is it important to us and why 22 are we here? 23 Well, I think the first point is that 24 there are truly limited opportunities to test new 25 agents in children so we have to be very careful in 26 1 what we bring forward. 2 We have to make this regulatory mandate 3 very practical. I think Steven was hinting at 4 that. We have to be careful that, from our 5 business partners in the pharmaceutical industry, 6 that we don't ask them to do things that are 7 unrealistic and impractical. We have to make this 8 mandate very practical for the benefit of us in the 9 research community, for the benefit of our 10 patients, and certainly for the benefit of the 11 industry. This has to be done in a very practical 12 way to make these agents then available for 13 children. 14 I think you are going to hear a little bit 15 of discussion today from various other presenters 16 about ways in which potentially we can address this 17 question of extrapolation of data by looking at 18 phenotype. I am a believer that an osteosarcoma in 19 a 10-year old is the same thing as an osteosarcoma 20 in a 25-year old. Maybe somebody believes 21 differently. We will hear that maybe today. 22 We could look at it from the genotypic 23 point of view, from the molecular point of view. 24 There may be common genotypes or molecular events 25 that make us believe that tumors are very similar 27 1 although histologically they may be very different. 2 [Slide] 3 So, my two rules then in trying to answer 4 this question are what two things am I going to be 5 looking for to help me decide whether things are 6 different or are similar enough that I could 7 consider them the same? I think in that regard the 8 two points that I hope we will hear some discussion 9 today of are, first of all, looking at the biology, 10 are there differences and similarities in the 11 biology? That is, what creates the disease 12 phenotype? If that is similar enough, are we 13 really talking about the same disease and the same 14 manifestations? 15 The second point is that as we try to 16 extrapolate data we need to look at the host, and 17 we need to look at differences and similarities in 18 the host because that may be critical in terms of 19 determining drug metabolism and toxicity and 20 relating to issues of safety, which is obviously a 21 primary concern. 22 [Slide] 23 Lastly, I want to present to you kind of a 24 general outline of how we may consider some of 25 these points in terms of extrapolating both the 28 1 biology and in terms of extrapolating host factors. 2 The progression and the malignant transformation 3 for the same tumor type may be very similar or may 4 be very different in children versus adults. There 5 may be common elements, such as drug resistance, 6 that tell us that the disease clinically behaves 7 the same way. Or, there may be differences in host 8 factors and enzyme polymorphisms that may lead us 9 to believe that, from the safety perspective, this 10 is an issue that we need to address in a different 11 population by looking at different pediatric 12 populations in a very unique way. 13 So, I wanted to finish here by just giving 14 you my perspective on this issue in a very general 15 sense. My intent was not to discuss every single 16 solid tumor and the challenges and implications of 17 that because I think that will be done later today 18 by other speakers. Thank you. Henry? 19 Challenges and Considerations 20 in Linking Adult and Pediatric CNS Malignancies 21 DR. FRIEDMAN: This is a special day for 22 me since I have never done power-point before and I 23 want someone to come up and show me something, and 24 to be sure this went well I sent the slides ahead 25 to Karen and to Steve, the FDA, living and dead, 29 1 Congress and the District of Columbia. So, there 2 are a lot of slides that are out there. 3 [Laughter] 4 DR. SANTANA: Remember, Henry, that 5 everything you say here will be in the public 6 record. Okay? 7 DR. FRIEDMAN: I always remember that! I 8 strive for that! 9 [Slide] 10 What I am going to try to do today is to 11 show some of the challenges and considerations 12 involved in linking adult and pediatric CNS tumors. 13 [Slide] 14 The question posed is what is the 15 relationship between adult and pediatric CNS 16 tumors? Are there compelling similarities or 17 differences in these tumors which can guide us in 18 the application of the Pediatric Rule of 1998? 19 [Slide] 20 This shows you the histologic 21 classification of tumors of the CNS taken from the 22 most recent WHO publication. You can see that 23 tumors are divided into neuroepithelial tissues, 24 astrocytic, oligodendroglial, mixed glioma and 25 embryonal, ependymal, choroid-plexus, neuronal and 30 1 mixed neuronal tumors and pineal parenchymal tumors 2 -- 3 [Slide] 4 -- continuing with meningeal tumors, 5 primary CNS lymphomas, germ cell, tumors of the 6 sellar region and metastatic tumors. So, the real 7 question is what is the difference in the adult and 8 pediatric population? 9 [Slide] 10 First off, malignant gliomas, meningiomas, 11 Schwann cell and pituitary tumors are the most 12 common tumors we see in the adult population as 13 opposed to benign gliomas, medulloblastomas/PNETs, 14 which is primitive neuroepidermal tumor, and 15 craniopharyngiomas which are the most common in the 16 pediatric population. 17 [Slide] 18 The vast majority of adult tumors are in 19 the cerebral hemispheres. In pediatrics more than 20 50 percent of tumors in children who are over a 21 year in age are infratentorial, but a majority of 22 tumors in children less than one year of age are 23 also supratentorial but they are different from the 24 adult tumors -- the chiasmatic-hypothalamic gliomas 25 and choroid plexus tumors. 31 1 [Slide] 2 So, are they differences between adult and 3 pediatric non-glial CNS tumors -- the 4 neuroepithelial, nerve sheath, meningeal, germ 5 cell, CNS lymphoma, sellar tumors? The bottom line 6 is that there is no compelling data which suggests 7 that there is a meaningful difference between these 8 tumors in adults and children. There may be 9 differences but at the biological level there is no 10 compelling data to say there is a difference. 11 [Slide] 12 Are there differences between adult and 13 pediatric gliomas -- ependymomas, pilocytic 14 astrocytoma, oligodendroglioma, subependymoma, 15 diffuse fibrillary astrocytoma? Again, no data 16 supports a meaningful, if any, difference between 17 these tumors in adults and children. I want to 18 acknowledge Peter Burger's help in looking at some 19 of these issues. He was very helpful in our 20 discussions. 21 [Slide] 22 So, we really resolve to are there 23 differences between adult and pediatric malignant 24 astrocytomas -- the anaplastic astrocytomas, the 25 glioblastoma multiforme? 32 1 [Slide] 2 This is taken from a number of different 3 sources, one of David Lewis' publications most 4 recently, showing you a number of the molecular 5 changes that occur in the development of a 6 pilocytic astrocytoma, the so-called secondary 7 glioblastoma multiforme and the primary 8 glioblastoma multiforme which has a hallmark of 9 EGFR gene amplification. But, again, how does this 10 help us with pediatric versus adult? You have 11 copies of all these slides. 12 [Slide] 13 So, a series of questions, the same 14 question slide after slide now: are there molecular 15 distinctions between adult and pediatric malignant 16 astrocytoma? Rickert et al., in American Journal 17 of Pathology, 2001, compared adult tumors. Plus 18 1P, plus 2Q, plus 21Q, minus 6Q, minus 11Q, and 19 minus 16Q were more frequent in pediatric malignant 20 glioma than in adult malignant glioma. 21 [Slide] 22 Sung, et al., in Brain Pathology, 2000, 23 pediatric malignant astrocytoma show a preferential 24 p53 pathway inactivation, 95 percent or more, 25 moderate RB pathway inactivation, 25 percent, and 33 1 no EGFR amplification. 2 [Slide] 3 Cheng, in Human Pathology, '99, pediatric 4 malignant gliomas have moderate rates of p53 5 mutation, a lack of EGFR amplification, a low rate 6 of PTEN mutation, and a moderate rate of 7 microsatelite instability as opposed to adult 8 tumors. 9 [Slide] 10 Pediatric malignant astrocytomas rarely 11 display EGFR amplification but frequently display 12 increased EGFR expression, from Bredel, et al., in 13 Clinical Cancer Research. 14 [Slide] 15 Pollock showed malignant astrocytomas in 16 children greater than four years of age display 17 TP53 mutations and p53 overexpression similar to 18 adult tumors. Both TP53 mutations and p53 19 overexpression were much lower in children less 20 than four years of age, showing a difference in the 21 true biology of older and younger children. 22 [Slide] 23 Again, malignant astrocytomas are more 24 similar than distinct in adults versus children 25 greater than four years of age. So, in the older 34 1 child, although there are obviously distinctions in 2 their molecular phenotype or molecular expression 3 of genes, the similarities are greater than the 4 distinctions. 5 [Slide] 6 I would like to modify this slide a bit. 7 The Pediatric Rule applies to all adult brain 8 tumors, including malignant astrocytoma, however, 9 as we have started to hear and will continue to 10 hear, the number of tumors in pediatrics -- the 11 resources are so limited that it is going to be key 12 that there not be just a reflex application of the 13 Pediatric Rule to any adult brain tumor, but that a 14 discussion with the representative groups that are 15 addressing this problem be held on a tumor by tumor 16 or trial by trial basis to make a decision whether 17 it is appropriate to actually extend the rule and 18 enforce it. 19 [Slide] 20 Advantages -- and I want to thank Steve 21 Hirschfeld for help with this -- to joint adult and 22 pediatric malignant gliomas, new and improved 23 therapies for the patients; a better understanding 24 of the biology of the diseases; development of 25 common, comprehensive prospective biological 35 1 studies; a better understanding of the effects of 2 therapy in poor and good prognosis groups; new 3 study paradigms; more efficient study accrual and 4 use of resources. 5 [Slide] 6 However, we may be making some assumptions 7 that are in error in children exposed to therapies 8 of no merit. There is always the concern of 9 adverse events in children having a greater pebble 10 in the pond effect than in the adult population -- 11 just intrinsically the way this country operates. 12 Requirement for cooperation and sharing of 13 resources may delay or confound study 14 implementation. I think the merger of POD and CCG 15 has formed one central organization. There is also 16 the Pediatric Brain Tumor Consortium. More groups 17 mean more committees; more committees means more 18 time, not necessarily time well spent. Potential 19 need for complex stratification and analysis. 20 But the bottom line is that we have an 21 opportunity when the situation is appropriate to 22 take advantage of the Pediatric Rule because I 23 don't believe, and we will see how the discussion 24 goes today, that we will see a situation where we 25 want to apply the rule and we don't have grounds to 36 1 apply the rule. Thank you. 2 Discussion 3 DR. SANTANA: Thanks, Henry. We now have 4 time for discussion of the three prior speakers if 5 anybody has any questions to Steven, to Henry or 6 myself or want to make any general comments about 7 where we are so far. Paul? 8 DR. MEYERS: Henry, I think you made a 9 very compelling case that the biology is strongly 10 in favor of linking the pediatric and adult brain 11 tumors, but you didn't address the issue of 12 toxicity and whether or not you think there are 13 specific toxicities for brain tumor treatment that 14 would impede that ability. 15 The other question I would like to ask you 16 is are there any clinical differences in the 17 behavior of these tumors? I recognize we should 18 all be looking at biology as the more fundamental 19 question but, for example, do these tumors progress 20 more rapidly in children and does that have an 21 implication for clinical trial design? 22 DR. FRIEDMAN: In terms of the second 23 question first, I don't know how to answer that 24 because therapies are so distinct that the clinical 25 course of the tumors is obviously going to be 37 1 influenced by the interventions you use, and the 2 approaches in the adult and the pediatric 3 population are frequently quite disparate. So, it 4 is hard to answer that question. I will turn it 5 over to others -- Roger perhaps -- in a second. 6 The first question, certainly, I think the 7 toxicities are going to be an issue. If there is 8 going to be an adult trial which is going to use 9 50,000 sonograde whole brain radiotherapy, perhaps 10 in pediatrics we might frown upon that kind of a 11 study. I am only kidding, folks; we are not going 12 to do that. But, certainly, there are going to be 13 situations where, because of the developing CNS, we 14 might be eager to avoid certain interventions. 15 If you are talking about things that have 16 unclear neurotoxicity, that will have to be 17 factored in. I mean, certainly if there are 18 interventions which you know are going to pose more 19 risk of damage and you know you have a more 20 vulnerable situation in the pediatric population, 21 you are going to have to think about it. That is 22 part of the rationale for a case by case type of 23 situation, or tumor by tumor. 24 DR. MEYERS: I guess what I am suggesting 25 is that Steve was looking to us to try to draw 38 1 general principles, and I am almost hearing from 2 you that you think that is unlikely to be a 3 possibility. You are really suggesting that we are 4 going to need to look at each of these agents 5 individually. 6 DR. FRIEDMAN: Correct, absolutely 7 correct. Roger? 8 DR. PACKER: I really want to comment 9 mainly on the second point. I think that one of 10 the mistakes potentially made is that there has 11 been a tremendous reservation to look at new agents 12 in pediatric brain tumors because of the potential 13 effects on the developing nervous system. There 14 are ways now to monitor those effects, to evaluate 15 them. There are certainly tumors for which we have 16 really very little to offer patients. We are 17 really hung up often by not being able to look at 18 those agents. If we monitor them appropriately -- 19 we have MRI; we have neuro-cognitive assessments; 20 we have ways to monitor toxicity -- it shouldn't be 21 the rate limiter to applying the rule, there may 22 just have to be better considerations for how you 23 evaluate toxicity. 24 The other component of that is that it is 25 a true marketing issue for many of the companies. 39 1 If they get into a toxicity that may delay the drug 2 getting to market, that is the major limitation. 3 And, as we are looking at the new drugs we are not 4 only looking at chemotherapies, we are looking at 5 biologics, we don't know how turning on and off 6 these genes is going to affect the development of 7 the nervous system. We are looking at new drug 8 delivery methods -- convection delivery for CNS 9 tumors, and we are worried about the volume of the 10 brain. There is always this tremendous difficulty 11 to get over the barrier as we work with new 12 companies, pharmaceutical firms, etc., of trying to 13 get them to apply these to pediatrics. 14 I don't have the answer, except I think 15 sometimes it is overblown where the damage is going 16 to be. If there is going to be damage it will 17 identify it if we choose the target population 18 appropriately in those children who have no other 19 options, which is where I think these things should 20 be started, then I think the issue of CNS damage, 21 though an important one, is often a secondary one. 22 DR. ELIAS: I just have a comment on 23 something Victor said, and that is that basically 24 we are talking really about Phase II/Phase III type 25 of indications. It is clear from your discussion 40 1 that Phase I cannot be bypassed. The pediatric 2 populations are sufficiently different in a variety 3 of way the PK, growth of the organism, and so forth 4 -- that you really cannot bypass the safety 5 considerations. But what we are really talking 6 about in terms of the Pediatric Rule, I believe, 7 would be the Phase II/III indications for market 8 basically. 9 But I also agree that the safety issues 10 represent a major stumbling block in terms of 11 developing drugs, new agents. None of the 12 pharmaceutical companies want toxicities associated 13 with their agent. 14 DR. HIRSCHFELD: I will make a comment, 15 and these are just general comments, and I will 16 also invite Dr. Pazdur to follow up if he wishes. 17 But I cannot think of a single example of the 18 85-plus drugs that we have approved where toxicity 19 has proved to be the stumbling block. It is always 20 the issue of potential benefit versus potential 21 risk. I think it is clear that we have put an 22 enormous number of highly toxic substances out on 23 the market -- not us per se, I mean the 24 pharmaceutical industry and the academic 25 investigators and everyone, but we have allowed 41 1 these products to be on the market despite, in some 2 cases, their substantial toxicities because there 3 is a perceived benefit that, at least based on the 4 available data, seems to outweigh the potential 5 risks. It is one of the reasons why there are 6 medical oncologists and pediatric oncologists, 7 because we require that there be physicians and 8 facilities which specialize in the treatment and 9 monitoring of the patients in order to administer 10 these therapies. 11 The other issue that I wanted to comment 12 on in terms of general points is that while we may 13 not have specific principles, I think that if we 14 would look for patterns, and I think by the end of 15 the day we may see some emerge, we should keep our 16 minds open as to what potentially may evolve. Dr. 17 Pazdur, did you want to comment? 18 DR. PAZDUR: Basically, if you take a look 19 at why NDAs do not get approved, it is not because 20 of toxicity but because of lack of efficacy, by and 21 large. The toxicity issues are usually answered 22 well in advance to the time they get into an NDA 23 situation as far as major toxicities. Unusual 24 toxicities, especially if they occur in a pediatric 25 population, could be handled in labeling 42 1 considerations or in further studies. 2 But this kind of fear that the FDA will 3 halt the development of a drug because we see an 4 unusual toxicity in a subpopulation I think may be 5 somewhat overblown. Yes, we are interested in the 6 toxicity. It may require further studies, but a 7 lot of that could be handled in labeling issues or 8 in really looking at the toxicities in 9 subpopulations. The major issue or approval or 10 non-approval of NDAs is not toxicity; it is the 11 lack of efficacy, and I think a sponsor should be 12 well aware of that. 13 DR. FINE: I think the only caveat I would 14 say in speaking about brain tumors in particular, 15 and later on in the afternoon I am going to address 16 some of the clinical differences between the 17 pediatric brain tumors and adult brain tumors, but 18 I think it is important to say that efficacy can be 19 defined, obviously, in very many different ways and 20 particularly for adult brain tumors, where we are 21 dealing mostly with malignant gliomas where the 22 prognosis is so poor and our therapeutic 23 interventions are so limited, we are more likely to 24 approve a drug with marginal benefit and with 25 issues of long-term toxicity hardly being an issue. 43 1 However, taking pediatric tumors as a 2 whole, and we will talk about the specifics as the 3 day goes on, generally, thank God, children tend to 4 do better as a whole than the adults, maybe not per 5 high grade tumor but as a whole. So, for a 6 marginal benefit, if there is some significant 7 long-term toxicity we may be more reticent to 8 approve that drug for a pediatric indication than 9 for an adult. I think that is the one caveat I 10 would say. 11 DR. FINKLESTEIN: I think our challenge is 12 to think out of the box, and thinking out of the 13 box and going back to the history probably of the 14 generation of this committee, the idea was how can 15 we bring new ideas, new agents, new drugs to the 16 pediatric population earlier so the lag time would 17 be shortened? Dr. Hirschfeld referred to that in 18 terms of the algorithms that he was showing. 19 So, I would prefer that we not discuss or 20 not use the phrase we are only considering Phase 21 II/Phase III studies. What we are considering and 22 what our challenge is, as I understand it, is 23 bringing the pediatric oncologic challenge to the 24 forefront and thinking of a different way of 25 getting our children to have an opportunity to get 44 1 new agents earlier on, and the contributions of 2 Henry are excellent because by thinking together in 3 a unison manner in terms of brain tumors this will 4 help us. Now, I understand there have to be some 5 exceptions, but I would really hope we will think 6 out of the box and not think of the old algorithm 7 because that is what we really want to get away 8 from. 9 DR. PRZEPIORKA: A question for Steven. 10 Victor and Henry both highlighted the fact that 11 these tumors are not real prevalent in the 12 pediatric population. Can you bring us up to date 13 on what the FDA is doing to logistically identify 14 the priorities within the pediatric oncology 15 community for drugs in pediatric solid tumors and 16 CNS malignancies? 17 DR. SANTANA: Maybe Malcolm will want to 18 comment. 19 DR. HIRSCHFELD: I will refer to Malcolm 20 but I will start by saying we wish we were in the 21 position of having to prioritize these, but we are 22 not. So, we are looking prospectively and 23 hopefully at the circumstances. 24 I will just make one more point and then I 25 will ask Dr. Malcolm Smith, who has taken a 45 1 leadership role in this arena, to address your 2 question in more detail. But the other general 3 point is that the '98 rule mandates that the drug 4 be made available for studies, or the biological. 5 It doesn't say it should be approved for children. 6 It doesn't say that it should be in any other way 7 disseminated but should be in a controlled 8 circumstance, made available for studies, and that 9 was the principle I wanted to emphasize. Can I 10 just turn it over to Dr. Smith? 11 DR. SMITH: I would emphasize some of what 12 Victor said, that there is the need for 13 prioritization. In terms of the prioritization 14 process, I think it needs to lay with the experts 15 in the pediatric cancers. So, we are trying to 16 facilitate the prioritization process through the 17 Children's Oncology Group and its Phase I 18 Consortium; through the Pediatric Brain Tumor 19 Consortium; through the disease committees of the 20 Children's Oncology Group. We think that is where 21 the prioritization needs to occur. 22 The kind of tools for prioritization -- 23 and again Victor mentioned some of these, you know, 24 if an agent looks super in an adult carcinoma maybe 25 it is good in a pediatric embryonal tumor. It is a 46 1 good question. But we are trying to develop ways 2 for prioritizing better, having additional data to 3 base some of these decisions about whether the best 4 drug for rhabdomyosarcoma is going to be a 5 rhabdomycin analog or proteose inhibitor or an 6 epidermal growth factor, etc., inhibitor or, you 7 know, SDI571, all of which are either in the clinic 8 in pediatrics or soon will be. So, we get to the 9 point Victor was making, how many of those will we 10 be able to study in Phase II in rhabdomyosarcoma or 11 osteosarcoma? Then, which of those will we select 12 to be our Phase III drug for the next four or five 13 years, the question of therapy that we are asking? 14 We are tying to work with the pediatric 15 research community to develop additional ways of 16 using preclinical data to inform those decisions. 17 We sponsored a meeting together with the Children's 18 Oncology Group Phase I Consortium yesterday to 19 begin assessing what tools there are available now 20 for preclinical models, and then how those tools 21 might be used in a more systematic way. I think 22 that will be a key component to the prioritization 23 process, and making more information available to 24 the people making the decisions in the Phase I 25 Consortium, the Brain Tumor Consortium, the disease 47 1 committees within COG. 2 DR. SANTANA: I want to take the 3 chairman's prerogative and ask anybody in the 4 audience from the pharmaceutical industry who wants 5 to comment on these issues, because I think we are 6 having a discussion here from the academic centers 7 and from the regulatory agencies but the third 8 point here in the triad is the business and 9 pharmaceutical. So, I know there are a couple of 10 representatives here and so I would invite anyone 11 from the industry who is here who wants to comment 12 on this particular issue to come to the podium. 13 Please take the invitation. You don't get many 14 opportunities. I will give you a couple of minutes 15 to get your thoughts together. 16 DR. HIRSCHFELD: I just want to make one 17 other clarifying comment on the general principles, 18 and this applies to both the Pediatric Exclusivity 19 Initiative and the Pediatric Rule. What we are 20 attempting to facilitate is the generation of 21 information, data, as it relates to pediatrics. 22 So, in the Pediatric Exclusivity program we are 23 willing to give a financial incentive for even 24 negative data because we consider it important that 25 there be credible data available for study in 48 1 children. The same with the Pediatric Rule, even 2 if the drug does not lead to approval or leads to 3 an indication, it will still provide useful data. 4 The mechanism that we have for 5 disseminating the useful data is in the product 6 label, and we would consider it an effort well 7 worth the undertaking if we were able to write 8 information which was of use to practitioners in a 9 product label, again, even if it didn't lead to an 10 indication. 11 DR. SANTANA: Roger? 12 DR. PACKER: A comment and then a question 13 to the committee. The comment is I am not 14 absolutely sure that prioritization is not an 15 issue. We have already run into the road blocks in 16 some of the new angiogenesis and biology drugs of 17 how we are going to prioritize those drugs and how 18 we are going to apply them to pediatrics. We have 19 also hit road blocks at the regulatory level, at 20 the government regulatory level of allowing those 21 drugs to go into pediatric trials for pediatric 22 brain tumors until there is some adult data showing 23 their efficacy, which is a real problem in some of 24 the things. I don't want to go into specifics but 25 just to say that at the regulatory level it isn't 49 1 all that black and white, that there are road 2 blocks at this point. 3 The question to the committee though is 4 that I understand, I think, fairly well how this 5 rule is applied in one direction and it hasn't been 6 that difficult for many of the investigators here 7 to take a drug in adult malignant gliomas and apply 8 it to pediatric malignant gliomas. I think the 9 drug companies understand that the regulatory 10 agencies understand it. Where I have difficulty is 11 how is this drug or biologic going to be applied 12 for tumors where there is not a tremendous interest 13 in adult trials? How are we going to apply it 14 where there aren't drug trials for low grade 15 gliomas, which is a major pediatric problem? 16 Whether or not drug trials for primitive 17 neuroectodermal tumors in adults, which is a major 18 pediatric problem -- what data will be utilized by 19 the FDA to make this rule apply to those tumors 20 that are not in trials in adults? 21 DR. LEVIN: I would like to expand on that 22 just a bit and clarify one aspect of it, and that 23 is that the same problems exist in the adult groups 24 for treating anaplastic astrocytomas because 25 getting access to new drugs is basically focused on 50 1 the fast market approach of looking at glioblastoma 2 and for many of these new drugs that is not the 3 target. The target is a much lower grade tumor. 4 So, we have the same problems that you do in 5 addressing anaplastic tumors and lower grade 6 astrocytic tumors. 7 I would like to make one more comment and 8 maybe put it in a different light, and that is 9 basically for the less common tumors what you are 10 really all talking about is developing at a 11 preclinical level target identification which would 12 justify the use of a pharmaceutical agent that will 13 be coming out. And, I think the goal should be to 14 get access to a drug irrespective of whether there 15 is an adult counterpart, but basing the access of 16 the drug on the need to address inhibition of a 17 target. 18 I think that that approach needs to be 19 utilized, but I would agree it is hard to imagine 20 that the pharmaceutical industry would be willing 21 to give you a drug that is, say, used in small cell 22 or being developed for small cell carcinoma and you 23 are going to mount a trial now in medulloblastoma 24 where you are basically going to have to do Phase 25 I, Phase II and everything. That probably should 51 1 be one of the major goals of this committee, to try 2 to work out a way that makes it easier, maybe gives 3 the pharmaceutical company some either regulatory 4 or financial incentive to let that drug out for the 5 use in pediatrics. 6 DR. PAZDUR: That is the whole pediatric 7 plan that we developed under the FDAMA 8 interpretation, our interpretation of FDAMA, which 9 allows the development of drugs in the pediatric 10 population in a Phase I population, and even if 11 there is prohibitive toxicity, if there is a good 12 faith attempt that a Phase I study is done, then 13 they get the carrot of six months exclusivity 14 attached to their entire product line. Likewise, 15 if they do a Phase II study and it turns out 16 negative, it is a good faith attempt in providing 17 what we require as needed information so they do 18 get that carrot. So, that has been built into the 19 exclusivity plan for the development of pediatric 20 drugs. 21 DR. SANTANA: Steven? 22 DR. HIRSCHFELD: Yes, I wanted to just 23 address the matrix issue once more. Rather than 24 necessarily thinking of a triad of investigators, 25 regulators and industry, I want to emphasize a 52 1 matrix. And, there are many other components, most 2 important patients and their families because they 3 are the ones who are the focus of all our efforts, 4 and many other people who have an interest in it. 5 I think that we have made an attempt to engage in 6 dialogue with as many people as we think have an 7 interest or, as they are called fashionably these 8 days, stakeholders in the problem, and I think it 9 will require efforts which will involve all of us. 10 At the last meeting that we had our 11 pharmaceutical industry colleagues had the 12 opportunity to conference over lunch and make a 13 statement after lunch, and I wouldn't necessarily 14 want to put undue pressure if they want a little 15 more time to consider some comments. 16 DR. SANTANA: Anthony, yes? 17 DR. ELIAS: Yes, I just wanted to talk 18 about a different matrix of sorts because we are 19 talking about what do you do with rare diseases. 20 One of the other matrices, of course, is that now 21 many of the tumors in adult oncology are going to 22 be subdivided. They are going to be subdivided in 23 major ways based on gene array and we are really 24 going to be starting to talk about pathways, what 25 pathways are important. So, you are going to have 53 1 maybe EGFR being an important pathway across 2 multiple disease histologies and maybe you will 3 have a drug that is going to be approved for any 4 tumor that is EGFR, that has that as an important 5 pathway. 6 Now, we also do know that some of these 7 pathways may be different within the context of the 8 cellular milieu but, nonetheless, I think we may be 9 completely reorganizing our oncology taxonomy and 10 really be talking about pathways, which pathways 11 are important. I think that may completely shift 12 the types of indications people are going to be 13 looking for and make what was once a very rare 14 tumor into something extremely common. 15 DR. SANTANA: Yes, I want to follow up on 16 that. I think, you know, historically the agency 17 and the sponsors seek an indication for a very 18 specific item -- you know, second-line salvage 19 therapy for metastatic breast cancer; that is the 20 indication; that is where they come forth. I think 21 what you are suggesting, and I think we have 22 thought a lot about that, is that maybe it is time 23 for all of us to rethink that; that there may be 24 some drugs or some biologics in which the 25 indication which the sponsor seeks and that the 54 1 agency is after is very different. It is not the 2 historical, traditional breast cancer salvage 3 therapy for metastatic disease, but maybe some 4 biologic event which this particular target agent 5 targets. 6 DR. PAZDUR: We welcome that, and we could 7 handle that by labeling. For example, a drug could 8 be approved if it inhibits this enzyme in a variety 9 of tumors. So, that can be handled by labeling. 10 So, that is not an insurmountable problem for us to 11 overcome and basically apply to a pediatric 12 situation if there are tumors in the pediatric 13 population that overexpress that -- 14 DR. SANTANA: Yes, the challenge is to 15 identify those. 16 DR. PAZDUR: But this has to be well 17 defined by the scientific community, that this is a 18 way to reclassify tumors. Remember, whenever we 19 are mandating a company to do something it is a 20 little bit different than just saying, "won't you 21 do it? It would be nice." This carries a stick 22 with it and repercussions for the company both 23 financially and from a regulatory point of view. 24 So, we have to have a sound scientific basis. It 25 can't be on the basis of one report or a feeling 55 1 that these tumors may overexpress this issue. It 2 has to be a recognition that there is a change in 3 the taxonomy of how we deal with these tumors and 4 the terminology. 5 DR. SANTANA: Yes, Donna? 6 DR. PRZEPIORKA: To follow up on a comment 7 that you made regarding labeling, using as an 8 indication inhibition of a particular enzyme or 9 pathway, would that be outside the context of doing 10 a full study to determine whether or not that 11 pathways in, as Anthony put it, the cellular milieu 12 is actually going to be effective? Would you still 13 not require a specific disease indication? 14 DR. PAZDUR: No. 15 DR. HIRSCHFELD: We may not. 16 [Slide] 17 I put up a slide, which I had in reserve, 18 which shows the type of principle and it echoes the 19 same thinking that Dr. Elias articulated which we 20 have been discussing for several months, and which 21 we have discussed in previous meetings of this 22 committee. It states in sort of broad terms that 23 if a lesion -- and we haven't stated what a lesion 24 may be but it could be a pathway, a translocation, 25 overexpression of a particular gene, point mutation 56 1 -- is necessary for establishing or maintaining the 2 malignant phenotype, and if a therapy is directed 3 against that lesion, then studies in tumors where 4 the lesion occurs and has the same critical role 5 are warranted. So, there are a number of 6 conditions. It shouldn't just appear in cells but 7 it must play some central role in the pathogenesis 8 of the tumor type. That is the type of general 9 thinking that we would like to be moving toward and 10 away from the more conventional, historical, 11 traditional approach. 12 DR. PAZDUR: But this is going to require 13 a great deal of work obviously and, you know, I 14 don't expect a sponsor to come in and say, "okay, 15 this is a target and we're just going to develop 16 the drug only in this target" because they are 17 subject to basically the same confines as we are -- 18 is this a well accepted change in the way 19 physicians look at tumors? 20 How I would expect this to occur over 21 time? Probably these targets will be identified in 22 a particular tumor. When confidence develops that 23 this is the way that the drug works, then this will 24 be extended and we will kind of divest ourselves 25 perhaps of the histological confirmation of tumors. 57 1 But I think it is going to be a multi-step process. 2 It is not just going to be a bang -- this is the 3 target and we will just develop drugs. I think it 4 is going to be a step-wise evolution in how we look 5 at things rather than a complete change in one 6 study. 7 DR. HIRSCHFELD: And just one other point, 8 our overriding and regulatory-derived principles 9 must show patient benefit. So, the indication, I 10 would expect, would never be for inhibition of EGFR 11 in such-and-such a cell type. The indication would 12 read for patient benefit for prolonging life in 13 patients who have tumors that overexpress EGFR and 14 have certain other characteristics, and all we 15 would be doing is moving from a histologic 16 description of the tumor to a more functional or 17 biological description but it absolutely must show 18 patient benefit. 19 DR. SANTANA: I think our colleagues from 20 industry want to go ahead and make some comments. 21 For the purpose of the record, please state your 22 name and your affiliation. 23 DR. RACKOFF: I am Wayne Rackoff, a 24 pediatric oncologist at Johnson & Johnson. I just 25 wanted to make one comment and then Raj is going to 58 1 make a number of others, just to support what Steve 2 said about the comment that Roger made about 3 adverse events. This has come up, and I make this 4 comment really as one of the co-chairs of the COG 5 Industry Committee. It has come up in repeated 6 conversations; it has come up in conversations with 7 children's advocates and in our committee and here, 8 and in the committee at COG it has come up and, 9 Steve, we just want to support what you say, that 10 there are no data that support that this has ever 11 been an issue. 12 I think, just talking among ourselves 13 especially with the number of pediatric oncologists 14 who have entered clinical research and development 15 within industry, it is not something that we hear a 16 lot. There is always a concern, especially from 17 our commercial counterparts, about how we will deal 18 with toxicities in labeling and then in 19 commercialization. But in research and development 20 and in looking especially at the necessity of 21 providing a clinical development plan for 22 pediatrics when we come before the FDA, we know 23 that there are pediatric oncologists within FDA who 24 are sensitive to the issue that the labeling will 25 have to reflect that a specific toxicity occurs 59 1 just in a subpopulation. 2 So, we hope that what Steve has said, and 3 we will reiterate that over and over again at 4 meetings as it comes up, that that is not and 5 should not be a concern in inhibiting 6 investigators, consumer advocates and families from 7 coming to us and suggesting a study that would be 8 appropriate in pediatrics. 9 DR. MALIK: I am Raj Malik, with Bristol 10 Myers Squibb, also a pediatric oncologist. Just a 11 couple of comments, and I am speaking on behalf of 12 the COG Industry Advisory Council, and that has 13 been a great forum for really establishing, I 14 think, a new paradigm of collaboration between the 15 COG, the NCI, CTAP, FDA, certainly patient 16 advocates in terms of really addressing all the 17 issues that are being discussed here. 18 I think one of the issues that was 19 discussed at our last meeting was really the issue 20 of prioritization, and I think it keeps on coming 21 up over and over again because it speaks to, as Dr. 22 Pazdur said, to the sound scientific rationale. It 23 speaks to how are we going to take these 400 agents 24 in development and pick up the best agents to 25 develop in children. And, that is certainly a 60 1 process in which industry is also very interested 2 in participating and I was very glad to hear from 3 Dr. Smith that the first such meeting has already 4 started and we, in industry, look forward to 5 participating in that dialogue as well. 6 So, in general, you know, we are also very 7 supportive of the efforts that are going on here 8 and having a core of pediatric oncologists in 9 industry right now I think makes for a very 10 collaborative environment. 11 DR. SANTANA: Thank you for those very 12 supportive comments. Yes? 13 DR. MELEMED: My name is Allen Melemed, 14 with Eli Lilly. I just want to add one thing that 15 wasn't stated. I hate to say this but we have 16 somewhat of a bias because we are some of the 17 larger pharmaceutical companies that are usually at 18 these so there is somewhat of a resource issue from 19 larger pharmaceuticals to smaller pharmaceuticals 20 in the sense that we have more people, more 21 pediatric oncologists in the company and they may 22 not have the same resources to get the clinical 23 trials, and they may not have the same resources as 24 far as the actual drug supply. So, there is 25 somewhat of a bias, obviously, with the larger 61 1 pharmaceuticals. So, it might be harder on the 2 small biotechs where they have these new drugs that 3 you want. So, that is one thing I wanted to say. 4 The other thing is the timing of the 5 studies. The Pediatric Rule is a mandate. Now, 6 the FDAMA is a bonus and an addition that you can 7 get on exclusivity. That is a patent extension and 8 that extension occurs at the end of the patent. 9 So, you want and obviously we want pediatric 10 oncology drugs now, but for FDAMA you can actually 11 do studies at the end of the patent life or when 12 the drug is already marketed. So, a lot of this 13 doesn't address the incentive; it addresses the 14 rule and that is why you have to be careful how you 15 administer the rule. 16 DR. SANTANA: Anybody else have any 17 comments? Malcolm? 18 DR. SMITH: I would have a question to 19 Henry and others relating to the slide that Steve 20 has put up. One of the slides mentioned a report 21 of EGF receptor expression in the majority of 22 pediatric gliomas but not the amplification of the 23 gene. So, what data do we need then to say that 24 this is a valid target for pediatric high grade 25 gliomas or that it is just unrelated; it is there 62 1 but it is not really doing something, and how do we 2 develop those data to inform us? 3 DR. FRIEDMAN: Specifically are you asking 4 is the amplification going to be an issue or just 5 the increased expression? 6 DR. SMITH: Well, that is my question. 7 DR. FRIEDMAN: Okay, what is the relevant 8 parameter for a drug being effective, an EGFR 9 inhibitor, for example, in this setting? 10 DR. SMITH: Right, how do we know? We 11 know expression and what do we need to know to be 12 more confident or to be confident that, in fact, an 13 EGFR inhibitor would be a good drug to try in this 14 population? 15 DR. FRIEDMAN: I think in any given 16 situation the hope is going to be that there are 17 trials being conducted to help answer that. In 18 point of fact, for that particular question there 19 are several trials, including one at Duke that 20 specifically we will know in the space of 12-15 21 months what is the relevance of EGFR amplification 22 wild type versus mutant and increased expression 23 without amplification versus activity of an EGFR 24 inhibitor. And, there will be studies like that I 25 think from a number of different sources. I am not 63 1 sure if that is going to be happening, Howard, with 2 you or not at NCI, but I think that as we get a 3 better idea of what biological parameter, in this 4 case expression versus amplification, is critical 5 we will be able to have the answer to your 6 question. For that particular question probably 15 7 months from now we will have the answer. 8 DR. SANTANA: Susan? 9 DR. COHN: Yes, I just wanted also to 10 follow up. Malcolm, I think the meeting that you 11 had yesterday, looking at these preclinical models, 12 is certainly one thing that we will be very 13 interested in looking at and seeing if that will 14 correlate. So, I am sure it will be relatively 15 simple to set up some preclinical models looking at 16 EGFR expression versus amplification and then 17 looking at efficacy of various targets to see if 18 these models respond or don't respond. I would 19 imagine that would be certainly a place to start in 20 terms of prioritizing. 21 DR. LEVIN: If I may make a comment, I 22 don't think it is so simple because the issue with 23 some of these new molecules is to understand how to 24 use them. I, for one, would say that it doesn't 25 make much sense to give one of these inhibitors for 64 1 an amplified target like EGFR because you have the 2 issue of conservation of mass. You have to knock 3 down too many receptor tyrosine kinase sites than 4 you can possibly do. 5 I think that a lot of the preclinical 6 research done by industry and, hopefully, done by 7 pediatric consortia and private academic 8 institutions has to address the issue of, one, is 9 the target really good; two, what is the optimum 10 dose of these agents that needs to be given to 11 inhibit the target, not what is the optimum dose to 12 be given to produce the toxicity, the MTD that will 13 then allow you to go forward. We need to 14 understand exactly how these drugs work in order to 15 use them well, and I think it is going to continue 16 to be increasingly the goal of most successful 17 pharmaceutical efforts and academic efforts to 18 learn how to use these drugs so that they can be 19 used in combination. I think that is going to 20 require a commitment from industry, academia and 21 the NIH. I do not think that the commitment need 22 come from the FDA. 23 DR. FINE: To echo that and to follow up 24 on the meeting that we had yesterday on the 25 preclinical model, I would propose that that is 65 1 really the challenge to the pediatric academic 2 community. If they want to have the Pediatric Rule 3 more commonly come into play for access to better 4 drugs, the onus is on us to actually show that 5 these targets for these new drugs are validated 6 targets for pediatric brain tumors and that the 7 preclinical data supports their use, at which point 8 then the Pediatric Rule simply comes into play. I 9 am not sure it is necessarily the onus of the 10 pharmaceutical industry to do that. So, if we want 11 drugs for our children, I think it is within the 12 academic community to make that preclinical data 13 come to fruition. 14 DR. WEINER: From the parents' and 15 patients' perspective, I think what we really want 16 is reassurance that the science will prevail 17 regardless of either the economic incentives or 18 disincentives or regulatory environment. When we 19 bring our kids into the clinic, it is the trust 20 that the science will dictate those decisions 21 rather than any other consideration and I think it 22 is absolutely imperative that that is what prevails 23 in this environment. 24 DR. SANTANA: Very appropriate comment. 25 DR. POMEROY: I think another aspect of 66 1 this that may be driven as we understand tumors 2 better actually has applied to histologically based 3 taxonomy of tumors as well, which is that there are 4 some tumors, such as glioblastomas and high grade 5 gliomas, that are very prevalent in adults where 6 the development of treatments is very rapid and, 7 yet, they are very rare in children. So, we end 8 up, because of a numbers problem, not being able to 9 conduct trials at the same pace. 10 I guess one question that will be raised, 11 as we have these new inhibitor compounds and a new 12 understanding at a molecular level of what is going 13 on in these tumors, is are there ways that we could 14 apply either statistically or by joint trials an 15 efficacy trial which I think we all agree, at least 16 I certainly agree, is the big issue for many 17 pediatric brain tumors, more than toxicity. How 18 can we include children in trials that move along 19 quickly so when a new compound comes along we don't 20 have to wait five years to test it? Because I 21 think things are going to be moving along pretty 22 quickly over the next ten years. 23 DR. SANTANA: Anthony? 24 DR. ELIAS: Yes, I would agree with 25 Howard. I certainly don't think that the science 67 1 is yet there to be able to say that, for example, 2 any time you see EGFR that is going to be an 3 important pathway. I think our experience, for 4 example, with anti-ras therapy with FCI is just a 5 humbling case where it probably is the case that, 6 in fact, the targets that we are targeting are 7 actually not perhaps the targets that actually will 8 work. 9 So, I think to a certain extent the 10 principle of developing things where EGFR is, in 11 fact, an important target or one other pathway is 12 an important target across histologies is at least 13 plausible. I think we are not there yet to be able 14 to know what the gene patterns are, the milieu and 15 so forth to be able to predict yet without actually 16 testing it. In the future the hope will be that, 17 in fact, certain gene patterns are going to be able 18 to predict for response to certain types of 19 interventions and that you will be able to tell but 20 I don't think we are quite there yet. 21 DR. SANTANA: Robert? 22 DR. BENJAMIN: I would like to echo what 23 Scott said from a sarcoma point of view. If we try 24 to deal with specific pediatric studies in specific 25 sarcomas, whether defined based on a molecular 68 1 abnormality or defined based on histology, there 2 will never be enough children to study. Therefore, 3 if a separate study needs to be done the children 4 will never get the drug. I think the alternative 5 strategy, which is really not addressed by the 6 rules as I see them, is allowing for participation 7 of human beings in studies of their cancers 8 regardless of their age. I think that would allow 9 children to get their drugs more quickly when it is 10 appropriate. 11 DR. HIRSCHFELD: I think we recognize that 12 and on a to be announced date we will specifically 13 look at that issue of trial design and trial 14 access. 15 DR. SANTANA: Roger? 16 DR. PACKER: I would certainly echo your 17 comments as long as we set up those studies, and 18 this goes back to trial design, to know what we are 19 monitoring; that we can't always be monitoring the 20 same things, such as lowering of blood count or 21 elevation of liver functions. If you are going to 22 be monitoring aspects of brain development and 23 brain function differently in that population, I am 24 certain on board with that. 25 I would still like to come back to that 69 1 principle that is up there, and the term that 2 really keeps jumping out at me is "malignant 3 phenotype." We are still missing a large grouping 4 of patients and if we are going to be basing 5 things, as we say, on a biologic basis and this 6 receptor or this chemical being elevated in the 7 specimen we are again going to be treating patients 8 relatively late in the course of their illness. 9 One of the other things that I would like 10 this committee to battle with and the FDA to help 11 us to work with industry is how do we apply these 12 things, again, at a time where they might be more 13 effective -- going back to Dr. Levin's comments -- 14 not only in pediatrics but in adults at a time when 15 the tumor has not mutated to GBM, where we may have 16 not picked up the same markers and where we may not 17 have strong biological rationalizations, except the 18 clinical story will tell us that if we have a low 19 toxicity molecule maybe we should apply it early in 20 the course where we don't have compelling data yet 21 that things are amplified? That is where I don't 22 see these models helping us dramatically in getting 23 that early application. 24 DR. LEVIN: I think you have to be a 25 little careful though because we should be the same 70 1 as industry in some ways and we should be focusing 2 on the target. So, say, for the lower grade tumors 3 you find a set of target molecules, you should 4 really be seeking your drug based on that. Some of 5 the molecules that are out there, for instance EGF 6 receptor inhibitors, might well work much, much 7 better in that subpopulation. So, it is going to 8 be up to somebody in academia to come forward with 9 a hypothesis that says I can test this in animal 10 systems or I can test it in cells, and it appears 11 as if this is more likely to be effective in the 12 subpopulation, therefore, I want access to the drug 13 to test it against that population. The 14 pharmaceutical company might say, well, there are 15 only 50 patients a year with that disease; it 16 doesn't financially pay, and what you are really 17 asking then is, is there another mechanism by which 18 you can get access to that chemical. 19 DR. PACKER: Let me just comment on that 20 one other time. We have talked about a 21 transformation of tumors from low grade to high 22 grade and that has already been presented. There 23 is a point in all of these tumors, we think, 24 especially as they march along to glioblastoma 25 multiforme, where they picked up some of their 71 1 transformation but maybe it is not high enough that 2 we have been able to pick it up in a Petri dish. 3 Those molecules may be extremely effective when 4 there is a very low amplification, and if we are 5 going to be stuck and have to wait until we can 6 prove that we are going to miss the opportunity to 7 impact on the disease early in the course, and we 8 do a very bad job on impacting on disease later in 9 the course and although these molecules may be 10 wonderful, nothing yet has proved to me that when 11 disease is rampant it is going to turn the disease 12 off. And, I just want to know how to get at it not 13 only early in a patient population but early in the 14 course of the illness to the patient. 15 DR. HIRSCHFELD: I would like to ask Dr. 16 Poplack if he could just address this because I 17 know he has thought very much about this, and there 18 are in the hematological malignancies conditions 19 which are called preleukemic states and I would 20 like you to make a comment as to whether therapy or 21 intervening in these preleukemic states has thus 22 far had any impact, or just how you would approach 23 the problem. 24 DR. POPLACK: I think that there is 25 certainly a need to apply therapy in some of the 72 1 preleukemic states. I am not sure whether we have 2 analogies in brain tumors that would be appropriate 3 for therapy, and I think probably appropriately we 4 are focusing on the situations of greatest need. 5 Whatever principle we adhere to or gets applied 6 needs to be assessed and proven through these 7 trials, and I think it would be more difficult, 8 Roger, for us to be applying therapies to suspected 9 or hypothetical situations where we don't have 10 biological evidence even if there is a need. So, I 11 am not sure how you would suggest that we would 12 apply an agent, without having biological data, 13 just because there is a need. 14 DR. SANTANA: Yes, and the challenge to 15 identify those populations because you are now 16 going to be targeting populations that don't have 17 the complete spectrum of the disease. You are 18 targeting at a very much earlier point and the 19 challenge is to be very careful to identify those 20 populations. 21 DR. PRZEPIORKA: In the hematologic group 22 I think the one example that comes to my mind, 23 because of recent action, is Gleevec where the 24 tyrosine kinase inhibitor works wonderfully in the 25 chronic phase of CML which we don't consider 73 1 potentially a full malignancy, but doesn't work 2 anywhere near as well in blast crisis when there 3 are so many other things that actually contribute 4 to the malignant phenotype. The challenge, as 5 Victor put it, is trying to identify what is going 6 to be important early on, and studying the 7 malignant cells will give us a whole array of 8 possibilities but we have to figure out what is 9 that one thing that early on we can step in there 10 and really deal with. 11 I just wanted to make one additional 12 comment. I think in planning the drug design 13 meeting it is important to think about the public 14 health interest in making sure the drugs are 15 available also in adults with diseases that are 16 prevalent in small numbers, the same way that we do 17 with the pediatric groups. 18 DR. SANTANA: Dave? 19 DR. PARHAM: I think one thing we are 20 going to have to come to grips with in this 21 discussion is that in the groups of neoplasms we 22 are discussing there is no analogy to preleukemia. 23 All of these tumors develop in a full-blown 24 malignant fashion, particularly in sarcomas. Even 25 in the brain tumors fibrillary astrocytomas are 74 1 very, very uncommon and by the time they announce 2 themselves as tumors they are full-blown 3 malignancies or else they are pilocytic 4 astrocytomas which very rarely later on develop a 5 malignant phenotype. So, I am not sure that 6 discussion is going to be helpful here because 7 there are no identified pre-malignant stages in 8 these tumors. 9 DR. SANTANA: Good. I am going to go 10 ahead and ask that we take a break. We have had a 11 very good discussion. Let me summarize two points 12 in very general terms that I perceived from the 13 discussion this morning with a lot of detail. One, 14 I think through this whole discussion through all 15 these meetings, it is important, like somebody has 16 reminded us, that the endpoints don't change 17 whether we are talking about the Pediatric Rule or 18 any other mandate. We are still looking at 19 bringing forth treatments that are scientifically 20 based with a good rationale and that ultimately 21 demonstrate some efficacy and some benefit for the 22 patients. So, I think that is a central point in 23 this discussion. 24 The second thing that I think is very 25 important to recognize is that it is encouraging to 75 1 hear that both the agency and other federal 2 agencies that deal with pediatric oncology and 3 sponsors are willing to start thinking outside of 4 the famous box in developing probably other models 5 with some of these new biologics and some new 6 principles that potentially could apply. So, it is 7 very encouraging to hear that we are moving into a 8 different phase and that the agency is willing to 9 consider these proposals in a very different way. 10 I think we have talked about the general 11 things this morning. After the break we will 12 specifically start addressing some tumor types. 13 So, let's go ahead and take a 15-minute break and 14 reconvene at 10:15. Thank you. 15 [Brief recess] 16 DR. SANTANA: We are going to go from the 17 general now to the specifics. The first session in 18 which we are going to try to address issues is on 19 sarcomas. Before we get started, I am going to ask 20 Karen to just briefly give us some instructions 21 about lunch. Then after that, any members who 22 joined us after we started this morning do need to 23 introduce themselves for the public record. So, I 24 will ask those of you who came a little bit late 25 who did not introduce yourselves this morning to do 76 1 that. Karen? 2 DR. TEMPLETON-SOMERS: We have made 3 arrangements for those of you at the table to be 4 allowed into the Parklawn Building. So, you can 5 pretend you are a regular federal employee and eat 6 in our cafeteria, which is the most convenient 7 place. You are not obligated to go there but it is 8 quick -- 9 DR. SANTANA: It is an honor! 10 [Laughter] 11 DR. TEMPLETON-SOMERS: It is an honor, 12 yes! Victor has been there before and he is 13 willing to go back. 14 DR. SANTANA: Stick with the salads! 15 [Laughter] 16 So, when we are done with the morning 17 session we will just walk over there and Karen has 18 arranged for some stickers because we have to go 19 through security over there too. 20 Any committee members that joined us late, 21 could you please introduce yourself for the public 22 record by stating your name and affiliation? 23 DR. KAYE: Frederic Kaye, from Centers of 24 Cancer Research, NCI and the Naval Hospital. 25 DR. SANTANA: Thank you. 77 1 MS. KEENE: Nancy Keene. 2 DR. SANTANA: Patient. Thank you, Nancy. 3 We are going to get started. Our first 4 presentation is by Mike Link, from Stanford. Mike? 5 Perspectives on Sarcoma 6 DR. LINK: Well, first I would like to 7 thank the committee. I am flattered to be asked to 8 speak here and, as I understood my charge, which I 9 may not have understood, I was going to give some 10 perspective on sarcomas to set the tone for some 11 discussion. 12 [Slide] 13 As such, I will give a brief tour of the 14 sarcomas to provide some background at least from 15 the pediatric perspective. I talked with Bob 16 before and I hope that he will fill out those 17 aspects that we don't like to deal with. 18 [Slide] 19 So, I am going to give you some themes. 20 This is not the conclusion slide, this is the 21 themes, sort of the punch line that I might as well 22 get to right at the start. First of all, sarcomas 23 are a heterogeneous collection of diseases and 24 families of diseases so that we shouldn't be 25 thinking of them as a group. 78 1 The individual diseases and families may 2 be defined molecularly and a molecular derangement 3 characterizes each tumor type usually so that in 4 the ones where it has been explored there is often 5 a particular molecular derangement which defines 6 the malignancy, and this derangement in most of our 7 minds, even if not in minds of all pathologists, 8 supersedes system morphology in defining the 9 disease. So, we are now defining the disease on a 10 molecular basis. 11 It is unlikely, however, that the 12 characteristic molecular derangement is the entire 13 story. So, obviously, one molecular derangement 14 doesn't make a summer, to paraphrase that, and I 15 think obviously we are learning from further gene 16 array studies that there is a lot more that goes on 17 beyond the initial event. 18 But one thing that is important for this 19 particular discussion is that I think that these 20 are prototypic diseases which span the child and 21 young adult age range. So, this is a disease of 22 children and young adults and so obvious for this 23 particular kind of discussion. 24 [Slide] 25 From that, I am just going to proceed to 79 1 the usual background talk. This is a small piece 2 of the action in children as it is in adults. So, 3 it is only those red things, about 11 percent of 4 all the tumors we are talking about are the soft 5 tissue and bone sarcomas. 6 [Slide] 7 The way that I think most pediatricians 8 think of them, although I will be glad to be 9 corrected by others in the room, is that we divide 10 them into essentially three groups of tumors, three 11 major groups, the osteosarcoma; the Ewing's family 12 of tumors which is bone and soft tissue tumor and 13 includes peripheral primitive neuroepidermal tumors 14 and others, and I will go into that to show you 15 that this is a family of tumors that has now been 16 unified by a molecular concept; and then a group of 17 tumors that has been disunited perhaps by every 18 factor that we can think of, the soft tissue 19 sarcomas, the non-rhabdomyosarcoma soft tissue 20 sarcomas, about which I will have very little to 21 say, relying on Bob for that; and rhabdomyosarcoma 22 which we know is heterogeneous in itself because it 23 includes embryonal rhabdomyosarcoma and alveolar 24 rhabdomyosarcoma which, I will show you, are very 25 different diseases even though we treat them with 80 1 the same treatment strategies, and other variants 2 which are probably less important because they are 3 very rare. 4 [Slide] 5 I do want to leave you the impression that 6 we have made progress in these diseases and, in 7 fact, some of the progress that we have made is one 8 of the problems in terms of new drug development. 9 This is the history of, let's say, the overall 10 five-year survival in the three major groups of 11 sarcomas, rhabdomyosarcoma, osteosarcoma and 12 Ewing's sarcoma which appear in childhood. This 13 was in an article in The New England Journal of 14 Medicine showing progress over time. As you can 15 see, with the current state of the art there are, 16 fortunately, fewer patients left who are candidates 17 for experimental therapies at least as front-line 18 treatment. 19 [Slide] 20 I am going to start with osteosarcoma and 21 not say too much about it because Bob Benjamin is 22 also an expert here, but I just wanted to 23 demonstrate that age of onset of the disease 24 probably tells the story, more than anything 25 better, why this is a disease that adults and 81 1 pediatric patients should be considered together. 2 As has been stated before, I don't know that there 3 is much difference between a child in the second 4 decade or an adult in the third decade of life in 5 the behavior of the disease, assuming that we are 6 talking about classic osteosarcoma. 7 [Slide] 8 There are some molecular derangements in 9 osteosarcoma, although I think that most of us 10 would agree that not a single one of them unifies 11 the disease in the way that I will show you for the 12 other sarcomas, but there are mutations in RB gene 13 and p53 mutations which are certainly 14 characteristic of a minority of patients; MDM2 15 amplification and, through this, inactivation of 16 p53 which occurs in a minority of patients and 17 overexpression of Her2 which is an important 18 therapeutic target, but not in all patients. I 19 think, again, no single molecular derangement 20 defines this group of diseases. 21 [Slide] 22 I understood that I was supposed to give 23 you the state of the art or the state of the 24 therapies that we have and I am going to give you 25 two slides which show the unfortunate circumstance, 82 1 as we talked about earlier, where we are able to do 2 perhaps in the best of circumstances a trial every 3 four to five years. We haven't necessarily always 4 been able to accomplish that but even when we have, 5 this is the outcome of a trial that I ran between 6 1981 and 1986 with a long-term event-free survival 7 of somewhere in the neighborhood of 57 percent but 8 a 4-year event-free survival, as you can see, of 9 somewhere near 60-some percent. 10 [Slide] 11 Then a trial that Paul Meyers, who I am 12 sitting next to, just finished running, from 1993 13 to 1997 and the overall outcome is pretty much 14 superimposable on the curves that I just showed 15 you. So, a couple of decades of work and not much 16 progress in terms of the number of patients that 17 are cured. 18 [Slide] 19 A group of patients who we also have not 20 made much progress against is patients with 21 metastatic disease. Staging of bone tumors is 22 pretty easy. They either have metastases or they 23 don't that are clinically evident. This is a group 24 of patients where about 20 percent of them are 25 cured. They fare poorly even with modern 83 1 treatments and are, obviously, appropriate 2 candidates for new approaches as first-line 3 therapy. 4 [Slide] 5 Now I am going to turn to the second 6 category, Ewing's sarcoma, similarly a disease of 7 young adults and children but where the curve is 8 shifted dramatically more to the left. So, I think 9 that most of the adult oncologists would agree that 10 we probably know more about it or at least have 11 more experience with it than our adult oncology 12 colleagues. 13 [Slide] 14 Here we have the first of a group of 15 diseases where there is a molecular derangement 16 which characterizes the disease and underpins 17 tumorigenesis. Ewing's family of tumors is 18 characterized on the right, as you can see, with a 19 chromosomal translocation between chromosomes 11 20 and 22 usually, which produces a fusion gene and 21 gene product which characterizes about 95 percent 22 of cases of Ewing's sarcoma in the tumor cells, and 23 is felt to be a felt and malignant transformation. 24 On the left you see an analogous transformation 25 which I will return to in discussing alveolar 84 1 rhabdomyosarcoma. 2 [Slide] 3 So, this is a reciprocal translocation 4 found consistently in all of the family of Ewing's 5 sarcomas. So, soft tissue Ewing's, PNETs tumors, 6 all of the diseases that have had various different 7 names but now are unified together. Through EWS is 8 fused FLY1 or ERG, the two common partner genes, 9 and this translocation results in a 10 tumor-associated fusion gene which can be detected 11 by a variety of techniques in virtually all cases 12 and, therefore, has become sort of a diagnostic 13 test which we use to diagnose the malignancy often 14 more rapidly than we can get an answer from our 15 pathologists. 16 [Slide] 17 What is the state of the art? Again, 18 about two-thirds of the patients with no evidence 19 of metastatic disease are cured compared to 20 patients presenting with metastases that are overt 21 where somewhere in the neighborhood of 20-15 22 percent of the patients are cured. Again, the same 23 theme as I said for osteosarcoma, a group of 24 patients where we need better approaches. 25 [Slide] 85 1 But there are some confounding variables. 2 This is a site-specific tumor. Patients with 3 certain sites do better than others. I am not 4 going to show all of them here but there are 5 obviously confounding variables in this related to 6 tumor size and presence of metastases, etc. which 7 contribute to this, but they have to be considered 8 separately and is one of the caveats when we talk 9 about just lumping patients together. 10 [Slide] 11 Here is another theme that will recur, 12 although we think they are the same diseases, I 13 believe, in older patients and younger patients, 14 but there is a theme where, again, younger patients 15 do better. Children less than nine years of age 16 fare significantly better than older adolescents 17 and young adults. I will get back to this -- I 18 don't know if it qualifies as one of the pitfalls 19 but is certainly one of the caveats that we have to 20 think about in terms of lumping tumors in older 21 patients and younger patients together even if they 22 have the same molecular underpinning. 23 [Slide] 24 Now, the soft tissue sarcomas -- 25 rhabdomyosarcoma is the most common soft tissue 86 1 sarcoma in children. 2 [Slide] 3 More so than even Ewing's sarcoma, this is 4 a disease of young children, although I don't know 5 if it shows up on this slide. Part of the problem 6 with this slide, of course, is that many of the 7 studies of rhabdomyosarcoma entered patients for a 8 while only up until age 21. So, I am not sure that 9 we really know what the incidence is. There are 10 clearly a lot of young adults out there with 11 rhabdomyosarcoma but they have not appeared on 12 clinical trials so they are essentially lost to us 13 in terms of understanding them very well. But here 14 you can see that the majority of kids are 15 presenting younger than age nine, and certainly the 16 overwhelming majority younger than age 15. 17 [Slide] 18 Here it is very clear that this is at 19 least two diseases, even just by histomorphology 20 and we know that there is an alveolar and embryonal 21 subtype. Although until now most of the principles 22 of therapy have been shared between the two, it is 23 pretty clear that these two diseases are quite 24 different, and it is not necessarily clear why we 25 lump them except that because of the problems of 87 1 limited numbers of patients we often do so for 2 convenience and to get more robust clinical numbers 3 for our trials. 4 But it is important, as you can see if you 5 look at the BOTR, which is a botryoid which is 6 another version of embryonal, and lump that yellow 7 curve with the green curve which is embryonal and 8 then compare that to the lowest curve, the gold 9 curve, which is the alveolar histology, you can see 10 that this is really a very significant difference 11 in outcome depending on histology. So, it is an 12 important difference clinically. 13 [Slide] 14 Of course, as I have shown you, the 15 alveolar variant is associated with a chromosomal 16 translocation and the production of a fusion gene 17 unique to alveolar rhabdomyosarcoma. 18 [Slide] 19 If you look at the lower half of this 20 slide, this translocation, 2:13, is similar or 21 analogous to Ewing's sarcoma fusion gene, PAX3 to 22 one of the fork-head transcription factor members, 23 and there is an infrequent similar translocation 24 that involves PAX7 and FKHR, which I will talk 25 about in a minute. So, there are two very, very 88 1 similar translocations which characterize alveolar 2 rhabdomyosarcoma, and there are some cases that 3 don't have or at least have no detectable 4 translocation at all -- very different from 5 embryonal rhabdomyosarcoma where certainly no 6 clear-cut gene has been identified that 7 characterizes the disease. 8 [Slide] 9 Now, even the difference in the 10 translocation has an impact on the outcome of the 11 patients. So, the more common PAX3 involved, the 12 orange curve -- if we just look at patients with 13 metastatic disease, those patients fare terribly, 14 whereas those that have the alternative 15 translocation involving PAX7 actually do quite 16 well. So, again, we have to be very careful in 17 terms of defining the disease based on a fusion 18 gene because we think has variations in the fusion 19 gene do make a difference. I think, although it is 20 not entirely clear that everybody believes it but 21 in the Ewing's sarcoma there are variants of the 22 translocation and it seems that different break 23 points in translocation are associated with more 24 favorable or less favorable outcomes. 25 [Slide] 89 1 Once again, we have made progress overall 2 in rhabdomyosarcoma but when we look at how we are 3 doing lately it is pretty much the same, about 4 65-70 percent of children presenting with 5 non-metastatic rhabdomyosarcoma are cured, although 6 in the results of our last study, which was 7 published just recently in The Journal of Clinical 8 Oncology, there is no difference in outcome. When 9 we use three different regimens all of the drugs 10 have activity but there is no improvement in 11 outcome by regimen. 12 [Slide] 13 Now, rhabdomyosarcoma is a disease that is 14 unique in one way, and that is the disease behaves 15 very differently depending on the site of 16 involvement, and this makes one of the difficulties 17 in talking to adult counterparts where they have 18 site-specific diseases like breast cancer or bowel 19 cancer. This is a different disease at any of the 20 sites and it occurs in a multitude of sites. 21 [Slide] 22 If you look at the outcome by site, and I 23 am not going to belabor each of these things but 24 you can see that the outcome varies from 90 25 percent, the top curve, to more like 60 percent for 90 1 other presentations and this putatively is the same 2 disease. So, again, we have the problem that 3 although we think we know how to define this 4 disease, it is very different in its behavior 5 depending on a number of different factors. 6 [Slide] 7 Then, a recurrence of this theme in terms 8 of the impact of age, we know that older patients 9 do less well, as I will show you, and part of the 10 reason for that is because if you look at the 11 incidence of alveolar rhabdomyosarcoma, which I 12 have shown you is an adverse prognostic factor, the 13 incidence of alveolar is higher in older children, 14 33 percent for example in children older than 10 15 years of age compared to only 18 percent in 16 children in the 1-9 age group. So, a highly 17 significant difference. 18 [Slide] 19 Even stage of presentation -- older kids 20 much more frequently present with advanced stage 21 disease, again accounting for why older children 22 may do less well. 23 [Slide] 24 If we summarize what happens in older kids 25 with rhabdomyosarcoma, they have a lot of things 91 1 that make them less favorable which may or may not 2 have to do with the underlying biology of the 3 tumors that occur in older children. So, they more 4 frequently have alveolar tumors; tumors arising in 5 extremity, which is a bad site; larger tumors; more 6 invasive tumors; more regional spread and more 7 metastatic spread. So, not surprisingly, they do 8 less well. So, the question is, is this a feature 9 of a different disease in older children or are 10 there really fundamental biological differences, 11 analogous to some of the things we saw in brain 12 tumors that Henry showed? 13 [Slide] 14 This is just to demonstrate the relapse 15 hazard. So, the lower this curve, the better the 16 patients do. As you can see, it goes up both in 17 very young children and older children, showing 18 that those patients are much more at risk to 19 relapse. 20 [Slide] 21 Now I am just going to make a brief foray 22 into an area where I know very little, and most 23 pediatricians don't know very much and I hope Bob 24 will talk more about these, but when we talk about 25 the soft tissue sarcomas of children and you take 92 1 out rhabdomyosarcoma and its variants and soft 2 tissue versions of the Ewing's family of tumor, we 3 are left with just a long list. I think Bob's is 4 longer than mine, but these are the ones that occur 5 in children and they are very, very heterogeneous 6 in their histologic appearance, their behavior, 7 etc., but the common ones that we see are synovial 8 sarcoma. The ones I want you to focus on are -- it 9 is not even up there, but a couple of the others 10 that are important and I will show you the reason 11 in the next couple of slides. 12 [Slide] 13 The reason is that similar to Ewing's PNET 14 and alveolar rhabdomyosarcoma, some of these soft 15 tissue sarcomas are now also molecularly definable. 16 So, we can group them. For example, desmoplastic 17 small round cell tumor, characteristic 18 translocation, characteristic genes involved and, 19 actually, they are kind of familiar because the EWS 20 gene is involved in this tumor as well although 21 fused to another partner, Wilm's tumor gene, so 22 another pediatric partner is chosen. Similarly, 23 synovial sarcoma and congenital fibrosarcoma also 24 have very characteristic translocations -- again, 25 titillating in terms of the fact that we can define 93 1 the diseases and also have a potential target for 2 intervention. 3 [Slide] 4 My last slide on soft tissue sarcoma, just 5 to show that, number one, children without 6 metastases do very well; number two, that 7 interventions beyond surgery and radiation therapy 8 haven't made much of an impact that we know about. 9 I suspect there has been some impact overall in 10 adults but for a pediatrician it would be difficult 11 to be convincing, although it may be convincing to 12 an adult oncologist. The differences are quite 13 small. 14 [Slide] 15 So, having said all that, what are the 16 considerations when we try to link pediatric and 17 adult patients with sarcomas? We can say that the 18 diseases occur in children, adolescents and young 19 adults, excluding, let's say, the 20 non-rhabdomyosarcoma, the soft tissue sarcomas 21 which occur in older adults as well, but these are 22 basically diseases in a group of patients which 23 span the adult and pediatric ages. 24 I think we could say that the diseases in 25 adults and children may be similar on a molecular 94 1 level. I don't think there is any evidence that 2 adults, at least for the fundamental 3 translocations, have a different translocation but 4 there is obvious heterogeneity even within each of 5 these major subclasses of sarcomas, even 6 histologically, biologically. There are different 7 outcomes. And, it is pretty clear that there are 8 other significant molecular derangements and 9 differences in gene expression which will be likely 10 to be determined, if they haven't already been 11 determined, which distinguish patients even within 12 a category and probably older patients from younger 13 patients. 14 [Slide] 15 What are some of the other considerations? 16 Well, as you have heard in the talks in this 17 session, there are limited numbers of patients 18 available to begin with. There are hundreds of 19 patients with these tumors, not thousands of 20 patients each year in the United States newly 21 diagnosed. We cure a relatively high proportion of 22 them with current therapy so that there is 23 limitation on what subjects are available for 24 experimental therapies. Not to say that we 25 wouldn't be interested in incorporating an 95 1 experimental therapy, but it does make it difficult 2 to try to decide how you are going to cut back on 3 what we know is curative for two-thirds of the 4 patients. Therefore, it seems obvious that we 5 should be combining efforts among adult and 6 pediatric patients where the disease really appears 7 to be a continuum encompassing pediatric and adult 8 patients. 9 [Slide] 10 So, what are some of the other problems? 11 Older patients fare less well in all varieties of 12 sarcoma virtually. How do you explain that? Well, 13 are there really true age-related biological 14 differences? In other words, are older age 15 patients associated with other features of the 16 tumor itself that may not be defined by the primary 17 translocation but other molecules that have yet to 18 be defined that may be different in older patients 19 and younger? It wouldn't be surprising. 20 Age remains independently prognostic in 21 the studies that I have shown you. This may be 22 also a reflection of host tolerance to therapy. 23 So, it is a difference in host rather than 24 difference in tumor. It may be a difference in 25 compliance with intensive therapy. We know that 96 1 improvements in outcome have resulted from 2 therapies which are pretty hard to give and if you 3 had a choice, which a child may not often have, 4 they may not always come in on time. And, there 5 may be differences in physician compliance with 6 intensive therapy. 7 So, it is not even a patient or a tumor 8 issue; it is a doctor issue, and the mind set of a 9 medical versus a pediatric oncologist, perhaps best 10 demonstrated in a trial of treating adolescents 11 with leukemia and the difference in results in a 12 pediatric trial or a cooperative group trial that 13 was presented at ASH in December are very 14 compelling results, which showed very, very 15 different outcomes, probably a difference resulting 16 from doctor rather than fundamental biologic 17 differences in the tumors. 18 [Slide] 19 I just wanted to conclude. So, these 20 molecules that we have seen, and some of them kind 21 of not primary targets for the therapies that have 22 been developed, certainly present themselves as 23 things that we ought to be interested in. For 24 example, osteosarcoma -- Her2 is expressed and in 25 those tumors Herceptin would seem to be a logical 97 1 potential intervention, not something that was 2 developed with osteosarcoma in mind. The PDGF 3 signal transduction pathway is blockaded by 4 STI-571, again not a primary reason for the 5 development of the drug but a reason to test it in 6 osteosarcoma. Of course, for those tumors that 7 have p53 and RB abnormalities, those might be 8 suitable targets. 9 In rhabdomyosarcoma the fusion genes would 10 be an interesting target either from immunologic 11 approaches or from small molecule approaches. A 12 similar case could be made for the Ewing's family 13 of tumors and its specific characteristic 14 translocation, and also in Ewing's the stem cell 15 factor c-Kit signal transduction pathway could be 16 blockaded by STI, again another application of a 17 drug not developed specifically for that. 18 Desmoplastic small round cell tumor is not 19 exactly a public health menace but it is a pretty 20 nasty thing if you have it. Again, PDGF is 21 putatively expressed in these tumors and might be a 22 target for STI. I showed you some of the fusion 23 genes involved in some of the other soft tissue 24 sarcomas which we obviously be potential targets 25 for new therapies. 98 1 Hopefully, I have given some of the 2 reasons why we should be thinking in terms of 3 unifying these but understanding, of course, that 4 there are differences in adults and children and 5 their outcomes which may present not necessarily 6 obstacles but just food for thought before we can 7 willy-nilly make the recommendation that these 8 should be combined. 9 DR. SANTANA: Thanks, Mike. We will hold 10 questions until we have the second presentation. I 11 am going to invite Dr. Benjamin, from M.D. 12 Anderson. 13 Perspectives and Background 14 DR. BENJAMIN: I use a Mac, which is 15 intuitively obvious rather than this machine which 16 is not. 17 [Slide] 18 This is just a picture of M.D. Anderson. 19 [Slide] 20 I am going to talk to you a little bit 21 about the adult soft tissue sarcomas. Mike and I 22 did talk in the beginning and I thought that, 23 rather than overlapping, I would give you a very 24 different perspective, and my perspective is that 25 everything that you are talking about for 99 1 pediatrics applies in spades to sarcomas in adults. 2 So, the question is how do you define these tumors? 3 Should they be defined by patient age, histologic 4 type, molecular abnormalities or whatever? 5 [Slide] 6 Sarcomas are extraordinarily rare tumors, 7 less than one percent of all malignancies. Mike's 8 slide showed you that it is about 10 percent of 9 pediatric malignancies, so a higher proportion but 10 smaller numbers. And, it is the smaller numbers 11 that really kills us in terms of progressing in 12 terms of knowledge in the treatment of these 13 diseases. 14 I made the comment once that you wouldn't 15 treat adenocarcinomas all the same way, would you? 16 And, that came back to haunt me at a meeting that I 17 was at in Europe, but no medical oncologist would 18 think of treating adenocarcinoma of the breast the 19 same way as adenocarcinoma of the colon. They are 20 totally different diseases. Yet, if you asked 21 people about treating soft tissue sarcomas, they 22 are one disease. 23 [Slide] 24 Well, here is the one disease; there are 25 probably 50. In fact, there has never been a study 100 1 which has adequately addressed the diversity within 2 soft tissue sarcomas in adults, let alone put in 3 the pediatric counterpart. Now, what was just 4 presented to you very elegantly by Mike Link is 5 that the pediatricians have done studies in 6 osteosarcoma, single disease -- group of diseases 7 but single group. They have done studies in the 8 Ewing's family of tumors, relatively homogeneous 9 group. They have done studies in 10 rhabdomyosarcomas, some heterogeneity but 11 relatively homogeneous group. The rest of the 12 studies, the studies in adults are all done in 13 "soft tissue sarcomas" and there are 25 different 14 varieties or 50, depending on how you define them 15 on a histologic level, not even at a molecular 16 level. 17 [Slide] 18 You have already seen an updated version 19 on this. Many tumors do have specific 20 translocations. The ones in the pediatric age 21 group tend to have more, but I can point out for 22 you myxoid liposarcoma, which is a disease which is 23 almost exclusively an adult disease but which has a 24 specific translocation; synovial sarcomas which 25 occur certainly more frequently in adults; 101 1 extraskeletal myxoid chondrosarcoma which is a 2 disease of adults. 3 So, some of the sarcomas do have specific 4 translocations that are identified. The vast 5 majority do not. A lot of the translocations 6 involve the EWS gene but there is a huge difference 7 between those which are in the Ewing's PNET group, 8 which are very sensitive to chemotherapy, and some 9 of the others such as desmoplastic small round cell 10 tumor which is a disaster and even some of the 11 myxoid liposarcomas which have an EWS 12 translocation, or extraskeletal myxoid 13 chondrosarcoma which is very resistant. So, the 14 presence of EWS as part of the translocation 15 doesn't mean that you are going to have the 16 sensitivity that we see in Ewing's sarcoma. 17 [Slide] 18 I was once asked what is the best regimen 19 for adult soft tissue sarcomas? And, the answer is 20 it depends on which sarcoma and which patient, but 21 we haven't done the studies to prove that. So, let 22 me show you the reasons why I think this is 23 important and I think the reasons why we got 24 confused. 25 [Slide] 102 1 When sarcoma chemotherapy started in 2 adults about 30 years ago, it started with a drug 3 called Adriamycin before it was called doxorubicin. 4 The response rates, which may be different in the 5 way they were done to the way they are done now, 6 are more or less the same across histologies. The 7 only exception, and it is not a soft tissue 8 sarcoma, was chondrosarcoma where the response rate 9 was lower. But if you look, for example, at 10 leiomyosarcoma, one of the common groups, and 11 synovial sarcoma, another one of the common 12 histologies -- more or less the same response. So, 13 I think we got into the mind set that sarcomas are 14 all the same and they all respond the same way to 15 chemotherapy. 16 [Slide] 17 Well, there aren't very many good drugs 18 for the treatment of sarcomas. This is one where 19 DTIC was added to Adriamycin and again you saw the 20 same sort of breakdown more or less by histologic 21 group, and there wasn't a big difference. 22 [Slide] 23 There was a big difference by primary 24 site. This was pointed out in 1975 by Jeff Gotlieb 25 who said that tumors that arose in the GI tract, 103 1 even though they were primarily called 2 leiomyosarcomas, had a much lower rate of response 3 than tumors that arose in the GU tract, even though 4 most of those were called leiomyosarcomas. He 5 suggested that there was some biologic difference. 6 [Slide] 7 But if you look at synovial sarcoma, now 8 with ifosfamide, the other real drug for adult soft 9 tissue sarcomas, in a number of studies -- this is 10 our data, combined second-, third-, fourth-line 11 therapy for synovial sarcoma, a higher response, 31 12 percent versus an average of about 20; and, 13 leiomyosarcomas, whether of GI or other origin, 14 only about 10 percent. Well, you know, that may be 15 just per chance so let's look at some other 16 studies. 17 [Slide] 18 Karen Antman's study, synovial sarcoma, 40 19 percent; leiomyosarcoma, 7 percent. Here, 20 reasonable numbers of patients, 27 patients with 21 leiomyosarcomas. So, you have to think that maybe 22 ifosfamide is not a particularly good drug for 23 leiomyosarcoma. 24 [Slide] 25 Here is Le Cesne's high dose ifosfamide 104 1 study, 11 patients with leiomyosarcoma and no 2 response; 4 with synovial sarcoma, 3 responses. 3 Again, small numbers but everyone doing the same 4 thing. Synovial sarcoma is more responsive to 5 ifosfamide and leiomyosarcoma is less responsive to 6 ifosfamide. 7 So, if you then look at combination 8 studies, and I am not going to get into a whole lot 9 of them but if you are looking at 10 Adriamycin-ifosfamide and you simply report the 11 data out as sarcomas, it is uninterpretable data. 12 You need to know what you have of what in that mix. 13 [Slide] 14 So, what we have used primarily at 15 Anderson over the past several years as a 16 front-line therapy is Adriamycin and ifosfamide 17 with attempts to maximize dose because these tumors 18 have very steep dose responses. Mike Link's 19 comment about whether a pediatric oncologist or 20 medical oncologist treats you is right on. The 21 pediatric oncologists give more intensive 22 chemotherapy. The medical oncologists are babies; 23 they don't like to make people sick. They don't 24 like to get calls in the middle of the night. So, 25 they don't treat their solid tumor patients as if 105 1 they had acute leukemia, except for those who just 2 do sarcomas who look and see, "well, wait a minute, 3 if you want to get a result you have to give those 4 high doses so you have to make them sick." And, we 5 are a lot more like the pediatricians. 6 [Slide] 7 That is supposed to be a "less than or 8 equal to" sign, 65. We have no exclusions for 9 children on our studies. As a matter of fact, our 10 front-line studies for osteosarcoma and Ewing's 11 sarcoma are joint studies between pediatrics and 12 sarcoma medical oncology. There is no difference. 13 We treat them the same. They are the same 14 diseases. The adults do worse; the pediatric 15 patients do better. They tolerate therapy better. 16 But we can give it. You can't give 75/10 to a 17 65-plus year old adult because even though they 18 appear totally normal, they have abnormal kidney 19 function because kidneys age, and what we found out 20 the hard way is that if you really push these 21 people it is very easy to cause renal failure. So, 22 we look for renal function. We look at questions 23 of whether or not patients have two kidneys because 24 a lot of people with retroperitoneal sarcomas have 25 had a kidney removed. So, for those people, even 106 1 though their renal function is supposedly adequate, 2 it is not adequate when you give them high dose 3 ifosfamide. So, we have to be worried about that 4 sort of issue in dealing with sarcomas. 5 But when we started our treatments and all 6 of our protocols for dose-intensive 7 Adriamycin-ifosfamide we excluded people with 8 gastrointestinal leiomyosarcoma. That was the 9 diagnosis at that time. We excluded alveolar soft 10 part sarcoma because it doesn't respond to either 11 Adriamycin or ifosfamide in the small number of 12 patients that have been treated. And, we excluded 13 clear cell sarcoma because it doesn't respond very 14 well. 15 [Slide] 16 Overall, we have a response rate of about 17 60-something percent in this group of patients, but 18 there is a difference based on histology, and that 19 is the point that I was going to get at. With both 20 Adriamycin and ifosfamide you would expect that 21 synovial sarcoma, the most ifosfamide sensitive, 22 would do the best. 23 [Slide] 24 In fact, it does. We get an 88 percent 25 response rate. Angiosarcomas are very sensitive to 107 1 both drugs. Unfortunately, they also recur very 2 rapidly. I will get back to angiosarcomas a little 3 bit later. Malignant fibrocystoma, which is 4 probably five or six different diseases 5 characterized by pleomorphic histology and large 6 cells, responds well. There may be differences 7 within the subgroups but they all tend to respond. 8 But even the non-GI leiomyosarcomas that 9 we put on this study have only a 50 percent 10 response rate. So, there is a difference based on 11 what kind of sarcoma you have, and none of the 12 studies in adults have addressed this. Can you 13 imagine if we tried to now do disease-specific 14 studies in pediatric patients with these 15 histologies? That is impossible. We are just 16 getting to the point where maybe we can do an adult 17 study in a specific histology. If we add on the 18 pediatric patients, we can add them into the 19 disease-specific studies but you could never do a 20 study. It would take you 50 years to do the study, 21 by which time they wouldn't be pediatric patients 22 anymore. 23 [Laughter] 24 [Slide] 25 So, I think we need to get into 108 1 disease-specific therapy and the hints of this are 2 just now coming. 3 [Slide] 4 As we move into the new future where we 5 are getting into genetically specific molecular 6 therapy -- this is a slide that John Edmonston made 7 up about ten years ago and I liked it because that 8 was the time when we were the dinosaurs. I mean, 9 you know, we are going to think back on this era 50 10 years from now and say we were barbarians. 11 But the patients are here, and they are 12 dying, and we have to treat them now. But we now 13 have the first hint that the genetically specific 14 molecular therapy will, in fact, work and that is 15 in GI stromal tumors. These are the things we used 16 to call GI leiomyosarcomas. As I said, 25 years 17 ago Jeff Gotlieb said they are different; they 18 don't respond to therapy the same way. 19 [Slide] 20 About five years ago the pathologists 21 started recognizing that they were different and 22 gave them a different name and called them GI 23 stromal tumors and the key to these tumors is that 24 they come from the interstitial cell of Cahal which 25 constitutively expresses c-Kit, and about 90 109 1 percent of those tumors which you would call GIST 2 based on light microscopy are c-Kit positive. 3 It happens that c-Kit is inhibited by 4 Gleevec and the preliminary data that were 5 presented at plenary session at ASCO were very 6 exciting. So, an intergroup study started and 7 every sarcoma investigator in the world is 8 participating either in the U.S. or the European 9 version of this intergroup study. So, in the past 10 six months we have entered probably more than 70 11 patients with GI stromal tumors because we had them 12 waiting in the wings. We had been keeping them 13 alive by doing surgery or by doing chemo 14 embolizations of their liver, and they are coming 15 out of the woodwork. 16 Preliminary data from our group of 17 patients -- if you use traditional criteria for 18 response at 8 weeks, there is about a 30 percent 19 response rate or 40 percent response rate. If you 20 use PET scanning, it is a 70 percent response rate. 21 I think if we continue to follow we are going to 22 see the higher response because that is what was 23 shown in the earlier studies. But that also raises 24 the question that I will get back to, that we don't 25 know how to measure response. 110 1 [Slide] 2 Let me just go over to a couple of other 3 specific tumors, myxoid liposarcomas, again a 4 specific translocation different from other 5 liposarcomas. So, this is a specific disease 6 within the liposarcoma family. It is the only one 7 where differentiation therapy with either 8 PPAR-gamma or retinoid-X receptor agonists seems to 9 be effective. So, again, a specific target for a 10 specific therapy. 11 [Slide] 12 Angiosarcomas, a group of very difficult 13 tumors because they respond well but they relapse 14 rapidly and often they occur in elderly people on 15 the scalp. These people can't tolerate the same 16 kind of aggressive chemotherapy that we give to the 17 younger people. Taxol, in a series from Memorial 18 that is not even a formal study -- responses in 8/9 19 patients. Taxol doesn't work in sarcomas. The 20 Memorial study which had 2/28, one of which was an 21 angiosarcoma -- it is 1/27 in the other 22 histologies. We did Taxol in 19 patients, no 23 angiosarcomas, no responses. Taxol is not a 24 sarcoma drug but it works for angiosarcoma. It is 25 a different disease. 111 1 We haven't done a formal study but we have 2 treated patients with Taxol and the people at 3 Memorial are right, it really works. So, this is a 4 therapy that can be given. Weekly Taxol is easy 5 for a 70-year old. It can be given. It can even 6 be given by the local medical oncologist because 7 they know how to do Taxol. 8 Epithelioid hemangioendothelioma -- a 9 weird disease; doesn't occur in children, I don't 10 think. I haven't seen one. Primary tumors in 11 liver, they have been treated with liver 12 transplantation. They also can undergo spontaneous 13 remission. But if you have a lesion that is 14 growing, embolization, cutting off the vasculature, 15 is very effective. Interferon is very effective. 16 The new angiogenesis inhibitors haven't been 17 studied -- beautiful target. 18 [Slide] 19 I can't show you the slide I wanted to 20 show you unless I take my Mac up and hook it up to 21 this, but our definitions of response are all based 22 on tumor shrinkage and sarcomas clearly do not 23 always shrink when they die and we miss a huge 24 amount by not using more sophisticated methodology 25 to assess the effectiveness of our drugs. 112 1 One of the things that can happen, which 2 this slide would show if you have enough 3 imagination, is a tumor in the mediastinum that 4 grew in size a little bit but became totally 5 necrotic on CT. We looked at it and we said this 6 is a great response. I have shown that slide to 7 groups of people around the country at various 8 places and I said what would you call it? And, 90 9 percent of medical oncologists would call it stable 10 or progressive disease. There are only a few that 11 will call it a response. Well, when you have 12 progressive metastatic disease, usually it is time 13 to give up and send the patient to hospice. We had 14 our thoracic surgeons go in, take out the tumor 15 along with aorta because that was what was 16 required, and the patient is alive and well five 17 years later. Less than one percent viable tumor in 18 the specimen. 19 We learned in osteosarcoma that if the 20 tumor is dead it is a good prognosis. It means the 21 therapy worked. We have to figure out ways of 22 measuring, short of surgery, when tumors are dead. 23 And, at least a hint from the GIST experiment is 24 that PET scanning is maybe a way to do that but 25 other techniques -- dynamic MRI, dynamic CT and 113 1 probably other things that I haven't even dreamed 2 of are an approach but we have to get out of the 3 box and think about shrinkage because we are going 4 to miss drugs that are active. 5 [Slide] 6 So, I think we need to start looking at 7 other approaches and that is important. 8 [Slide] 9 So, getting back to pediatrics, where does 10 all of this fit in? In the adult sarcoma community 11 we are moving more and more towards accepting that 12 these many tumors are very different, that we 13 really do need to do studies where we address each 14 of the different groups and then follow-up on leads 15 on the groups that are positive. As we get 16 molecular markers of these groups, we will move 17 into molecular markers as ways of going. But they 18 are not all the same. We are going to have to get 19 separate trials, and I would include children with 20 the specific diseases on these trials. To try to 21 do a separate disease in children I think would be 22 fruitless. Thank you. 23 DR. SANTANA: Thank you, Bob. We have 24 other individuals in the committee with expertise 25 in this area so I would invite Paul, Anthony and 114 1 others with expertise in this particular area of 2 discussion to make comments now. 3 Discussion 4 DR. MEYERS: I would like to make several 5 comments. I think the point that Bob is making 6 about thinking of novel ways to evaluate tumor 7 response is extraordinarily important. Not only is 8 the technology that we use important, but the 9 timing. Dr. Elias was heavily involved in the 10 development of a drug, ET743, where we learned that 11 if we had used the conventional time point to 12 evaluate that drug we would probably have discarded 13 it early on, and the patients with soft tissue 14 sarcomas continue to respond in a manner very 15 different from our conventional use of cytotoxic 16 chemotherapy. You can see a very modest response 17 after one or two cycles, and if you continue the 18 drug for three, four, five, six, seven cycles you 19 continue to see responses and sometimes ultimately 20 achieve the conventional definition of a partial or 21 complete response for these patients. We need to 22 be sure that we don't discard some of these novel 23 compounds, which may be working by different 24 mechanisms from conventional cytotoxic 25 chemotherapy, by using too early a time point and 115 1 discarding a drug that may still have activity. 2 Again, I would concur very much with what 3 Bob said, that I have not seen any convincing 4 evidence -- just as Henry did looking at the brain 5 tumors, I have not seen any convincing evidence 6 that these sarcomas, when carefully defined, 7 ideally defined by a consistent chromosomal 8 translocation, behave any differently in children 9 from adults. I think that our decisions about 10 therapies and which therapies to employ and which 11 new agents to bring forward into clinical trial 12 should be based on the biology of the tumors 13 whenever possible. 14 I do need to comment, however, just 15 briefly because in our first session this morning 16 we heard I think some very encouraging comments 17 that we were going to use the efficacy to drive the 18 process much, much more than toxicity, and handle 19 toxicity perhaps appropriately through labeling, 20 and point out that I had an opportunity -- and this 21 is a trial that Dr. Benjamin is also involved in -- 22 to attend a meeting of the Recombinant DNA Advisory 23 Committee just two weeks ago. We have proposed a 24 trial for a gene therapy approach for metastatic 25 recurrent osteosarcoma, and the RAC was unwilling 116 1 to accept what we just discussed here today, that 2 the trial should be open to patients with 3 metastatic recurrent osteosarcoma regardless of 4 age. They felt that that approach could not be 5 offered to patients under the age of 18 until the 6 safety of the approach had been established in 7 patients greater than 18. So, the consensus that I 8 am getting from many of the individuals around this 9 table is not universally shared in the regulatory 10 community. 11 DR. SANTANA: Dr. Elias, do you want to 12 make any comments? 13 DR. ELIAS: Well, I would like to also 14 agree with what Dr. Benjamin has said. I mean, it 15 is quite clear that the individual histologies have 16 enormous differences in terms of response to the 17 conventional chemotherapy agents, and that has been 18 known for years, but the real difficulty has been 19 that no one institution and even groups of 20 institutions have sufficient numbers. 21 I think it is extraordinarily heartening 22 to see the amazing productivity of the GIST trials, 23 and the ability to mobilize a whole community 24 worldwide to actually target this. I think one of 25 the issues with sarcomas is, because they have a 117 1 more simplified genome or alteration in genome 2 relative, for example, to common epithelial tumors, 3 they can represent a situation of proof of 4 principle so that you have a more discrete pathway, 5 lesion, etc., what-have-you so that at least from 6 the pharmaceutical standpoint you could, in fact, 7 justify developing the drugs in these diseases. 8 DR. SANTANA: David? 9 DR. PARHAM: I agree that histology is a 10 very key thing in sarcomas, but I think it is also 11 equally important to take in the effects of grade 12 because within the confines of grade a low grade 13 sarcoma will do relatively the same if it is 14 localized, whether it is a synovial sarcoma or a 15 peripheral nerve sheath tumor or fibrosarcoma. So, 16 I think that would be a key thing to keep in mind 17 when we talk about treating things according to 18 histology. The histology may not be as important 19 as grade. 20 DR. SANTANA: Do you want to make a 21 comment, David? 22 DR. POPLACK: Yes, I am a little perplexed 23 by the comment that you made, Paul, regarding the 24 unwillingness to accept the similarity of a tumor 25 between adults and pediatrics. I think the issue 118 1 that you were speaking of with the RAC is a safety 2 issue, and that is not being necessarily addressed 3 in these discussions. Is that not the case? 4 DR. HIRSCHFELD: I would concur with Dr. 5 Poplack that the safety issues are not the topic, 6 but I think I will let Dr. Meyers answer but I 7 think he was just raising an axillary point. 8 DR. MEYERS: I think it is inseparable, 9 and this is the question that I was placing to 10 Henry. As we prioritize, moving forward with novel 11 agents, it will be both an issue of what agents 12 give us the greatest potential for benefit and what 13 is the risk/benefit ratio that we perceive for one 14 of these agents. I was encouraged to hear that we 15 were placing a strong emphasis on the first half of 16 that balance. I think that the emphasis was 17 perhaps over-weighted in terms of the risk side of 18 that equation at the hearing that I attended. 19 DR. SANTANA: I don't want to get into a 20 public discussion of this auxiliary issue because I 21 think the points have been made. I just think we 22 have to be sensitive that there are environmental 23 issues of current things that are happening in that 24 regard that I am sure influence a lot of these 25 discussions in other committees. Malcolm? 119 1 DR. SMITH: Several speakers have 2 addressed the issue of studying these tumors that 3 cross the pediatric-adult line together and the 4 benefits of doing that. There is a paradigm for 5 doing that, and that is the leukemia world with 6 acute promyelocytic leukemia. Since the early 7 1990's, the first intergroup trial for APL that 8 studied all transretinoic acid was amended to 9 include pediatric patients, and the then Pediatric 10 Oncology Group and the Children's Cancer Group 11 participated in that adult cooperative group-led 12 trial. 13 The current APL trial is examining arsenic 14 trioxide and one of the randomizations is 15 plus/minus arsenic trioxide. When that trial began 16 there wasn't much data concerning the safety of 17 arsenic trioxide in children, but those data have 18 emerged since the trial was initiated and the trial 19 is being amended so that children over five years 20 of age will be able to participate in the arsenic 21 trioxide trial randomization. So, there is a 22 paradigm for when there is a similarity at the 23 molecular level between the pediatric and the adult 24 condition, how those can be studied together 25 appropriately in the same clinical trial. 120 1 Maybe Paul and others can comment on this, 2 but CTAP has really been encouraging the bone 3 sarcoma and the soft tissue sarcoma committees and 4 COG and then the adult cooperative groups to work 5 together to study these cancers as they do cross 6 the adult-pediatric age distinction. That is an 7 artificial barrier and a number of efforts are 8 being made to try to stimulate such collaborative 9 research. 10 DR. SANTANA: I think those comments are 11 important. I think there has been a merging of the 12 consensus that at least when it relates to sarcomas 13 in adults and children there may be more 14 similarities since, at least in pediatrics, a good 15 portion of these patients are cured. And, the 16 challenge of the number of patients can only be 17 dealt with by a collaborative effort between adults 18 and pediatric studies. At least from my 19 perspective, I think that is the way to move in 20 this particular disease category. Any other 21 comments before we get to the questions? 22 [No response] 23 We have a series of questions that we have 24 to answer or give advice to the FDA on -- yes, we 25 may continue the discussion if somebody has another 121 1 question. Go ahead, Donna. 2 DR. PRZEPIORKA: Two questions for Dr. 3 Link. I enjoyed the slide, the ten-year old slide 4 that Dr. Benjamin showed about the dinosaur age, 5 and him mentioning that, you know, we still have 6 patients and although we like molecular therapy we 7 are still kind of in the dinosaur age. So, my 8 first question to you is how do you choose drugs 9 for your patients nowadays? When you have somebody 10 with recurrent disease and you have to treat them, 11 on what basis do you choose drugs to develop? 12 DR. LINK: In the recurrent situation -- I 13 think that would also apply to patients who have 14 very high risk disease. So, we view those as 15 similar categories, patients who are candidates for 16 more experimental therapies. Most of us I think 17 would participate in -- because, again, the 18 Pediatric Cooperative Group is such an 19 all-encompassing thing and most of our patients are 20 on clinical trials -- the majority of patients -- 21 well, I should back up, many patients who develop 22 recurrent disease have become candidates for Phase 23 II or Phase I trials that are usually 24 CTAP-sponsored trials so that they are entered on 25 those trials. 122 1 Now, one of the problems, particularly 2 with rhabdomyosarcoma, is that there are a lot of 3 active agents that have not proven useful when 4 added to the standard combination. It is an irony 5 which is unfortunate. So, the standard combination 6 which we use today, although it has been tweaked 7 many times, is the same combination of drugs that 8 has been available since the 1970's. I was a 9 fellow when we were using the same therapy. 10 Since then many drugs have come along 11 which show obvious activity, and many pediatric 12 oncologists feel that in a patient who develops 13 recurrence you sort of have to go through what is 14 available as treatment before you sort of begin to 15 use an investigational agent. I mean, I think that 16 is a philosophical problem rather than anything 17 else. Many of us would try a Phase I trial and 18 then put them on a standard agent. So, there are 19 some problems there. 20 In most of the other diseases, like 21 refractory Ewing's sarcoma or osteosarcoma, I think 22 that those patients are candidates for either a 23 biological or whatever experimental agent, mostly 24 in the context of a clinical trial. 25 DR. PRZEPIORKA: There is nothing specific 123 1 about the disease, however, that points you towards 2 one set of drugs versus another set of drugs 3 empirically? 4 DR. LINK: Well, there are now. For 5 example, the rationale for Gleevec is leading to 6 the initiation of a trial to study those patients. 7 So, for example, the results in GIST tumors, a very 8 refractory tumor that responds to this -- I suspect 9 that many patients will end up on a trial like 10 that. But I don't know how one would pick and 11 choose otherwise, except for the fact that they 12 have been prioritized one way or the other, either 13 just because it is the standard Phase II drug that 14 is being studied and that is the next candidate on 15 the list, or sometimes when there is a particular 16 drug of interest which is being prioritized by a 17 specific disease committee, they want to try that 18 but we often will have a specific retrieval 19 protocol mandated and that would be the next trial 20 that the patient would be eligible for. 21 DR. PRZEPIORKA: You mentioned a number of 22 new translocations that are very useful for 23 diagnostic purposes in the pediatric sarcomas, 24 especially in the Ewing's family, and the questions 25 I have for you are, are the functions of the fusion 124 1 transcripts known? If so, what are they? And, 2 secondly, you also mentioned a number of biologic 3 correlations, such as PGF and Her2 expression, but 4 I didn't hear anything about any preclinical data 5 that would suggest that inhibition of those 6 receptors actually has any function in inhibition 7 of growth of the pediatric sarcomas. 8 DR. LINK: I will address the second 9 question first because I have to think about what 10 your first question was. The mutated c-Kit 11 expression in Ewing's sarcoma and in Ewing's lines 12 has been shown in vitro. You can get abrogation of 13 cell growth, or whatever the appropriate endpoint 14 would be in vitro. So, there is something more 15 than just that it has the c-Kit and so we should 16 target it because it worked in GIST. I mean, there 17 is more data than that. I am not party to all of 18 it, but it is available. So, some of that 19 preclinical stuff has been done but not in all 20 tumors. 21 Her2 was your specific question? What was 22 it? 23 DR. PRZEPIORKA: The function of the 24 transcripts in the Ewing's family? 25 DR. LINK: Oh, I mean some of them are 125 1 known and they are clearly downstream. These are 2 transcription factors so that there is clearly a 3 whole myriad of downstream genes which are turned 4 down by these. Some of it is fairly well 5 characterized, but I think we don't know the whole 6 gamut of what the consequence of the translocation 7 is. Other people may want to comment on that but 8 some of it is known but I think the entirety of 9 what the consequence is unknown. 10 DR. SANTANA: There is great effort also 11 in creating some knockout models of some of these 12 transcripts and looking at what the phenotype is in 13 animals if you do those kind of experiments. I am 14 aware of some work in rhabdomyosarcoma in that 15 regard. 16 DR. POMEROY: Some of the transcription 17 factors are very difficult targets for soluble 18 small molecules. I think the value of tyrosine 19 kinase inhibitors is that these molecules are 20 relatively accessible on the cell surface, and 21 things that work within the nucleus are much more 22 difficult to target. So, although we can 23 understand in some cases specific biological 24 mechanisms of how tumors grow, they won't all be 25 equal in terms of how they might be attacked. 126 1 DR. KUN: Just one slightly related 2 question, in the case of APL it is not difficult in 3 adult oncology to get pathologists to send those 4 samples off for molecular diagnostics. My 5 experience has been in a number of hospitals for 6 adult sarcomas which are not that common but we see 7 them enough, pathologists are very reluctant. It 8 seems that they feel the gold standard still is 9 their histology or histopathology so when they call 10 it alveolar, embryonal or synovial sarcoma that is 11 sufficient, and we often don't get au courant 12 molecular diagnostics on these patients and we are 13 missing out on a lot of information I think that 14 leads to this. 15 So, one recommendation might be just to 16 stress the importance of these molecular 17 diagnostics, which will be essential in this time 18 where there are clearly adult and pediatric links. 19 But there is a lot of information that is not being 20 gotten because pathologists -- certainly off 21 clinical trials, which most of them in the adult 22 world still are, aren't getting this information. 23 DR. SANTANA: David, do you want to 24 follow-up on that? 25 DR. PARHAM: Right. First, I would like 127 1 to address one question, that is, there are at 2 least 20 upstream and downstream modulators of that 3 fusion gene that you just asked about. There is a 4 cottage industry of literature appearing on that. 5 But as to the question of diagnosis, this 6 is becoming a greater question with each passing 7 year because pathologists are becoming more and 8 more efficient at arriving at a histologic 9 diagnosis using fine-needle aspiration biopsy and 10 similar things which have much less morbidity for a 11 patient. So, if the criteria for putting a patient 12 on study is simply histologic diagnosis, we are 13 going to be in a situation where we are getting 14 less tissue, not more. I don't think pathologists 15 have any problem with sending tissue off, but there 16 always is that question of how much is adequate. 17 There have been some recent things coming 18 out from CIOP. I think it is going to be in the 19 upcoming Ped. Onco., about how to handle 20 fine-needle aspirations for biologic studies. But 21 the point is pathologists are willing to send it 22 out, but it is always a question of how much do you 23 get and this is an issue that is going to have to 24 be addressed if we want to do biologic studies. 25 DR. SANTANA: I know from other 128 1 conversations with investigators that I have had 2 that the COG has had a major recent effort in their 3 sarcoma working group, and Mike may want to 4 comment, specifically looking at this issue in a 5 subcategory of patients with soft tissue sarcoma to 6 establish a biology type protocol to try to resolve 7 this issue. So, it is done in a group-wide effort 8 because we just don't know -- people don't know 9 where to send the samples; they don't know who to 10 contact. So, they are trying to do it in a 11 collaborative effort. So, I know that at least in 12 the pediatric community there is a major effort 13 being placed on this particular question in the 14 soft tissue sarcoma field. 15 DR. LINK: I would just make the comment 16 that the answer to your question is you have to 17 start doing things the way pediatrics do it. 18 [Laughter] 19 Because basically what you do, first of 20 all, you get a monopoly on the market. So, all the 21 kids are seen in places where they are all put on 22 clinical trials. The clinical trial becomes the 23 standard of care. Then you up the ante and say you 24 want the kid to go on a clinical trial; you have to 25 get the tissue or the kid is not eligible for the 129 1 trial; or you have to get a letter from your mother 2 or something like that which says why you didn't 3 get the tissue. That is how we do it. That is why 4 if you look at kids with a variety of solid tumors 5 -- neuroblastoma, Ewing's sarcoma -- we are now 6 recapitulating what went on in lymphoblastic 7 leukemia where, admittedly, it is much easier to 8 get the stuff. But we wouldn't let a patient on a 9 trial unless you got the cytogenetics and got all 10 the stuff that you need to risk stratify the 11 patient. 12 That is happening in neuroblastoma now. 13 We don't even know how to treat -- this is now 14 standard of care. I mean, Sue could comment on it 15 better than I, that we don't even know how to treat 16 a kid with neuroblastoma unless you do the biologic 17 studies because that determines the outcome. So, 18 that is an editorial pitch but maybe that is the 19 answer. 20 DR. COHN: Yes, I was just going to say in 21 terms of neuroblastoma, I mean the advantage that 22 we have in neuroblastoma is we define the therapy 23 according to the molecular genetics. We don't care 24 about their stage and age anymore. Now, in this 25 new biology study we are going to be obtaining 1P, 130 1 11Q, 17Q, 14Q. So we have hopefully the whole 2 gamut covered. But the difference is that you need 3 the information to determine the therapy. So, 4 there is the carrot and the stick. You can't 5 decide how to treat this patient without knowing 6 all the genetic abnormalities. 7 DR. LEVIN: What happens if you have all 8 the genetic abnormalities and the therapy that you 9 envision is something that requires two 10 experimental drugs from two companies? To give you 11 a good example, osteosarcoma -- it looks like what 12 you should do is you should take a PDGF receptor 13 inhibitor like STI -- Gleevec, and you should take 14 a pan SARK RTK inhibitor, which would get the 15 receptor as well. So, maybe that is what you 16 should do. So, the question is how can you 17 expedite that kind of a process and move it 18 forward? But that would be based on genetic 19 information; that would be based on signaling logic 20 and it is testable. It is more valuable probably 21 than testing one of those receptors. 22 DR. HIRSCHFELD: I think excellent 23 diplomatic skills is going to be the way to solve 24 that one. 25 [Laughter] 131 1 DR. ELIAS: Just one comment. I was 2 intrigued in terms of the discussion of what do you 3 do for recurrent Ewing's? Namely, you take the 4 drugs that previously showed activity. I think we 5 are getting into that problem in adult medicine. I 6 mean, in breast cancer we have drugs that are 7 developed and approved for second-line, for 8 third-line. We are working on fourth-line. 9 Ultimately, while that is a very good 10 strategy for the pharmaceutical companies and for 11 the drugs to develop a niche to get approved, what 12 it does also do is mean that in a sense there is 13 some mandate to require that a patient, before they 14 get to an experimental agent, has had their first, 15 second, third, fourth, whatever. And, this is an 16 increasing problem, and I think there is not data 17 that one has to use a particular sequence. On the 18 other hand, this is what is being used, such that 19 only two percent or so of adults actually go on 20 clinical trial, and I think that is even going to 21 get worse as we get into this fixed sequence of 22 trials based on what is FDA recommended and, 23 therefore, what the insurance companies are going 24 to pay for. 25 DR. HIRSCHFELD: I just want to clarify, 132 1 we don't recommend or endorse trials particularly. 2 We allow them to proceed on the basis of safety 3 evaluation. 4 DR. ELIAS: I am sorry, I am not 5 disagreeing with that but, you know, they are 6 approved for second- or third-line use. In other 7 words, they have developed a specific niche so that 8 taxotere is approved for second-line use in 9 non-small cell lung cancer or breast cancer, and so 10 forth. NTA, I believe, is being developed for 11 fourth-line Zeloda refractory patients because 12 Zeloda is now approved for third-line. So, we are 13 getting an increasing, sort of regimented, set of 14 treatments and these are the approved indications 15 and the insurance companies are not paying for 16 anything that isn't approved in a sense. So, there 17 is a difficulty there. 18 DR. PAZDUR: This is a manifestation of 19 accelerated approval. Okay? And, this is a game 20 that many of the drug companies that come in play 21 with us to define what is an unmet medical need in 22 order to get their drug approved on basically the 23 least amount of information possible and the 24 smallest population, and to try to get a more and 25 more refractory patient population. For example, 133 1 if we approve at ODAC a drug in third-line breast 2 cancer or something with a 10 percent response 3 rate, the following week we have an army of people 4 coming in wanting to know what is the minimum 5 response rate it will take and the smallest number 6 of patients and in the most refractory population. 7 We are trying to discourage this strongly, 8 believe me. It doesn't serve anybody any good in a 9 situation -- yes, it could get a drug approved but 10 as far as moving the science forward I really 11 question it. It may not even be doing the drug any 12 good because as you study drugs in more refractory 13 populations the chances of missing activity are 14 also there. 15 We are trying to re-encourage people to 16 take a look at accelerated approval as it was 17 meant, that there would be randomized trials that 18 were ongoing in a reasonable indication such as a 19 first-line indication, and if it looked like their 20 drug was better in a randomized setting against the 21 standard therapy, they would get approved on a 22 surrogate endpoint, awaiting survival data to come 23 up. 24 But this is a manifestation I think of 25 companies looking at what are niche areas to get 134 1 their drug approved on a single-arm trial and we 2 really are trying to discourage that to get more 3 randomized trials in place. It gives better 4 characteristics as far as toxicities of the 5 therapies. We actually get the drug approved in an 6 indication that is meaningful. 7 DR. ELIAS: I am sorry, these are 8 randomized trials. I mean, there is nothing wrong 9 with the science or anything like that. But what I 10 do find difficulty with is just that by the time 11 you get to a mandate to be able to use experimental 12 therapies you have already exhausted all of your 13 standard approaches. That, at least 14 philosophically, I think is the wrong way to go 15 about it. 16 DR. LEVIN: And those can be mediocre 17 therapies. 18 DR. ELIAS: Yes. 19 DR. BENJAMIN: But this issue is not an 20 FDA issue. This is an insurance reimbursement 21 issue. The problem is the way medical care is paid 22 for and the pediatricians have learned the right 23 way. They do have a monopoly. Everybody goes on 24 trial and it gets paid for. The adult oncology 25 community has not learned how to do that. So, the 135 1 only things that get paid for are the 2 non-experimental treatments and that encourages an 3 anti-clinical trial approach. So, I think that is 4 not an issue that we can solve here. 5 DR. SANTANA: I think we were commenting 6 on that same thing on this side of the table, that 7 the issue is that pediatrics is a completely 8 different model of how to surmount some of these 9 problems. 10 Questions 11 I want to go ahead and get to the 12 questions before we go to lunch because, if not, 13 the FDA will tell me I haven't done my job. So, 14 let's go ahead and address the questions. They are 15 on the second page of the handout. Some of these I 16 think we may have already answered but we will go 17 through them. 18 Specifically for sarcomas, which is what 19 we have been discussing for the last two hours, 20 what general principles could be used to relate 21 sarcomas in adults to sarcomas in children? Mike? 22 DR. LINK: I think I would endorse your 23 extra slide, your spare slide, as sort of a 24 starting point, you know, that if we can define a 25 molecular basis, which is a unifying theme, that we 136 1 should take advantage of it. 2 DR. SANTANA: I would go a step back and I 3 would say that I think for sarcomas still the issue 4 of histology, grade and molecular characterization, 5 in that order, are the principles that define the 6 similarities and the differences. I am taking what 7 you said and extending it a little bit further. 8 David? 9 DR. PARHAM: I would emphasize grade over 10 histology because I think grade is sort of an 11 expression of biologic status of that tumor using 12 morphologic parameters. So, to me, that should be 13 one of the overriding things because a low grade 14 tumor, if completely excised, is cured. A high 15 grade tumor, even if completely excised, probably 16 deserves additional therapy. I know there are some 17 caveats between grading pediatric tumors and 18 grading adult tumors but I don't think these should 19 exclude the possibility that it cannot be done and 20 that we should really concentrate on grading 21 sarcomas in order to stratify them for new 22 therapies, given the fact that a large proportion 23 will be cured adequately by surgery. 24 DR. FINKLESTEIN: I am not the scientist 25 that is sitting around this table but it seems to 137 1 me that I have to disagree with my colleague to my 2 left -- both colleagues to the left because if we 3 are really talking about new generation, it seems 4 to me the microscope may have played a significant 5 role in the past but is not molecular biology the 6 role for the present? And, is grade moving 7 backwards and molecular biology moving forwards or 8 is it vice versa? 9 So, I do like Steve's slide. I think 10 there is a role -- I can give you an example. 11 Embryonal rhabdomyosarcoma, if it sits in the 12 bladder, looks like grapes and if it sits somewhere 13 else -- and you will have to correct me, Mike, if I 14 am wrong -- and it may be the same molecular lesion 15 but if it sits somewhere else it is going to look 16 different. Grossly it will. So, I really would 17 like us to reconsider do we use the microscope or 18 do we use and call upon our molecular biologists 19 for the present and the future? 20 DR. SANTANA: Jerry, I think my comment is 21 that they are complementary because we know a lot 22 about the former; we are just beginning to learn 23 about the latter and how it correlates to the 24 former. So, I don't think they are exclusive; they 25 are complementary. And, you may be right, we know 138 1 a lot more about the molecular characterization of 2 alveolar rhabdomyosarcoma but we still don't know 3 what that means in terms of what drugs we should be 4 using for those patients. So, that is why my 5 comment was more of complementary, not abandoning 6 one and bringing a new one forward. Bob? 7 DR. BENJAMIN: Yes, actually, Victor, I 8 would like to support your contention. I think we 9 may get to molecular definition and that is well 10 and good, but we don't have molecular definition 11 for most of the tumors that we treat. 12 But I would strongly disagree with Dr. 13 Parham that grade by itself unifies. Grade unifies 14 aggressive behavior. High grade osteosarcoma and 15 high grade embryonal rhabdomyosarcoma are totally 16 different diseases. They are treated with 17 different drugs; they respond differently. If we 18 just say, well, everything high grade gets mixture 19 A-B-C we are going to mix the things which are 20 important. We know that there are differences in 21 these tumors that are based on their biology and 22 right now the best definition of the biology is 23 what it looks like under the microscope. 24 DR. SANTANA: Mike? 25 DR. LINK: Yes, I think that we are 139 1 quibbling over something which is probably not 2 useful because grade is irrelevant in 3 rhabdomyosarcoma. You know, I mean we can argue 4 this some other time but it is certainly 5 irrelevant. It is very relevant in osteosarcoma 6 but we all acknowledge that those are different 7 diseases when you have a periosteal osteosarcoma. 8 So, it is relevant in a group of tumors which is 9 very unlikely to occur in children. So, in the 10 soft tissue sarcomas where there is no molecular 11 definition, which are the diseases that actually 12 don't occur in children anyway so it is a 13 non-issue, you can use whatever it takes. But the 14 point is that where we have actually got the 15 tissues and we have the diagnosis -- and we are not 16 going to say that the pathologists -- I think you 17 are taking it a little too far, Jerry, to say that 18 the pathologist has out-served his -- first of all, 19 we will never get another tissue sent for another 20 study ever again -- 21 [Laughter] 22 So, what was it? delicate politics or 23 diplomatic? Even I recognize that that wouldn't 24 have been approach and I am not known for my 25 delicate politics particularly. So, I wouldn't 140 1 have said that, but I think that we have a fairly 2 good one from our chairman that, you know, we have 3 to encompass everything that is important and where 4 we have the molecular things and we have an agent 5 that really is related to the molecular event, then 6 obviously that is paramount. When we are talking 7 about an agent that has more to do with grade or 8 more to do with proliferative rate for which there 9 may be an agent, that is the unifying thing rather 10 than a particular kinase that it inhibits. Then 11 maybe grade will be the paramount thing in terms of 12 determining where it should be studied. 13 DR. FINKLESTEIN: I accept your diplomatic 14 interpretation of my remarks. 15 [Laughter] 16 DR. SANTANA: That is the other difference 17 between pediatricians and adult oncologists! 18 Larry? 19 DR. KUN: There are differences here and 20 one of our charges -- I mean, if you look at is it 21 appropriate to study new agents, then grade is 22 clearly an indicator of prognosis. On the other 23 hand, if you are really looking at how you identify 24 new agents, it might be applicable or justified for 25 a particular new agent, then grade is almost 141 1 meaningless. I mean, it may be quite important in 2 your initial prognosis but it is not going to 3 determine what agent you are anxious to try. 4 DR. POMEROY: I would like to add that the 5 issue of obtaining tissue for developing a 6 molecular taxonomy is not a trivial one, as my 7 pediatric brain tumor colleagues will attest in our 8 recent meetings. If there are diseases, which is 9 true for all pediatric brain tumors at this point, 10 where you don't have a molecular marker that 11 impacts a treatment decision, then so far it has 12 not been mandated that we collect tissue to be part 13 of a clinical trial. 14 In the current ethical and regulatory 15 environment, it sounds like we are going more in 16 the direction that we need consent, we need to have 17 approval to be able to collect these tissues and we 18 cannot ethically put patients on trials where we 19 don't have a decision to be made for collecting 20 their tissue with tissue collection as mandatory 21 because that would be coercing them to give tissue 22 to get treatment when they don't have any tangible 23 benefit for themselves. So it is a tricky issue 24 that we have been struggling with a lot in 25 pediatric brain tumors and I can only imagine in 142 1 all tumors as well. 2 DR. SANTANA: Paul? 3 DR. MEYERS: It is not quite as ethically 4 suspect as that. If you are talking about a Phase 5 III trial which involves conventional agents, then 6 denying a patient participation in the trial is not 7 denying them potential benefit. They can have all 8 of those agents without trial participation. So, 9 in fact, for the osteosarcoma and the Ewing's 10 sarcoma trials we have required specimen submission 11 for entry into the clinical trial. If they decline 12 to submit tissue we are not denying a child 13 potential benefit. They can receive all of the 14 therapy according to but not enrolled in the trial. 15 In fact, with this particular strategy we surprised 16 ourselves with the increase in the amount of tissue 17 submission in both of those clinical entities. We 18 are now getting excellent submission of biological 19 material. 20 I do think we are just a tiny bit off the 21 topic though because the question was what 22 principles should we be using to relate sarcomas in 23 children and adults, and I think that they are 24 stated, that we should use histology and molecular 25 pathology, but I have not heard anyone disagree 143 1 with the principle that we should encourage, design 2 and carry out trials which ignore the age of the 3 patient as much as possible and concentrate on the 4 biology of the tumor. 5 DR. SANTANA: Donna? 6 DR. PRZEPIORKA: Actually, that was almost 7 essentially the comment that I wanted to make and 8 both speakers very eloquently stated that an 9 osteosarcoma in a kid isn't like an osteosarcoma in 10 an adult, and no one has presented any information 11 that sarcomas in children are different than 12 sarcomas in adults if you get down to the 13 histology. Even if we go on to question B, any 14 specific type of sarcoma doesn't appear to be -- 15 or, there was no data presented to suggest that 16 different types of sarcoma are different in adults 17 and pediatric patients, unless anybody has any 18 other information that wasn't stated. 19 DR. SANTANA: Because nobody disagrees, we 20 have actually covered A and B together. Mike? 21 DR. LINK: I agree. I just want to say 22 that the caveat is that, as a statistician would 23 say, no difference doesn't mean that there is no 24 difference. It just means you haven't detected it. 25 So, we know that older patients do less well. 144 1 There is a bunch of theoretical reasons why that 2 is, you know, blaming the oncologist and blaming 3 the tumor and everything in between. But it may be 4 the tumor and I think that until we have array data 5 that shows that they really are identical, that all 6 the downstream effects of having that EWS-FLY1 7 transcript are the same in an older patient and a 8 younger patient I don't think that you can -- I am 9 not as certain because there has to be some reason 10 why an 18-year old treated by a pediatric 11 oncologist in the same center as a 10-year old does 12 less well. Then, remember, the IRSG data that I 13 showed you, those aren't patients that are treated 14 by adult oncologists. Those are patients treated 15 in the same centers, by the same people, with the 16 same willingness and putatively the same compliance 17 with therapy, although that is an issue -- not 18 entirely the same host but we think that there is 19 not much difference between an 18-year old and a 20 12-year old in terms of their tolerance for 21 therapy, yet the outcome is quite different. 22 So, I think that you may or may not be 23 right, but I think I still agree with Victor. You 24 know, I don't think that changes the answer to the 25 question. 145 1 DR. HIRSCHFELD: Then the question, once 2 again, is when should studies be undertaken, not 3 when should the therapy necessarily have findings 4 extrapolated from one population to another but 5 should it be studied. 6 DR. SANTANA: So, I think we have answered 7 A and B, unless Steve or Richard want to address 8 the issue differently. 9 DR. BENJAMIN: Well, I think that 10 rhabdomyosarcoma needs to be studied specifically 11 in a pediatric population because that is where it 12 exists. 13 DR. SANTANA: Oh, yes. 14 DR. BENJAMIN: And, I would be happy to 15 include adults on the pediatric trial but it is a 16 pediatric disease and we are not going to be able 17 to study it in adults. There must be studies in 18 children on these tumors. For Ewing's 19 osteosarcoma, you know, they go across the bridge. 20 We are going to continue to study them. I think we 21 should study them in the same way and learn 22 whether, in fact, we can determine what the factors 23 are which make the 18-year old different from the 24 10-year old. But I am certainly not going to treat 25 the 18-year old patient with rhabdomyosarcoma on an 146 1 osteosarcoma protocol because he is 18. I mean, 2 his therapy is determined by the disease. 3 DR. SANTANA: We agree. The last 4 question, I had trouble with this one, Steve. You 5 may have to give us some more guidance in trying to 6 answer this. So, are there pediatric sarcomas that 7 have an adult counterpart that is not commonly 8 defined as an adult sarcoma but as some other type 9 of adult malignancy such as carcinoma? Help me 10 through that. What do you want from us on that? 11 DR. HIRSCHFELD: Okay. Every once in a 12 while, because of historical reasons or taxonomic 13 reasons based on histology, a disease ends up with 14 a different name. So, since we asked about 15 extrapolation from adult sarcoma to pediatric 16 sarcoma, we wanted to look at the obverse question, 17 is there a pediatric sarcoma that, now that we have 18 a different understanding of biology, has some 19 other name? 20 I will give a potential example. A 21 potential example might be the GI leiomyosarcomas 22 which have been called different things in 23 different eras and now we might think of it as 24 something different. So, we just wanted to make 25 sure that if we are reviewing proposals from 147 1 companies and they say they want to cover tumor X 2 that we would say, oh, well, we were advised every 3 time we saw tumor X to think that there also might 4 be a role for asking for pediatric studies. So, it 5 is just an attempt to be complete. 6 DR. SANTANA: Others may want to comment; 7 I don't have any comment. David? 8 DR. PARHAM: Dr. Benjamin actually raised 9 that question indirectly because we do have an 10 adult tumor, which is rhabdomyosarcoma, pleomorphic 11 rhabdomyosarcoma, and that diagnosis has gone in 12 and out of favor. But I think from a biologic 13 basis they are nothing like pediatric tumors but I 14 would bet that question as to whether you would put 15 those tumors on a rhabdomyosarcoma program. They 16 are really defined only by the fact that they make 17 muscle but biologically they are different. 18 DR. BENJAMIN: But I think that is also 19 true of embryonal and alveolar. That is one of the 20 reasons why some of the differences in the older 21 rhabdomyosarcoma patients come out, because there 22 is a higher percentage of the bad-acting group, but 23 the bad-acting group is defined as the bad-acting 24 on the therapy given for the entire group. It may 25 well be that we need different therapy for alveolar 148 1 rhabdomyosarcoma than we need for embryonal 2 rhabdomyosarcoma, and alveolar can be defined 3 molecularly so that is a group that needs specific 4 targets. 5 I agree completely with you about the 6 pleomorphic. I don't know how many of them you see 7 in pediatrics. We see relatively few still in 8 adults, at least at Anderson. Again, that depends 9 on definitions and what you accept as the 10 definition of pleomorphic rhabdomyosarcoma. I 11 mean, we do see some and, frankly, we treat them 12 more like an adult sarcoma. You are right. We 13 would like a pediatric rhabdo. 14 DR. HIRSCHFELD: I would like, Victor, to 15 have just one clarification before we break for 16 lunch, and that is in part B, number 4. I just 17 want to make sure that we are not being too 18 efficient. Specifically, should gastrointestinal 19 carcinomas be excluded or included? I just want a 20 little discussion on that point. 21 DR. LINK: Colon cancer? 22 DR. HIRSCHFELD: Or stomach cancer, yes. 23 DR. SANTANA: A rare entity, Steve. 24 DR. HIRSCHFELD: No, no, I want it 25 addressed. What we are looking for here is if 149 1 someone comes in and says we are studying a 2 gastrointestinal carcinoma, that we would be 3 comfortable saying no, we have advice and we feel 4 that there is no mandate for pediatric studies, 5 that this should be waived. 6 DR. SANTANA: I think 4 specifically are 7 very, very diseases -- 8 DR. HIRSCHFELD: But not in adults. 9 DR. SANTANA: In kids. That is what I am 10 telling you. So, in terms of attributing the 11 waiver, it has to apply. It is just a very few 12 number of patients. 13 DR. HIRSCHFELD: Right. 14 DR. SANTANA: Logically, I would have to 15 assume, and I may be incorrect both in my logic and 16 my assumption -- logically, I would have to assume 17 that the counterpart is the adult disease. 18 Anthony? 19 DR. ELIAS: Just one caveat to that, 20 namely, if you are talking about chemoprevention of 21 gastrointestinal tumors in familial syndromes, for 22 example, there certainly may be a rationale for 23 that particular situation to want to be able to 24 study children as well. But otherwise I totally 25 agree that the actual carcinoma is not a pediatric 150 1 disease. 2 DR. SANTANA: That is a very good point, 3 Anthony. Yes, that is a very good point. Donna? 4 DR. PRZEPIORKA: The other situation is 5 biologics and inhibitors of biologic markers, 6 addressing C in a slightly different approach in 7 that if you have a gastrointestinal tumor with a 8 specific tyrosine kinase and that tyrosine kinase 9 is also present in another pediatric tumor and does 10 the same thing, that inhibitor should probably be 11 tested in pediatric patients. For example, the 12 PDGF inhibitors are being tested now in prostate 13 cancer. There is no prostate cancer in kids but, 14 clearly, PDGF is a big thing in pediatric tumors. 15 So, yes, that drug should be tested in pediatrics. 16 DR. SANTANA: But ultimately the end 17 result is what is the medical indication being 18 sought that would drive that. Am I correct? 19 DR. HIRSCHFELD: It is going to depend on 20 how we, again, define the word "indication." 21 Although conventionally we have stuck to histologic 22 definitions, we are open and evolving in terms of 23 how we define that. 24 DR. LINK: But isn't that circumstance a 25 little bit of a non-issue because putatively if you 151 1 have shown that it works in a pediatric tumor you 2 have already studied it in kids, and then you want 3 to open the indication up to adult tumors. Maybe I 4 have it wrong. The other way around? All right, I 5 withdraw my comment. 6 DR. BENJAMIN: But the comment regarding 7 the PDGF inhibition is that we do not know in which 8 circumstances PDGF is the critical driving factor 9 of the malignancy. We know that it is present in a 10 number of different tumors of vastly different 11 histology. We also know that c-Kit is present in a 12 number of tumors of vastly different histology. We 13 don't know that inhibition does anything except 14 just where c-Kit is critical to the development of 15 the tumor. It may or may not; it may work in some 16 and not others. We may find out that PDGF is 17 present in osteosarcoma and in Ewing's sarcoma but 18 that blocking it has an effect in osteosarcoma and 19 not in Ewing's sarcoma. We need to do all of those 20 studies to find out, and we may come back a few 21 years from now and say, okay, we know that this 22 pathway, if it is ever blocked, will always be 23 therapeutic and we can define an indication based 24 on a pathway. But until we have the data I don't 25 think we can say that. 152 1 DR. SANTANA: Pat? 2 DR. REYNOLDS: I think with respect to 3 that issue the pediatric preclinical testing 4 consortium that Malcolm spoke about is going to be 5 a valuable asset to providing data, and if there 6 was consideration from FDA as to whether or not 7 they should grant a waiver and there was a common 8 target among pediatric tumors, presumably that 9 consortium could quickly address that and tell you 10 if there was preclinical basis for saying no, this 11 should be studied or there was a lack of 12 preclinical data and, therefore, a waiver should be 13 amended. 14 DR. SANTANA: Dave? 15 DR. POPLACK: I think that Steven quite 16 eloquently represented that in his slide, that the 17 presence of a common pathway doesn't make or define 18 the need for a trial. There has to be a definition 19 which encompasses that the pathway is related to 20 the development and progression of the disease. 21 DR. SANTANA: Howard? 22 DR. FINE: And that just bets the issue 23 that I often talk about. Some of us do believe 24 that as molecular targeting becomes more than just 25 a catch-all phrase but becomes a reality, there 153 1 will be drugs ultimately, hopefully, that will be 2 approved based on their target rather than their 3 histologic subtype in the future, but that gets 4 down to this term we talk about as far as target 5 validation. The mere presence of a target is very 6 different from target validation. Again, that gets 7 back to what I spoke about before. I think that is 8 where the academic investigators -- it is incumbent 9 upon you if you are interested in a particular 10 tumor type, such as a pediatric tumor, it is 11 incumbent upon us to validate that target in order 12 to make a case for that drug to come into trial. 13 DR. SANTANA: Last comment, Joe? 14 DR. GOOTENBERG: From the biologics 15 viewpoint, what we are really talking about here -- 16 and from a drug viewpoint also but biologics I 17 think is where we are going to come into these 18 issues -- we are talking about one narrow question. 19 When a manufacturer comes in and says they want a 20 license whether or not we grant them a waiver for 21 doing pediatric studies, not licensing or this or 22 that, and the question that I think, Donna, you 23 posed very well and that we would like some 24 consensus on from the group would be that if there 25 is a common pathway but the two diseases are very, 154 1 very different historically should we or should we 2 not grant the waiver? Just like the PDGF that you 3 are talking about. 4 DR. PRZEPIORKA: The answer would be no. 5 DR. SANTANA: Anybody disagree with that? 6 [No response] 7 Then we are done for the morning. We will 8 try to reconvene at 12:45. It says 12:30 on the 9 schedule but since we ran a little bit late we will 10 meet at 12:45. 11 [Whereupon, at 12:00 noon, the proceedings 12 were recessed, to be resumed at 1:58 p.m.] 155 1 AFTERNOON PROCEEDINGS 2 DR. SANTANA: This afternoon is a 3 continuation of the Pediatric Subcommittee of the 4 Oncology Drugs Advisory Committee. Since we 5 introduced ourselves this morning and I don't think 6 there is anybody new at the table, we don't have to 7 go through the introductions again. 8 As is customary, if there is anybody in 9 the audience who wants to address the committee, we 10 have some time allotted for an open public hearing. 11 Anybody in the audience who wishes to address the 12 committee, please come to the microphone and 13 identify yourself. Nobody? Let's go ahead and get 14 started. 15 The first session this afternoon is going 16 to address issues of extrapolation for lung tumors 17 and neuroblastoma, and Frederic Kaye will be our 18 first speaker. Frederic? 19 Perspectives on Lung Tumors and Neuroblastoma 20 DR. KAYE: Thank you. 21 [Slide] 22 I was asked last week to come over here 23 and just give a brief overview of adult lung 24 cancer. It is a little unusual term for me to put 25 up there, but that is what I will try to do. 156 1 [Slide] 2 Starting off by looking at the different 3 histological types of lung cancers, there are 4 different ways you can approach it but one common 5 way is to separate out these lung tumors into 6 neuroendocrine lung tumors and non-neuroendocrine 7 lung tumors. Of the neuroendocrine lung tumors the 8 most common type is small cell lunch cancer. I 9 might add that the term for small cell may be a 10 little bit of a historical term because it is also 11 called oat cell carcinoma. There is some feeling 12 that at least part of that might have been crush 13 artifact. These cells are very fragile in general, 14 and they have a tendency for crush artifact on the 15 edges. I know Dr. Matthews, who is a 16 well-respected lung cancer pathologist over at NIH, 17 felt that many of these tumors were medium sized 18 epithelial carcinomas and the oat cell phenotype 19 might have been contributed by crush artifact. 20 At any rate, the one distinguishing 21 feature that we had noted in our lab about a decade 22 ago is that small cell lung cancers were very 23 tightly correlated with RB gene inactivation. This 24 is unusual because there is almost no other tumor, 25 that is, besides the pediatric retinoblastoma tumor 157 1 and then, in the case of a subset of sarcomas that 2 might arise in the phenotype of familial 3 retinoblastoma. So, it is distinctly unusual to 4 see a tumor that will have in excess of 90 percent 5 have mutationally targeted the RB gene. 6 Now, other common types of neuroendocrine 7 lung tumors are pulmonary carcinoid tumors, 8 non-small cell lung cancer with neuroendocrine 9 phenotype -- these are usually called large cell 10 tumors with neuroendocrine phenotype but they can 11 be in almost any histological type of non-small 12 cell. The feeling is that since this is the 13 largest subtype of lung cancer seen in about 10 14 percent of these lung tumors have striking markers 15 for neuroendocrine phenotype, there was a sense of, 16 well, how do you distinguish these tumors from 17 small cell lung cancer, and perhaps they may have 18 similar biological features, response to treatment, 19 etc. 20 That really hasn't resolved itself. They 21 do have distinct, it appears, genetic background 22 from small cell lung cancer, and the response to 23 treatment is higher but not quite like small cell. 24 So, there is some confusion as to exactly what the 25 meaning is of non-small cell lung cancer with 158 1 neuroendocrine phenotype and it gets back to the 2 cell of origin of these tumors. There is still 3 some controversy as to whether or not there might 4 be a stem cell that is targeted that can 5 differentiate into different types of lineages that 6 might include neuroendocrine and non-neuroendocrine 7 types or the specialized differentiated cell is the 8 one that is targeted for the mutational events. 9 This is partly driven by the fact that 10 pathologists have seen different histological types 11 of lung cancer in the same tumor biopsy, where they 12 might see tumor types that will have small cell 13 features along with either squamous or 14 adenocarcinoma. It is not a large number but there 15 is a small subset. 16 These are the main types that are thought 17 to be lung tumors. There are other types that 18 might be included. The ones that I put up there 19 are primary undifferentiated carcinomas and, of 20 course, peanut type tumors tend to occur in 21 non-smokers and in younger patient types, and some 22 of them may have, if we look for them, 23 characteristic molecular characterizations that 24 might put them into other categories. Nonetheless, 25 since they have neuroendocrine features and since 159 1 we are not exactly sure how best to manage them, 2 and many of these patients can present with 3 metastatic disease just like small cell lung 4 cancer, there has been a sense that as long as the 5 lung tumor has a neuroendocrine phenotype there is 6 a certain aggressive biological behavior associated 7 with it, with the exception of typical pulmonary 8 carcinoid. There are certain chemotherapy drugs, 9 mainly of the cisplatin category, that are utilized 10 for treatment. 11 Of course, there is carcinoid of unknown 12 primary that can sometimes have neuroendocrine 13 phenotype or germ cell type element that can be 14 central or mediastinal, and those are also often 15 recommended to be lumped together to treat them 16 with small cell lung cancer-like regimens. 17 I put up here just for interest that there 18 are few clearly small cell lung cancer-related 19 diseases which are non-lung tumors. The best 20 characterized are what we call extrapulmonary small 21 cell. Small cell lung cancer -- and we have seen 22 it and maybe a number of pathologists have seen it 23 -- can arise primarily in the prostate gland; in 24 the bladder; in the cervix; in the thyroid and a 25 variety of other organ tissues. They are not 160 1 common but when they do appear they seem to have a 2 lot of biological features and they include the 3 characteristic feature of mutationally inactivating 4 the RB gene. So, there is a close connection with 5 this, and why they are arising in other tissues and 6 not in the lung is still uncertain. 7 The other tumor type is Merkel cell tumor, 8 which is a characteristic cutaneous tumor but it is 9 highly aggressive and even though it might present 10 locally in the skin, we often recommend that after 11 that is excised that they receive a small cell lung 12 cancer type regimen. 13 The other box I included there is purely 14 animal model type information. Using the clue that 15 the small cell lung cancer has unusually targeted 16 in almost every case the RB gene, if you make a 17 mouse that is defective for the RB gene and in that 18 sense make a familial retinoblastoma mouse, the 19 mouse doesn't develop retinoblastoma tumor although 20 you can under other experimental circumstances. 21 But, what the mouse gets is a series of 22 neuroendocrine tumors which are really fascinating 23 and is a subject of some work, and they get 24 spectral tumors which overlap between the MEN1 and 25 MEN2 syndromes. They certainly get pituitary 161 1 tumors with almost 100 percent penetrance, but they 2 will also get medullary thyroid cancers. They will 3 get islet cell tumors of the pancreas. They will 4 get pheochromocytomas, etc. 5 So, there is an important link there with 6 mutationally targeting the retinoblastoma gene and 7 even though the RB gene is essential for transit 8 through the cell cycle in all eukaryotic cells, 9 there is something specific about it as a single 10 hit leading to retinoblastoma tumors in humans, 11 which is a neuroendocrine tumor which does resemble 12 in cell culture and other things in retinoblastoma. 13 But also in animal models it is telling us that 14 there is an important link that is still undefined. 15 [Slide] 16 I am saying lung tumors because many of 17 these aren't considered lung cancer but they are 18 tumors that arise in the lung -- taking just the 19 malignancy ones and not benign lesions, of course, 20 the most common is what is called non-small cell 21 lung cancer, just the general category of many 22 different histologic types -- adenocarcinoma, the 23 squamous cell, large cell undifferentiated are the 24 most common type. 25 I put bronchoalveolar carcinoma as an 162 1 additional one because even though histologically 2 it belongs to adenocarcinoma, it has a number of 3 very distinct clinical features, although also some 4 histologic features. 5 As I said, about 10 percent of these 6 non-small cell lung cancers will have a 7 neuroendocrine phenotype. They will appear maybe 8 as two different populations or as a population 9 that partially resembles neuroendocrine cells, and 10 they will especially express a number of markers. 11 It is very popular to use synoptifisine and a 12 variety of other neuroendocrine markers to tell 13 this. 14 There is some thought that maybe this 15 could guide treatment, and there is a push to 16 perhaps consider cisplatin-like regimens although 17 these regimens are pretty much used routinely now 18 for all types of lung cancer. So, that is really 19 not as much of an issue as it was five or ten years 20 ago. 21 Mesothelioma is a very different type of 22 lung tumor. It is of interest to us because it 23 will have a specific genetic marker that will help 24 in molecular diagnosis and that should be coming 25 out in the future. 163 1 There is pleuro-pulmonary blastoma which 2 is primarily a pediatric disease. Then, of course, 3 there are sarcomas of a variety of types that can 4 occur primary to lung but they are unusual and we 5 might see only an occasional patient a year. 6 [Slide] 7 I will just jump ahead and just show one 8 slide on the RB gene because it makes a few 9 important points. First of all, identifying that 10 about 90 percent, or slightly in excess, of small 11 cell lung cancers have targeted the RB gene for 12 mutational inactivation, and these can be just 13 single codon substitutions, but the biochemical 14 result is that they function as null. It is as if 15 they weren't there. 16 One interest we had in the lab is that 17 there are about five or ten percent of small cells 18 that still retain wild type RB function. We 19 thought it may be a DNA tumor virus or a variety of 20 other exogenous that might be targeting it. It 21 turned out that an upstream gene, called the p16 22 gene, was the one that was mutating those. When we 23 went back to look at non-small cell lung cancer we 24 found that almost all of non-small cell lung cancer 25 had targeted the p16 gene and not RB. Small cells 164 1 target RB and not p16. They both target the same 2 pathway that is referred to as the RB/p16 pathway. 3 You only need to knock out one or the other but not 4 both. So, 100 percent of lung cancers are 5 targeting this pathway. It is still the undefined 6 question as to why this neuroendocrine type is 7 picking RB gene while the non-neuroendocrines pick 8 p16. 9 This pathway is particularly interesting 10 because it converges on a set of enzymes that by 11 themselves will entirely drive the cell cycle, 12 particularly the transition between G1S and then 13 the transition through mitosis as well, and this is 14 the cyclin-dependent kinase family. A variety of 15 other epithelial tumors -- you can find in 16 melanoma, for instance, a specific mutation in a 17 cyclin-dependent kinase molecule. That is shown on 18 the slide as CDK. The mutation is exclusively at 19 the site where the p16 inhibitor will bind to CDK. 20 So, when you mutate the CDK molecule, its enzyme 21 kinase activity is completely intact. But what it 22 can no longer do, it can no longer bind to p16. 23 So, it is as if p16 wasn't present. So, it 24 resembles other tumors where p16 has been mutated. 25 When you look at these tumors you find 165 1 that RB is wild type, p16 is wild type, but the 2 single mutation is in this one residue on the 3 cyclin-dependent kinase molecule. So, this shows 4 you that while there is really an RB, CDK, p16 5 pathway, you only need to mutate one but not any of 6 the other ones to disrupt the pathway. And, 7 certain melanomas will target CDK. Other melanomas 8 will target p16 non-small cell lung cancer, and 9 many other cancers will target p16. Small cell 10 lung cancer targets RB. 11 Finally, cyclin-D is the other partner in 12 this, and cyclin-D overexpression has been noted. 13 As a matter of fact, it was first identified as a 14 translocation partner in parathyroid tumors and it 15 was initially called the Prad-1 gene. It has also 16 been found in a number of other circumstances, 17 BCL1. In breast cancer it is overexpressed. In 18 those where it is overexpressed, it appears again 19 as if that is the only target in this whole 20 pathway. 21 That is again particularly interesting 22 because there is a series of papers, one that was 23 reported yesterday, in which you look, again, in 24 animal models which tell us a lot. Looking at the 25 tumor patterns to identify these pathways are so 166 1 much more precise, I believe, than in vitro 2 laboratory experiments, and when you look at a 3 mouse model, they have known for about five or six 4 years that you can have a cyclin-D1 null mouse and 5 that mouse grows up normally without any tumors. 6 It has a few other really specific histological 7 abnormalities but when they cross the cyclin-D1 8 mouse with a transgene mouse that gives a high 9 penetrance for breast cancers -- when they cross it 10 with a ras transgene mouse that will give you 11 breast cancers in a wild type background otherwise, 12 when they cross in a cyclin-D1 null mouse you get 13 no breast cancers. So, it is strongly arguing that 14 this mutated ras is acting again through a 15 CDK-cyclin-D1 pathway and that gives important 16 clues. 17 When they take a neu mouse, which is 18 another word for the Her2 neu since neu was 19 originally described in neuroblastoma tumor, that 20 also gives a high penetrance of breast cancers. 21 When they cross that with a cyclin-D1 null mouse, 22 they get no breast cancers, again arguing that at 23 least a new pathway appears to be also funneling 24 in, in some way to cyclin-D1. 25 On the other hand, when they crossed that 167 1 with a classic myc transgene mouse that give a high 2 incidence of mammary tumors, they saw no reduction 3 in the number of breast tumors in the cyclin-D1 4 null mouse. So, it is giving some clues and using 5 this key point of the G1S phase as a funnel but, 6 again, giving some clues as to different pathways 7 and perhaps now incorporating ras, neu and other 8 things. That is sort of a general editorial 9 comment. 10 [Slide] 11 So getting back to adult lung cancer, a 12 number of labs -- our lab and other labs have tried 13 to collect what might be some defining phenotypes, 14 and there are a number of caveats with this. One 15 caveat I might add is instead of myc overexpression 16 it is myc amplification of one of the different 17 members. Myc overexpression will be considerably 18 higher. 19 But, again, there are also caveats with a 20 number of these percentages put here but they give 21 a rough idea, showing how small cell lung cancer, 22 the neuroendocrine, is genetically different from 23 non-small cell lung cancer. If you look at certain 24 genes, they are very similar when you look at other 25 genes. Certainly, small cell lung cancer is very 168 1 different from other lung neuroendocrine tumors 2 such as carcinoid if you look at some genes but not 3 others. 4 [Slide] 5 This shows the genes which are probably 6 best implicated in lung cancer, RB, p16, CDK-cyclin 7 pathway -- myc, there seems to be, I think, 8 substantial strong circumstantial evidence, 9 likewise for ras, p53. Also, I believe for p10 10 signal transduction pathway and perhaps for erbB. 11 All these tumors are notably characterized 12 by chromosomal instability and have a high 13 incidence of telomerase activation. On the right 14 side there is a long list of many other candidate 15 lung cancer genes, particularly candidate genes on 16 the short arm of chromosome 3. I might add that 17 c-Kit is activated in a large number of these lung 18 tumors but, as far as I can tell -- I am not 19 involved in these studies, there really haven't 20 been dramatic responses with the Gleevec agent but 21 I am not the source to report on that. 22 [Slide] 23 I am going to end right there just with a 24 brief introduction, and these are just some 25 off-the-cuff thoughts I had when I sent in the 169 1 slides yesterday. Extrapolation will need to focus 2 on cell biology and genetics. As has been 3 mentioned here before, you know, assuming that we 4 know enough for rational therapies, the specific 5 treatment. 6 The overwhelming thing that hits you when 7 you look at lung cancers is the neuroendocrine 8 phenotype because they do appear to be a 9 characteristic feature of a large number of adult 10 lung cancers. If you take small cell lung cancer 11 by itself, they say it would be the fourth or fifth 12 most common cancer. But this decision, as I say, 13 would require a case by case evaluation. I suppose 14 that is part of what this committee is here for, to 15 look at that. 16 If you take specific treatments you can 17 decide with it is one of these unusual fusion 18 translocations that you only see in a certain type 19 of tumor, like in some of the subsets of sarcomas 20 etc., or you can take something like the p53 gene 21 which is seen in almost all epithelial cancers and 22 my sense is if there were a way to express wild 23 type p53 function in these tumors, you would stop 24 them and you would induce either growth arrest or 25 cell death very consistently, and that would be an 170 1 overwhelming choice, just in my opinion. So, if 2 you had a therapy that you could consistently show 3 would reactivate wild type p53 functioning cells, 4 that would be a good choice for a whole range of 5 what you might otherwise think are disparate 6 histologic subtypes. That doesn't exist right now 7 but the hope will be that APL, which is probably 8 the best paradigm because there is a certain type 9 of treatment linked with a certain translocation, 10 might be applicable. 11 One last sort of side comment, and this 12 has to do with the discussion that we had before 13 about molecular diagnostics in sarcoma, I just want 14 to emphasize, not as a pediatrician and not working 15 in sarcoma, that I feel that it really is 16 critically important to get the molecular 17 information despite the practical issues, and I am 18 sure in brain cancer they are exponentially 19 important. But if you look even in the APL 20 situation, as far as I understand it, there are a 21 number of different types of translocations that 22 can be seen in APL with different binding partners, 23 and not all APL leukemias respond to retinoic acid 24 and there is some suggestion that the specific 25 translocation in APL is the one that really will 171 1 target to you which ones of those APLs will respond 2 to retinoic acid. Using that, again, as a 3 paradigm, it continues to emphasize the need to be 4 collecting this data and putting it into the 5 database. 6 Then, the last thing is that these are 7 still the tentative days of directed treatments, as 8 has been pointed out here before. There is a track 9 record of empirical success, and we just have to 10 keep in mind that a lot of our rational therapies 11 will appear empirical down the road. Thanks. 12 DR. SANTANA: I think we will have time 13 for discussion and questions later. I want to ask 14 Pat Reynolds to discuss issues of neuroblastoma as 15 they may relate to some adult counterparts. 16 Neuroblastoma and Small Cell Carcinoma of the Lung: 17 Differences and Similarities 18 DR. REYNOLDS: I would like to thank Vic 19 and Steve and Karen for asking me to talk on this 20 topic. This particular topic is one that has 21 fascinated me since I was a medical student, and 22 that is really, is there any relationship between 23 small cell carcinoma of the lung, a tumor that 24 occurs in older adults, and a pediatric tumor, a 25 neuroblastoma? 172 1 I work at a children's hospital and I 2 couldn't find any of my colleagues to make any 3 profound statements about lung cancer -- 4 [Laughter] 5 [Slide] 6 -- but since my children's hospital is 7 located in Hollywood, I relied on some of the local 8 talent to point out to us that lung cancer is 9 primarily a disease of smokers. 10 [Laughter] 11 That is clearly one of the major 12 differences between neuroblastoma and small cell 13 lung cancer, and that is that it is a disease in 14 which the etiology of small cell cancer is almost 15 exclusively related to tobacco use whereas, 16 clearly, that is not related, at least as far as we 17 know, in any way to the etiology of neuroblastoma. 18 [Slide] 19 These tumors share a common ancestor. If 20 you look on this rather complex slide, the neural 21 crest stem cell gives rise to a whole variety of 22 different neuroendocrine cells within the body. In 23 fact, it is this ability that is required of the 24 neural crest stem cell to migrate out and spread 25 throughout the body that is thought to confer some 173 1 of the biological features of tumors derived from 2 the neural crest stem cell, namely, the propensity 3 for rapid and widespread metastasis early in the 4 course of progression. That is certainly true for 5 neuroblastoma and small cell lung cancer, as you 6 will see. 7 Neuroblastoma probably arises from the 8 neural crest stem cell, or from a cell that is just 9 immediately downstream from it, because there is a 10 variety of these different phenotypes that can come 11 out in a differentiated pattern from neuroblastoma. 12 Small cell lung cancer is thought to arise from one 13 of the neural crest stem cell derivatives that 14 gives rise to these APUD cells, the various 15 neuroendocrine cells that spread out in certain 16 organs and these cells, termed as Kulchitsky's 17 cells, are thought to be potentially the cell of 18 origin for small cell lung cancer. So, in that way 19 these two tumors do share a common ancestor in the 20 neural crest stem cell. 21 [Slide] 22 Now, if one looks at the staging for 23 neuroblastoma, we see a very distinct set of stages 24 that are clearly related to prognosis. In fact, 25 these stages are probably very directly related to 174 1 biology as well such that localized tumors, even 2 the partially resected stage 2s, will do quite well 3 with no chemotherapy, indicating that they are a 4 distinct biological subgroup from the more 5 widespread tumors and the more aggressive tumors. 6 There is no counterpart to these in small 7 cell lung cancer. What we see in small cell lung 8 cancer are basically two stages that are defined by 9 the adults, one of which is extensive, widespread 10 disease and the other is more local, regional 11 disease and they do have a prognostic impact. The 12 more localized tumors do significantly better than 13 the more widespread tumors. Those probably 14 correspond to these two stages in neuroblastoma, 15 high risk stage 3 which is a bad biological feature 16 of local, regional tumor and then the more 17 widespread or completely widespread stage 4s. As 18 you will see, this is our major problem in 19 neuroblastoma, the stage 4 patients that present 20 over one year of age. 21 For completeness, another staging 22 component that was initially identified by Chick 23 Coop, Audrey Evans and Dan DiAngelo, the stage 4S 24 tumors, are widespread disease that can 25 spontaneously regress with no therapy at all. That 175 1 occurs only in infants in neuroblastoma and clearly 2 has no counterpart that we know of in small cell 3 lung cancer. 4 [Slide] 5 If one looks at sites of disease, clearly 6 there are differences in the sites of disease, 7 mainly the primary tumor. The small cell lung 8 cancer presents in the lung, whereas neuroblastoma 9 presents anywhere where there is sympathetic 10 nervous tissue but a very common site of it to 11 present is in the adrenal. 12 However, if one looks at the metastatic 13 sites, there is almost complete overlap. 14 Neuroblastoma and small cell lung cancer both 15 frequently present at diagnosis with bone marrow 16 metastases and that is a common site of recurrence 17 for both of these. They both commonly have liver 18 metastases and whereas small cell lung cancer can 19 present -- and in fact sometimes the initial 20 presenting symptoms, in fact, the first small cell 21 patient I ever saw as a medical student, that is 22 how he presented with a seizure from a CNS 23 metastasis, that then through a chest x-ray showed 24 us that he had a small cell lung cancer. That is 25 not seen in neuroblastoma where CNS metastases at 176 1 diagnosis are exceedingly rare. 2 However, at relapse, now that we are 3 starting to control the disease, we are, 4 unfortunately, starting to see a significant 5 increase in the number of CNS metastases in these 6 relapse patients. So, there is some degree of 7 overlap in that site of disease as well. 8 [Slide] 9 In neuroblastoma we see a spectrum of 10 differentiation. You can see this in an individual 11 patient if you serially biopsy particularly stage 12 4S patients when they are aggressive. You can see 13 highly undifferentiated and metastatic cells that 14 mature through these differentiated phenotypes with 15 pseudorosettes all the way to a benign 16 ganglioneuroma, which is very reminiscent 17 histologically of a sympathetic ganglion. 18 One does not see this kind of 19 differentiation in small cell lung cancer and so 20 there is clearly a difference between them there. 21 As you will see, therapeutically we have been able 22 to apply this differentiation in neuroblastoma and 23 it probably can't be applied in small cell. 24 [Slide] 25 If one looks at localized disease in 177 1 neuroblastoma, this is essentially a surgically 2 cured disease and, as you see from these data from 3 the cooperative group, with a fairly good long-term 4 follow-up period, these patients with no 5 chemotherapy are doing quite well. So, this is 6 another clear difference between small cell and 7 neuroblastoma in that localized disease patients do 8 quite well. 9 [Slide] 10 I would like to turn to some of the 11 molecular features, in particular the myc oncogenes 12 which, as you remember, were initially identified 13 by Michael Bishop and Harold Varmus, the v-myc gene 14 being the viral version of the cellular homolog 15 c-myc. Manford Shrawin, in Michael Bishop's lab, 16 was then able to look at neuroblastomas which were 17 well-known to have some sort of amplified gene 18 because they carried double minutes, and the MYCN 19 gene, which has homologous sequence to c-myc and is 20 found to be amplified in a large proportion of 21 neuroblastomas. Almost half of the high risk 22 patients have amplified c-myc. 23 Then Marion, now working with John Minner, 24 over at NCI Navy, were able to do exactly the same 25 thing. Knowing that there were amplified sequences 178 1 of myc genes in small cell lung cancer, they were 2 able to fish out another homolog of c-myc, the 3 L-myc gene. So, those are the three myc genes, one 4 being primarily derived from neuroblastoma but, as 5 you will see, that is also amplified in some small 6 cell patients, and one being derived primarily from 7 small cell lung cancer. 8 [Slide] 9 This amplification occurs, as you will 10 see, at the chromosome 2 region where NMYC is 11 located. It is believed that there is an excision 12 of the gene which leads to plasmids that turn into 13 double minute chromosomes and that those can be 14 then integrated back into chromosomes as 15 homogeneous staining regions, but regardless of the 16 cytogenic manifestations, the multiple copies of 17 the gene are seen in about 25 percent of all 18 neuroblastoma primary tumors and, as you will see, 19 that has significant prognostic outcome 20 relationship. This amplification basically 21 provides a large amount of NMYC RNA which then 22 overcomes the short half-life for NMYC and 23 generates a large amount of NMYC protein. 24 [Slide] 25 This shows, from a study that was reported 179 1 by Marilee Schmidt in JCO in 2000, one of the more 2 dramatic demonstrations of the impact NMYC 3 amplification in neuroblastoma. We are looking 4 here at patients who have stage 4 neuroblastoma 5 that present as infants. Those patients that have 6 no NMYC amplification get relatively modest, not 7 superintensive, chemotherapy and do extremely well 8 whereas, in spite of whatever aggressive therapy 9 you can try to get into these infants, those 10 patients with MYCN amplified disease do extremely 11 poorly. 12 [Slide] 13 In my laboratory we have been spending a 14 lot of time trying to characterize drug resistance 15 mechanisms, and this is some work I wanted to share 16 with you from Nina Kashlava where she has looked at 17 a variety of different neuroblastoma cells lines, 18 and here are just some representative ones 19 established at diagnosis, then some established at 20 PD-IND at progressive disease during ararfrin 21 induction therapy, then the PD-BMT are cell lines 22 that were established at time of progressive 23 disease after myeloablative therapy. Shown on this 24 axis is the amount or resistance of these cells to 25 various agents. The two platinum compounds, 180 1 carboplatinum and cisplatinum, nafolem, then 2 doxorubicin in red, and finally in yellow 3 etoposide. As you see, we go from diagnosis where 4 there is extreme sensitivity, as you will see and 5 we also see this extreme sensitivity in the 6 patients, as you go through the various stages of 7 therapy, as this therapy gets more intensive those 8 recurrent tumors that we then place in a culture 9 have a sustained, very high level of drug 10 resistance. 11 We looked at a variety of different 12 mechanisms for this and we weren't able to pin 13 anything on it. But, what Nina did was then to 14 examine p53 function and what she found is 15 summarized in the next slide. 16 [Slide] 17 The loss of p53 function, primarily by 18 mutation which is virtually never there -- only two 19 percent of all neuroblastomas as primary tumors 20 have mutation at p53, but in these cell lines that 21 are highly drug resistant there was an incredible 22 correlation with loss of p53 function, again, 23 mostly by mutation. If she knocked out, as you see 24 in the yellow squares here, p53 function by 25 transducing in the 16EC6 protein, then on this axis 181 1 we see the LC90 for a variety of drugs. The red 2 bars indicate the clinically achievable levels and 3 we go from responsive cell lines that can be killed 4 by clinically achievable levels to, as you see in 5 the yellow symbols, those p53 non-functional lines 6 are virtually never killed by clinically achievable 7 levels of the drug. 8 Now, there are some exceptions with new 9 agents which are p53 independent, but for those 10 agents -- the alkylators, the platinum, etoposide, 11 the agents we commonly use for neuroblastoma -- a 12 loss of p53 function appears to be one of the 13 mechanisms by which drug resistance occurs. 14 [Slide] 15 If we go to the bottom of this table, we 16 see that that can be related back to small cell 17 lung cancer where p53 mutations are present in a 18 high proportion of these. From the literature it 19 is not clear whether these are mutations detected 20 at diagnosis or after chemotherapy. In talking 21 with Dr. Kaye, it was clear that he feels that a 22 number of these tissues were procured at various 23 points in time during therapy so this may be a 24 mixed bag and not just at diagnosis. 25 With neuroblastoma, again, there is a 182 1 large body of literature showing that essentially 2 if you take it all and do a meta-analysis two 3 percent of the tumors are mutated. We are starting 4 to see from the cell lines in the tumors we are 5 looking at clearly more than 20 percent, we don't 6 know what the exact number is going to be, in the 7 post-chemotherapy neuroblastomas p53 mutations. 8 So, just on that basis alone, one of the 9 things I am going to try and do is draw a very 10 strong parallel between relapse neuroblastoma and 11 small cell lung cancer in terms of its behavior. 12 In fact, the clinical behavior of these diseases is 13 quite identical in that they both do poorly 14 eventually with chemotherapy. Even though there is 15 some response, both relapse neuroblastomas and 16 small cell lung cancer are incurable diseases. 17 There are other parallels that one can see 18 in molecular biological features. MYCN 19 amplification, as we said, occurs in half of the 20 neuroblastomas but it also occurs in small cell 21 lung cancer, with at relapse or at diagnosis. 22 Unlike small cell, we don't see c-myc amplification 23 or L-myc amplification in neuroblastoma but both of 24 those genes can be amplified in small cell as well. 25 Neuroblastoma is an adrenergic tumor and, 183 1 therefore, secretes catecholamines quite 2 frequently. That is not seen in small cell lung 3 cancer. But other neuroendocrine features, 4 chromogranin expression, PGP9.5 expression, NSE, 5 the leu-7 antigen -- a variety of neuroendocrine 6 features are seen both in neuroblastoma and small 7 cell lung cancer. They appear to have different 8 tumor suppressor loci, however, whereas deletion of 9 3P is the most common deletion seen in small cell, 10 and it is a deletion of the short arm of 1 in 11 neuroblastoma, although there are some P1 deletions 12 that are reported in small cell lung cancer. 13 [Slide] 14 This is a curve showing the CCG data at 15 two periods of time. initially the 1978 to 1985 16 studies in the CCG were stage 4 neuroblastomas 17 presenting over one year of age, and then the data 18 that was obtained in the period from 1986 to 1995. 19 You can see that there is a clear-cut and 20 statistically significant improvement in survival 21 for stage 4 neuroblastomas presenting at greater 22 than one year of age. 23 There are probably two major reasons for 24 this, one of which is the application of very 25 intensive therapy, as I will show you in a moment. 184 1 The other is the application of 2 differentiation-inducing therapy. This shows you 3 an NMYC amplified neuroblastoma in culture, which 4 is then treated, in the panel on the right, with 10 5 micromole retinoic acid and it shows you the 6 remarkable growth arrest and differentiation that 7 can be achieved with that agent. There is no known 8 parallel to this in small cell lung cancer and we 9 don't know of any differentiation inducers that are 10 effective like this. 11 Going clinically with that, in the CCG-389 12 study we were able to show that the combination of 13 intensive meyloablative therapy, supported by 14 autologous bone marrow transplant, or ABMT, 15 followed by 13 cis-retinoic acid gives the highest 16 survival rate that you can get for this particular 17 form of neuroblastoma, the high risk disease. That 18 is, in fact, what is now being applied essentially 19 worldwide for treating this tumor -- myeloablative 20 therapy followed by a differentiation inducer. 21 Other types of therapies to go along with 13 22 cis-retinoic acid are being tested. For example, 23 monoclonal antibody therapy is being tested in 24 Europe and will soon be tested here, in the U.S. 25 [Slide] 185 1 Now, if one looks at the response rates to 2 induction chemotherapy for both small cell lung 3 cancer and neuroblastoma, they are identical. Both 4 of these diseases with combination chemotherapy get 5 a response rate of 80-90 percent. There are 6 clearly more CRs that are achieved in 7 neuroblastomas than there are in small cell lung 8 cancer but they both get an almost identical 9 response rate. 10 [Slide] 11 What I find even more striking is to look 12 at the clinically activity of the drugs. Shown in 13 yellow are all those agents that are used as 14 standard parts of therapy, components of standard 15 therapy now for neuroblastoma and small cell lung 16 cancer. You see that those agents are identical. 17 By empirical clinical studies the exact same agents 18 have been shown to be useful in these two diseases. 19 Now, other agents that are used less 20 frequently in these diseases, such as ifosfamide, 21 topotecan and paclitaxel -- we don't know for 22 paclitaxel; certainly we know for ifosfamide and 23 topotecan that they are active in both of these 24 diseases. For melphalin there is not enough data 25 to say whether it is active in small cell lung 186 1 cancer, presumably it would be. And, 13 2 cis-retinoic acid presumably would not be but we 3 don't have that data. 4 But in spite of those two, there is this 5 enormous overlap and I would like to propose this 6 as one paradigm for trying to look at diseases, 7 disparate diseases. Here we have a lung cancer in 8 adults, caused by smoking, and an embryonal neural 9 system tumor and if you look at the pattern by 10 empirical studies of drugs that have been found to 11 work, they are almost identical. If you take that 12 as a paradigm for trying to apply the Pediatric 13 Rule I think it makes a lot of sense, that if by 14 empirical observation we find that the pattern of 15 agents has been the same perhaps the next agent 16 that is going to be tested, unless it is targeting 17 some specific pathway not known to be in the 18 particular tumor, could be also useful. 19 One can extend this beyond what we are 20 talking about here to diseases such as embryonal 21 carcinoma of the testes where there is a 22 considerable overlap with this pattern as well of 23 the agents being active, and we also know that the 24 same is true in terms of p53 -- embryonal carcinoma 25 of the testes virtually never mutated at diagnosis 187 1 but at relapse tumors do have p53 mutation. So, I 2 think we can draw a lot of parallels. 3 [Slide] 4 I wanted to end by saying we can go back 5 the other way. These are data from my laboratory 6 looking at fenretinide, a cytotoxic retinoid, in 7 combination with safingol. You can see in these 8 neuroblastomas which include post-BMT relapse 9 neuroblastoma that these cell lines are totally 10 resistant to virtually every agent we can throw at 11 them. And, if we use this combination of 12 fenretinide plus safingol we get this striking 13 multi-log cytotoxicity at dose levels here, at 14 least in tissue culture, that are totally non-toxic 15 for normal myeloprogenitors and fibroblasts. So, 16 we are very interested in developing this therapy. 17 In fact, the NIH is supporting developing it 18 through a grant. 19 [Slide] 20 My colleague has looked at this in small 21 cell lung cancer, and this is one of several lines 22 he has looked at, and he sees exactly the same 23 striking synergy with these agents. So, it may be 24 that agents that are developed in the pediatric 25 community could be then brought back forward to 188 1 adult diseases, and I think we ought to think about 2 going both ways in this area when we are trying to 3 link diseases. 4 [Slide] 5 So, I would like to end by saying that I 6 think there are substantial similarities between 7 neuroblastoma and small cell lung cancer. These 8 similarities include metastatic sites, the 9 neuroendocrine markers and antigens that are 10 expressed on these tumors; their molecular 11 biological features; their initial response rates 12 to chemotherapy, which I think is very important; 13 and especially their profile of clinically active 14 drugs. 15 Based upon those, I think these 16 similarities suggest that drugs developed for 17 either disease should be strongly considered for 18 clinically testing in the other. Thank you. 19 Discussion 20 DR. SANTANA: Thank you, Pat. We have 21 time now for discussion. I want to get started now 22 myself. The question kind of relates to this 23 analogy that you are proposing between aggressive 24 neuroblastoma and aggressive small cell lung 25 cancer. The question is, yes, there may be many 189 1 similarities but the similarities may be truly 2 coincidental; may have to do with what ultimately 3 causes cancer in a very simplistic way and totally 4 unrelated to the two diseases independently. But 5 there are also a lot of differences. So, the 6 unifying principle is not quite there because there 7 are as many differences between the two diseases as 8 there are similarities. So, I wanted you to expand 9 a little bit on that and where you think those two 10 cross so that we can then propose when somebody 11 comes to the agency with small cell lung cancer 12 drug development that they consider neuroblastic 13 tumors in that development too. So, do you want to 14 tackle that one? It is a very general question, 15 not very specific. 16 DR. REYNOLDS: What I am trying to do here 17 is to show that certainly there are differences but 18 there are also similarities and the real question 19 at hand, as I understand it from the FDA's 20 consideration of the Pediatric Rule is whether or 21 not there is enough evidence to link an adult 22 cancer to a pediatric cancer to indicate that a 23 study is warranted. And, I believe that based upon 24 not only the biological features, even you consider 25 that there are differences between them, but 190 1 especially based upon the history that we find by 2 empirical drug development in small cell lung 3 cancer and neuroblastoma, given that so many of 4 these agents have been shown to be active for both 5 of these tumors, that would be, at least from my 6 perspective, strongly suggestive that a study 7 should at least be considered in neuroblastoma for 8 an agent that is being brought forth for as an 9 indication for small cell also. 10 DR. SANTANA: Jerry? 11 DR. FINKLESTEIN: Pat, if you did the same 12 kind of comparison with, say, malignanct melanoma 13 what would you show, and have you done it? I say 14 that because that is another neural crest cell 15 tumor which many of us have grown up thinking in 16 terms of neuroblastoma. What do you think you 17 would find? 18 DR. REYNOLDS: Well, I haven't done that 19 comparison at the depth I would like to, to answer 20 that question but there are certainly some 21 similarities. They are both neural crest cell 22 derived. They share some common antigens. In fact 23 the NTGD2 antibody which was developed for 24 neuroblastoma therapy is being tested in melanoma 25 and in small cell as well. So, there is clearly 191 1 some overlap there. The striking similarities in 2 terms of similar histology is there; metastatic 3 pattern is not there, nor is the response rate or 4 identical profile of drugs. But I think there is 5 overlap there and it certainly should be looked at 6 carefully. 7 DR. LINK: You had me going until the last 8 business about this comparison of the drug 9 sensitivity profile as a way of relating the two 10 because, you know, everything in pediatrics is 11 sensitive to dactinomycin, cyclophosphamide and 12 Adriamycin. So, if you took those three drugs and 13 said that Ewing's is like rhabdomyosarcoma because 14 they both respond to those three drugs you would 15 sort of set back all of this splitting that we have 16 been doing in defining a molecular underpinning for 17 the specificities of the cancer. In fact, there 18 are those people who say small round cell tumor -- 19 just give them cyclophosphamide and Adriamycin and 20 it should go away, like our pathologists at my 21 institution when they refuse to do these molecular 22 tests. 23 [Laughter] 24 So, I think you have a very cogent 25 argument. And then the last slide with the 192 1 fenretinide -- but I think we have to be very 2 careful about sort of putting on record that 3 anything that responds to three drugs is likely to 4 be the same because ultimately those are the only 5 three drugs that are useful for most adult tumors, 6 and platinum I guess. So, you are lumping a lot of 7 things where we would be going both ways. That is 8 my fear. 9 DR. REYNOLDS: I agree with you totally, 10 Mike. My point with this was not that we could 11 lump small cell lung cancer in with neuroblastoma 12 and do trials. Obviously, you can't do 13 myeloablative therapy in these 16-year old patients 14 and there is a whole variety of reasons why you 15 can't treat them the same. What I am suggesting 16 is, is an agent that is active in one likely to be 17 active in another? That is the real question that 18 we are getting at for the purposes of this 19 committee. I think based upon those data, there is 20 a strong history that would suggest that if an 21 agent was tested in one and was active, it was 22 likely to be active in the other. 23 DR. COHN: I was just going to say that I 24 think, again, we have to think about this. It is a 25 neuroendocrine tumor much like a melanoma is a 193 1 neuroendocrine tumor and there are certainly some 2 antigens that are similar; there are some 3 amplifications, some genes that are similar; and 4 there are certainly presumably some biologic 5 pathways in terms of cell growth and development 6 that are probably similar. I think, once again, we 7 need to take into account Steve's slide. Rather 8 than trying to lump small cell and neuroblastoma 9 together, I think it is much more important to say 10 what is the drug? What is the pathway targeting? 11 Is the pathway prevalent in both small cell lung 12 cancer and neuroblastoma? If the answer is yes, 13 for example, GD2 is an antigen and you could 14 potentially use that particular therapy in both 15 diseases. Then, you know, I think it makes sense. 16 But I agree that to just say, you know, that a drug 17 that destroys DNA and basically kills cells the way 18 chemotherapy does to be sensitive in both probably 19 isn't a rational approach. 20 DR. ELIAS: Just a comment because it is a 21 struggle I think. The devil is in the details, so 22 to speak, in terms of when we try to define a 23 treatment according to its biologically targeted 24 activity. If we take the example, for example, of 25 Herceptin in breast cancer and we say, okay, now 194 1 with FISH we think we have a negative predictive 2 factor in breast only, if that is not amplified 3 then Herceptin doesn't seem to work. If it is 4 amplified, it does work. However, when we think 5 about it working, it really only works, let's say, 6 20 percent of the time, at least by itself. 7 So, the point really is, number one, 8 should we spend an enormous amount of time trying 9 to validate the target by looking at the assay 10 used? Number two, it is clear that even when the 11 target is validated the results are very 12 spectacularly heterogeneous, and we don't 13 understand why a Her2 overamplified breast cancer 14 doesn't response. The third, it is also being 15 defined specifically for Herceptin and it is fairly 16 clear -- well, no, it is not fairly clear but it, 17 hopefully, will become clear that if you use a 18 different targeting method or a different molecule 19 you might, in fact, get a completely different 20 answer in terms of what is important biologically 21 in that pathway. 22 So, yes in small cell we have GD2, we have 23 GD3, we have a number of overlaps particularly in 24 neuroendocrine type pathogens that are relevant in 25 neuroblastoma and melanoma, for that matter -- I 195 1 mean, we can go and look at a lot of similarities 2 and if we have a specific target it might be 3 relevant but I think it comes back to what Howard 4 said, that we have to be absolutely sure that these 5 are meaningful in that disease and that they are 6 going to have a biological effect. And, I think 7 that is where we fall apart a little bit. I don't 8 think we know that. 9 DR. SMITH: I have a biological question 10 for Dr. Kaye. You know, Pat shared his data with 11 the p53 and how that decreases the 12 chemosensitivity, yet, in this very chemoresponsive 13 or initially chemoresponsive cancer there is a high 14 percentage of p53 mutations and not in 15 neuroblastoma in response to the same drug. Is 16 there any explanation of how you can have such a 17 high percent of p53 mutations and yet be so 18 chemoresponsive, as opposed to the situation in 19 neuroblastoma where once those mutations appear you 20 lose much of your chemosensitivity? 21 DR. KAYE: Small cell lung cancer is a 22 really tough disease for oncologic research because 23 it tends not to be a surgical disease. It is 24 almost never a surgical disease. So, there is 25 almost no primary material to deal with, and a lot 196 1 of the biopsies are on needle aspirates or small 2 bronchscopic biopsies. So, I am not so certain 3 that how much of the mutational data for the tumor 4 was done pre-therapy and you are going to be skewed 5 a little bit more for more advanced stage. 6 There is an interesting point with myc 7 amplification. I showed a very low incidence of 8 myc amplification in lung cancer. Earlier data 9 showed a much higher incidence. There is a 10 question as to why you don't see as much myc 11 amplification now as you did earlier on. There 12 have been a few studies that tried to say -- again, 13 that data was often done after patients had been 14 subjected to chemotherapy -- so, there was an 15 argument, and I am not sure how tenable it is, that 16 cytoxin-Adriamycin, which was the most common 17 regimen that was used for small cell lung cancer 18 earlier, might be pressuring these cells to undergo 19 myc amplification while cisplatin-etoposide might 20 not have the same genotoxic stress and that might 21 be a reason why. But, again, a lot of this data is 22 done post-treatment. So, it is plausible that p53 23 is not targeted early. 24 The counter thing is that, again, if you 25 target especially the RB gene, those cells with 197 1 wild type p53 undergo apoptosis almost immediately. 2 So, it is impossible to find any sample or cell 3 line that has a mutant RB relating to wild type p53 4 function. So, I think there is no answer to that 5 question. 6 Questions to the Committee 7 DR. SANTANA: Any other comments? If not, 8 I want to go ahead and start with the questions so 9 we can remain on time. As I did earlier this 10 morning, I want to take a first pass proposal to 11 answer the first question. 12 This relates to what general principles 13 should be used -- I am going to change the 14 question, not "could" but "should" be used to 15 relate malignancies in adults to neuroendocrine 16 malignancies in children? 17 I would propose the following, clearly, as 18 has been demonstrated well by Pat and Dr. Kaye 19 today, there are many similarities between the 20 general spectrum of neuroendocrine malignancies in 21 adults and in children, and specifically maybe with 22 the two examples that were shown, small cell lung 23 cancer and neuroblastoma. But, there are also many 24 differences. I am not comfortable stating that the 25 similarities outweigh the differences so that I 198 1 think we should take a unified approach of lumping 2 these things together when it comes to interpreting 3 the regulations, with the exception that if a 4 product is coming forth, as has been alluded to, 5 which is specifically indicated for a biologic 6 target and that biologic target has been 7 demonstrated to be important and relevant both in 8 small cell lung cancer and also demonstrated 9 preclinical and biologically in neuroblastoma that 10 in that case there should be a link and you should 11 request pediatric studies but only in the context 12 of where there has been a predefined common 13 element, a targeted therapy that is biologically 14 relevant and suggests that it may be effective 15 would I consider that the two diseases be unified 16 in terms of the regulation. That would be my 17 answer to this question. I don't know if the rest 18 of the committee agrees so please speak up. 19 DR. PAZDUR: Would you advocate that that 20 target should be actually measured in a 21 subpopulation? 22 DR. SANTANA: Yes. 23 DR. PAZDUR: Say, if somebody was 24 developing a drug for lung cancer but they were 25 targeting and measuring a specific enzyme that they 199 1 were inhibiting, and it was only going to be used 2 on that specific population so it really targeted 3 -- 4 DR. SANTANA: Yes. 5 DR. PAZDUR: -- rather than a more general 6 -- you know, this may inhibit enzyme, cure, 7 whatever. 8 DR. SANTANA: Correct. I think it needs 9 to be defined very precisely and targeted very 10 precisely. Mike? 11 DR. LINK: Just more of a generic 12 question, and I am not sure it is directed to you 13 but do you really want to have separate principles 14 for sarcomas and separate on -- I am a little 15 nervous that you had kind of a nice proposal for 16 sarcomas -- at least I thought it was nice -- and 17 now you are kind of dancing around a little thing 18 here to try to accommodate a very different 19 approach. Maybe we should try to go back to the 20 sarcoma one and amend it a little bit to looking at 21 a pathway that might be targeted, which wasn't 22 included in the sarcoma thing, rather than trying 23 to make a totally different thing here for a 24 different class of tumors. 25 DR. SANTANA: Mike, I will give you my own 200 1 bias. My own bias is that in sarcomas the diseases 2 clinically, pathologically, etc., are very similar 3 and the differences are minor, whereas in this 4 example that we are being given now I think the 5 similarities and differences are very obvious -- 6 DR. LINK: I agree totally with what you 7 said -- 8 DR. SANTANA: -- so, I think I want to 9 propose -- 10 DR. LINK: -- the diseases are the same or 11 the pathway is the same, not that you have a 12 principle for sarcomas and a different principle 13 for neuroendocrine tumors and now we are going to 14 have brain tumors and we are going to have a third 15 different principle there -- you know, kids' brains 16 are fully developed or brains are not fully 17 developed. It seems more rational to have a 18 unifying principle which is either a targeted 19 pathway or that the tumors are identical on a 20 genetic -- 21 DR. PAZDUR: I think there could be 22 differences here. I think an uncomfortable feeling 23 that we are having here in dealing with small cell 24 lung cancer versus neuroblastoma is that even 25 though if somebody had a similar mechanism here 201 1 that was proposed, the clinical differences between 2 these two tumors are so different. So, for us to 3 go back and, don't forget, mandate a company to say 4 that this is the same indication would be very 5 difficult to do and we could be challenged on this. 6 DR. SANTANA: I think, Mike, the 7 principles are basically the same. It is just that 8 the diseases are different and they have to be 9 taken on a case by case basis. I think that is 10 what we are saying. In this particular case the 11 differences are so obvious that I would feel 12 comfortable saying the disease is technically the 13 same and, therefore, whenever anybody from industry 14 comes to the FDA saying I have a new drug or a new 15 product for small cell lung cancer that the agency 16 would mandate that they do studies in 17 neuroblastoma. To me that would be a step -- 18 DR. LINK: Too big a step. 19 DR. SANTANA: Too big a step. 20 DR. HIRSCHFELD: Unfortunately, our 21 knowledge is not the state of physics where I 22 think, much as we might like to have a unifying 23 principles, we couldn't come to that. So, that is 24 why we left open the possibility for nuances or 25 corollaries of some general schema, which is why we 202 1 asked the same question multiple times. 2 Now, to refine this a bit further, and it 3 might help looking at part B of this, should we 4 then think of, for instance, the refractory setting 5 and might that be different than the first-line 6 setting? 7 DR. SANTANA: I will get to that. I think 8 Anthony had a comment or a question. 9 DR. ELIAS: Not a major one. I think it 10 is just where the burden of proof lies. I think 11 the principles are the same and I agree with your 12 statement, Victor, but basically these two diseases 13 are so different that all you can really rest on is 14 if you have commonalities in particular pathways. 15 In the sarcoma situation you obviously have a lot 16 more similarities and the burden of proof is not 17 that you have to prove that these share the 18 commonality pathway; you can make that assumption 19 reasonably. 20 DR. SANTANA: Steve, I want to explore 21 your comment a little bit further. You are 22 suggesting that in the relapse setting the 23 principle should be different? Run that by me one 24 more time. 25 DR. HIRSCHFELD: I was just raising the 203 1 question that perhaps in the relapse setting we 2 might have a different perspective on it than in a 3 more global addressing of the two disease entities 4 or of these neuroendocrine tumors. 5 DR. SANTANA: Malcolm, think about that 6 one. 7 DR. SMITH: Yes, I thought that the 8 purpose of the exercise was not to describe how an 9 agent should be studied in children or population 10 that should be studied. So, I wouldn't see the 11 purpose of this committee to say you should study 12 it in a relapse setting but not in a newly 13 diagnosed setting but say it does or doesn't 14 warrant evaluation for neuroblastoma. 15 DR. HIRSCHFELD: Right, but that is if you 16 believe that all neuroblastomas are of the same 17 flavor. But if you postulate that the diseases 18 that lead to relapse are different than the ones 19 which don't, then you could I think logically 20 extend to saying, well, that would be something 21 else again and we happen to call it neuroblastoma 22 but maybe we should call it neuroblastoma variant, 23 or some other thing. I don't want to get into a 24 semantic argument; I just want to raise the 25 question. And, if the answer is, no, we should 204 1 continue to consolidate, then that is the 2 recommendation. 3 DR. SANTANA: I feel very uncomfortable 4 with that, Steve, and I can't give you a strong 5 argument. I am going to have to think through it, 6 but my gut feeling is that I feel very 7 uncomfortable with that train of thought. I think 8 Donna had a comment and I will get back to you in a 9 minute, Mike. 10 DR. PRZEPIORKA: Trying to get back to the 11 request to keep the unifying principles the same 12 throughout, I think that can be done because I 13 think what we had talked about in answering 14 questions A and B with the sarcomas in the design 15 of the clinical trial was would you put pediatric 16 and adult patients with such-and-such sarcoma in 17 one study, and our experts said, gee, we would 18 treat them the same way and they act the same way, 19 why not? So, in lumping sarcomas as a term, it 20 appeared that from a clinical perspective they were 21 truly the same disease. 22 I think in this instance we are talking 23 about a much larger pot. So, I would not conceive 24 of somebody coming to the agency and saying, well, 25 we have a drug for a neuroendocrine tumors and then 205 1 lumping pediatric and adult neuroendocrine tumors 2 together. I think this is a situation where the 3 neuroendocrine tumors in the pediatric population 4 clinically are different rather than just 5 pathologically and histologically and molecularly. 6 So, there may be some rationale to keep those 7 diseases on different protocols, but if there is a 8 molecular target in the adult situation which is 9 the same as in the pediatric population, that is 10 where the rule should be mandated to do additional 11 studies, not put them in the same protocol. 12 DR. SANTANA: Mike, do you have a comment? 13 DR. LINK: I guess I am confused now. If 14 you had a cytotoxic drug that had an 80 percent 15 response rate in non-small cell lung cancer would 16 you mandate that they do pediatric trials because 17 this is such a great drug? You wouldn't care? 18 DR. PAZDUR: That is not the question. 19 DR. LINK: I understand the question but I 20 am just saying in general principles, if a drug is 21 active -- 22 DR. PAZDUR: Of course, we would care. We 23 have to follow the law. Okay? And, the law is not 24 what we want it to be; it is what is written on the 25 books here and it clearly states that the 206 1 indication has to be the same. So, although we 2 would encourage sponsors to do it -- here, again, I 3 think this is a principle that I would like to get 4 across, remember, we are mandating companies to do 5 this so they can question us in a court of law 6 regarding our interpretation of this and, believe 7 me, if we stretch this it would lead to litigation 8 regarding this. I guarantee you. 9 DR. SANTANA: You would have to serve as 10 expert witnesses. 11 DR. PAZDUR: So, what we want and what we 12 think is academically interesting, for example, 13 yes, if a drug had activity in small cell lung 14 cancer I would like to see it studied in 15 neuroblastoma. I think it would be potentially an 16 interesting drug and perhaps an active drug, but 17 can we mandate that they do this? That is a 18 different situation and we have to live within the 19 confines of the law rather than what we think would 20 be academically interesting. 21 DR. HIRSCHFELD: And it has to be 22 something that is reviewed under that. So, even if 23 it is active in non-small cell, the company has to 24 request a marketing license for non-small cell in 25 addition. 207 1 DR. SANTANA: Susan? 2 DR. WEINER: I guess part of what makes me 3 so anxious about this conversation is that we 4 started with the elegant statements of the 5 accomplishments of the pediatric cooperative groups 6 and now, suddenly, it is a question of mandating 7 studies -- who is responsible for mandating studies 8 of drugs that companies are proposing for other 9 indications. I guess I just would like some 10 reassurance that the relationship between the 11 pediatric cooperative groups and the 12 decision-making would be pretty seamless about 13 this. 14 DR. SANTANA: I think both Malcolm and 15 Steve can speak about that. 16 DR. SMITH: I would just second Susan's 17 concern that I am not sure what the decision-making 18 process will be, but whatever it is, there needs to 19 be input from the research community about these 20 decisions. 21 DR. SANTANA: Dr. Kaye? 22 DR. KAYE: It is sort of a semantic issue 23 but another way of looking at the two principles 24 just has to deal with our confidence in the level 25 of evidence between the two. For instance, in the 208 1 sarcomas when you look at a rare, specific 2 translocation it is such compelling evidence 3 linking those diseases. On the other hand, every 4 drug that comes out now, it seems to me, is going 5 to have some mechanism of action because there is a 6 big push for that. How you get the same confidence 7 and the level of evidence that that is doing it, it 8 is often intuitive and for a lot of the agents that 9 are out there right now, that have been out there 10 previously for the past couple of years there is a 11 certain feeling, yes, it is probably not targeting 12 what we initially thought it was. So, it is more 13 likely, given the complexity of biology, that they 14 may not be quite right on the mechanisms of these 15 agents than being right. So, it is just something 16 that you have to keep in mind. I think that is 17 probably what is in the back of the mind -- you 18 feel confident with the translocation when they 19 come out with a tyrosine kinase inhibitor that says 20 this is specifically what it is doing. I think our 21 confidence this year is going to be not as great. 22 It just brings in again, you know, empirical 23 treatment. If I knew of a drug that was 80 24 percent, 85 percent effective in small cell lung 25 cancer I would certainly want to try it on any 209 1 disease, and that is sort of the empirical nature 2 and I think there is a bandwagon right now on 3 molecular targeting that is -- you know, I think 4 the push for that has always been present. Those 5 entities have always been present but there is a 6 bandwagon that I think may be blinding us. 7 DR. HIRSCHFELD: Victor, I just want to 8 say that the recommendations that would be useful 9 would be to say, yes, the rule should be invoked; 10 no, it should be waived; or we don't know yet and 11 let's continue to examine this. 12 DR. SANTANA: I would vote for the latter. 13 We don't know yet, and I think you have to take 14 each case individually for these particular 15 diseases. 16 DR. REYNOLDS: That is exactly what I was 17 saying. If you recall my last slide, I didn't put 18 on there I think that the Pediatric Rule should be 19 invoked; I said that studies should be strongly 20 considered. I think "by strongly considered" it 21 means that we should gather a little more data in 22 the process of doing this, and I think that is 23 consistent with what you are saying. It is 24 basically saying that if the targets are the same 25 and if you can get the clinical data suggested, 210 1 then perhaps the Pediatric Rule might need to be 2 invoked in this case. 3 DR. SANTANA: I think we have reached a 4 consensus on that one. Does the agency feel that 5 way? 6 DR. HIRSCHFELD: Right. I would like some 7 clarification down the list, if there are any 8 recommendations regarding waivers. 9 DR. SANTANA: Well, you know, I haven't 10 treated or seen a lot of mesothelioma but I think 11 they are probably the same disease. It is a 12 pediatric disease but it is the same disease as in 13 adults. That is what I was implying. I think the 14 pediatric mesothelioma, as rare as it is, is 15 probably the same disease as mesothelioma in 16 adults. I am trying to answer the questions. I 17 think probably the same is true with bronchiogenic 18 tumors. With the exception we have had about small 19 cell lung cancer, I think small cell lung cancer 20 and non-small cell lung cancer are not pediatric 21 disease and I don't want to go any further on that. 22 DR. PAZDUR: Let me just ask a technical 23 question because I was unaware of the mesotheliomas 24 and there are applications that we have looking at 25 drugs for this disease. Are there sufficient 211 1 numbers of patients to even invoke this rule? 2 DR. SANTANA: I mean, in the whole history 3 of St. Jude I think there have been ten patients. 4 So, it is very, very rare. It is very rare. 5 DR. PARHAM: Very rare, five cases. 6 DR. SANTANA: How about endocrine tumors? 7 We really didn't talk about those in the general 8 context, but I would propose that thyroid carcinoma 9 are probably the same diseases in adults as they 10 are in kids. Anybody disagree with that comment? 11 [No response] 12 Then adrenal tumors other than 13 neuroblastoma, Pat, do you want to comment on that? 14 DR. REYNOLDS: Well, I would suggest that 15 fibrochromocytoma is probably the same regardless 16 of its age. 17 DR. LINK: Except that that is a tumor 18 that occurs in people who are progenitively 19 predisposed. 20 DR. SANTANA: But when it gets manifested 21 it is variable, as you well know. So, the 22 pediatric disease is probably the same as in adults 23 in terms of the genetics. It is just a matter of 24 when it gets manifested. 25 Then, are there other pediatric 212 1 neuroendocrine tumors that have an adult 2 counterpart that is not commonly classified as an 3 adult neuroendocrine tumor but as some other type 4 of adult malignancy such as a carcinoma? It is the 5 same question as this morning which I had 6 difficulty with. Anybody want to comment on that 7 one? I can't think of any. David, any thoughts on 8 that? 9 DR. PARHAM: I can't think of anything. 10 DR. SANTANA: Okay. Have we satisfied 11 those questions for the agency? Let's go ahead and 12 talk for the rest of the afternoon about the CNS 13 malignancies. So, I invite Susan to come to the 14 podium, and Dr. Burger is going to join us on the 15 telephone. So, give us a second to get the 16 telephone connection. 17 DR. BURGER: Hello. 18 DR. SANTANA: Dr. Burger, can you hear us? 19 DR. BURGER: Yes, I can. 20 DR. SANTANA: Welcome. For the purpose of 21 the record, please state your name and your 22 affiliation. 23 DR. BURGER: Yes, this is Peter C. Burger. 24 I am from Johns Hopkins University, Department of 25 Pathology. 213 1 DR. SANTANA: Thanks, Peter. We are going 2 to have two short presentations, one by Susan and 3 one by Howard, and we are just going to go ahead 4 and do the presentations and then we will open up 5 for discussion. Okay? 6 DR. BURGER: Fine. 7 DR. SANTANA: Susan? 8 Perspectives on CNS Malignancies 9 DR. STAUGAITIS: Thank you. 10 [Slide] 11 I am going to give some of my perspectives 12 on CNS malignancy, and I will be reiterating many 13 of the points that were brought up already today 14 and I will emphasize some of the unique opinions 15 that I may have compared to the rest of the group. 16 [Slide] 17 The background that I come from is as a 18 neurobiologist with an interest in development and 19 also as a neuropathologist. I do not have the 20 breadth of experience as my colleagues, like Dr. 21 Burger, in terms of how much I have seen in CNS 22 malignancies, neither am I an oncologist, and I 23 have been encouraged to speculate to provoke 24 discussion and so as a disclaimer in the beginning, 25 I want to say that I am going to throw out a lot of 214 1 crazy ideas. These are not recommendations; they 2 are for my clinical colleagues to respond to and 3 determine whether or not they have any weight. 4 I am going to talk about CNS neoplasms by 5 reshuffling the deck in different ways. First, I 6 will go through the classical dogma of the general 7 classification of tumors as defined by histology, 8 then I will describe them in other ways, group them 9 in other ways as defined by physiology, for 10 example. 11 [Slide] 12 Just for some background, the diagnosis of 13 brain tumors is very different now than it was many 14 years ago. Imaging has enabled us to identify 15 smaller lesions, subclinical lesions. Biopsies are 16 smaller. And, if we are talking about whether 17 different malignancies are the same, a 18 neuropathologist often wonders whether the tumor is 19 the same when they are two centimeters apart from 20 each other in the same patient. 21 One of our roles is in terms of specimen 22 adequacy, and one of the issues that was brought up 23 earlier in terms of can we do all of the genetic 24 studies that we would like to do on the tissue that 25 we are provided, and sometimes that is just not 215 1 possible, although we would like to be able to 2 obtain as much tissue as we can. 3 Classically, the neuropathologist looks at 4 tumors from the point of view of histologic 5 phenotype and also grade and, as we have mentioned 6 throughout the day, we have additional information 7 in terms of gene expression. Immunocytochemistry 8 is now a standard of care in pathology in general, 9 and genomic alterations and molecular diagnosis is 10 on its way there. 11 [Slide] 12 One of the things that the pathologist 13 contributes with these molecular studies is that it 14 is up to us to tell the molecular biologist where 15 the tumor is and what to sample. I don't want 16 anybody to really lose sight of that aspect of our 17 responsibility. 18 The morphologic classification of CNS 19 neoplasms is based upon a resemblance of neoplastic 20 cells to normal cells. Throughout the ages people 21 have used this to infer a cell of origin. I am 22 very hesitant to say that. I will basically be 23 talking about the phenotypes of different cells, 24 not necessarily the specific cell that neoplasm 25 might be derived from because I think that we 216 1 probably don't know all of that information. 2 And, the cell of origin is important 3 because this becomes the basis of in vitro 4 experimental models on which initial compounds are 5 tested. So, for example, do mature human adult 6 astrocytes in culture represent a model for all 7 kinds of astrocytomas? I am not completely sure. 8 There could be progenitors, other kinds of 9 precursor cells that may reflect the physiology of 10 the cell that becomes transformed. 11 [Slide] 12 In terms of just outlining the different 13 tumors, I am going to describe them in terms of 14 their sites of origin, CNS parenchymal accessory 15 structures and the CNS coverings. The largest 16 group are the CNS parenchymal neoplasms and, as I 17 alluded to earlier, I am dividing this into cells 18 with a glial phenotype, a neuronal phenotype and an 19 embryonal phenotype. 20 Among the glial phenotype astrocytomas, 21 oligodendrogliomas, the neoplasms look like the 22 normal cells in many of the instances but it does 23 not necessarily imply a cell of origin. 24 Astrocytomas tend to have a high 25 propensity to progress to higher grade lesions, 217 1 whereas with some of the other neoplasms -- 2 oligodendrogliomas -- we can have a higher grade 3 progression to that although it is less likely. In 4 ependymoma cytologic malignancy often is not 5 correlated with the clinical behavior on the 6 patient. So, even within this classification there 7 are many differences. 8 [Slide] 9 The neoplasms with the neuronal phenotypes 10 tend to be more within the pediatric population. 11 They tend to be more low grade, and the most common 12 of these are the ganglioma/gangliocytoma family. 13 The other neoplasms with names like neurocytoma, 14 dysembryoplastic neuroepithelial tumor lead us to 15 say that we really don't know what we are talking 16 about with these lesions. They express certain 17 antigenic phenotypes that make us infer that they 18 might have properties of neurons or neuron-like 19 cells or progenitor-like cells, but there is still 20 a lot to be learned about these. Fortunately, many 21 of these are very benign lesions and often not an 22 issue for drug development. 23 [Slide] 24 The third category are the embryonal 25 neoplasms, such as medulloblastoma, the 218 1 supratentorial PNET tumors and the atypical 2 teratoid/rhabdoid tumor. 3 [Slide] 4 The accessory CNS structures include the 5 lesions of choroid plexus, the pineal gland and 6 pituitary. 7 [Slide] 8 The lesions arising in the coverings 9 include the meningeal tumors such as meningiomas, 10 hemangiopericytoma, other sarcomas and melanocytic 11 neoplasms, as well as the peripheral nerve sheath 12 tumors. 13 [Slide] 14 Now I would like to rearrange these in 15 terms of who gets what. For the most part, 16 virtually every age patient can get these different 17 CNS tumors but some are much more commonly found in 18 adults; some more commonly found in pediatrics; and 19 some are almost exclusively pediatric. 20 [Slide] 21 For example, most gliomas are found to a 22 much greater extent in adults. Histologically, to 23 my knowledge, the fibrillary gliomas in adults and 24 the pediatric population histologically are 25 essentially the same. So, perhaps they could be 219 1 treated as the same. 2 Similarly, for the other neoplasms that I 3 list here, the pineal parenchymal neoplasms, the 4 embryonal pineal blastoma are more common in 5 younger people but histologically the tumors are 6 the same. Similar, for the tumors of the 7 coverings. 8 [Slide] 9 In terms of pediatric being much greater 10 than adult, we have the unusual low grade 11 astrocytoma, such as pilocytic astrocytoma and 12 pleomorphic xanthoastrocytoma, the intraventricular 13 ependymoma, the glial and glial neuronal neoplasms 14 and the embryonal neoplasms, such as 15 medulloblastoma and, as you can see on the slide, 16 choroid plexus, germ cell and craniopharyngioma. 17 These are the ones where I think we really have to 18 try and find criteria for including this with other 19 neoplasms because it is unlikely that drugs would 20 be developed specifically for these, given that 21 there are small populations of people who are 22 actually affected. 23 [Slide] 24 Finally, there are a few neoplasms that 25 are virtually unheard of in adults, such as the 220 1 desmoplastic infantile astrocytoma or ganglioma, 2 atypical teratoid/rhabdoid and supratentorial PNET. 3 [Slide] 4 We mentioned a lot about the effect of 5 mutations and alterations, and I want to take a 6 moment to think about what the genetic alterations 7 that we can detect mean in terms of the biology of 8 the tumor. For example, a mutation or 9 rearrangement affects a specific gene in a specific 10 way and we can see how it is reflected in gene 11 expression. Whereas, a gain or a loss of genetic 12 material can involve huge areas of the chromosome 13 and it may be difficult to predict the behavior or 14 the responsiveness of a therapy based on loss of 15 chromosome 1P because, for example, loss of 16 chromosome 1P in an oligodendroglioma may have a 17 different effect on a tumor than a loss of 18 chromosome 1P in a neuroblastoma, and so forth. 19 [Slide] 20 In thinking about the cell of origin of 21 the neoplasm is does the physiology of the 22 precursor cell that is transform affect the 23 behavior of the neoplasm, and does that affect the 24 way that drugs interact with it? For example, once 25 a precursor cell is transformed by genetic 221 1 alteration, do its normal physiologic processes 2 matter or don't they? Is it important to think 3 about the cell of origin at all? 4 I think with higher grade tumors that 5 acquire more and more mutations, that becomes less 6 important. The low grade, these elusive tumors 7 where we don't have specific molecular markers for 8 early intervention, those tumors may actually have 9 more of a relationship to the precursor cell. 10 [Slide] 11 Another thing that I would like to 12 consider in my talk is the relationship of familial 13 syndromes that are associated with CNS neoplasms. 14 Many of the neoplasms, such as the astrocytomas and 15 the meningiomas that one sees in the pediatric 16 populations are superimposed on a genetic syndrome. 17 As you can see from the different syndromes that 18 are listed here, some tumors are increased in 19 incidence on very different genetic backgrounds. 20 For example, astrocytomas have been associated with 21 neurofibromatosis Type 1, neurofibromatosis Type 2 22 with the Li-Fraumeni syndrome in TP53 alterations, 23 with APC mutations. Are all of these tumors the 24 same? Histologically they look identical but 25 because potentially different pathways are involved 222 1 and this is the substrate upon which these tumors 2 are superimposed, can we really make predictions as 3 to whether the indications are the same? 4 [Slide] 5 Let me reshuffle the deck again a little 6 bit more. We talked about histopathology. What 7 about the growth properties of transformed cells? 8 Can we lump histologically disparate tumors 9 together based upon, say, proliferation, survival, 10 migration, motility and angiogenesis? I would just 11 like to throw out a few examples here for 12 discussion. 13 For example, some of the rare, highly 14 malignant tumors that are very common in the 15 pediatric populations such as medulloblastoma, the 16 other PNETs and high grade gliomas, choroid plexus 17 carcinomas are rapidly dividing tumors and the 18 strategy in oncology for years has been just to 19 target the rapidly proliferating cells. If we can 20 identify specific molecular targets that interfere 21 with a particular aspect of the cell cycle, that 22 could be effective and less toxic and that is 23 advantageous. But this is sort of an approach 24 where we are lumping together tumors based upon 25 their growth properties, and I think it also ties 223 1 in with the comments that were made earlier about 2 grade. 3 [Slide] 4 Another way that we might be able to link 5 neoplasms is in terms of their ability to 6 infiltrate into the central nervous system. One of 7 the aspects of CNS malignancies that make them 8 really refractory to treatment is the ability of 9 single cells to migrate long distances, and if 10 there was an agent that could interfere with the 11 motility of one type of transformed glial cell, 12 might it also be able to interfere with the 13 motility of another type of transformed glial cell? 14 Similarly, if one were developing 15 mechanisms by which therapies can home to tumor 16 cells that infiltrate widely, perhaps that can be 17 applied to many classes of neoplasms. 18 [Slide] 19 Another example would be angiogenesis 20 inhibitors. For example, both high grade 21 astrocytomas, such as glioblastoma multiforme and 22 low grade pilocytic astrocytomas, show 23 histologically similar vascular proliferation 24 patterns. Do the same mechanisms promote this 25 proliferation and, if so, can drugs designed to 224 1 target the vasculature in high grade astrocytomas 2 be effective in unresectable pilocytic 3 astrocytomas? A pilocytic astrocytoma resected 4 from the cerebellum is essentially cured but there 5 are many, many patients who have very deep lesions 6 around the hypothalamus that can not be adequately 7 resected and the vascular proliferation that is 8 associated with these neoplasms may be a target for 9 therapy and extending the rule. 10 [Slide] 11 We have mentioned p53 mutations a number 12 of times and I will just reiterate some of the same 13 points. Many, many of the neoplasms in the CNS 14 have mutations in p53. One thought is to find 15 agents that will stimulate the function of p53. On 16 the other hand, there are also agents being tested 17 that will inhibit the function of p53 in normal 18 cells so that normal tissues can be protected 19 against the genotoxic stress of therapies. This 20 may be particularly important to test in the 21 pediatric population where we are very concerned 22 about the developing nervous system and the effect 23 that different radiotherapies and chemotherapies 24 can have. So, I think we have to keep our minds 25 open and also think about agents that protect the 225 1 normal tissues. 2 [Slide] 3 We have mentioned the PDGF receptors many 4 times already today. There is evidence that 5 PDGF-alpha receptors are overexpressed in a number 6 of gliomas, including fibrillary astrocytoma, 7 oligodendroglioma, ependymoma and pilocytic 8 astrocytoma. If it can be shown that the 9 expression of this receptor and the activity of 10 this receptor and pathway is critical to the 11 neoplastic phenotype, I would agree with what we 12 have already said before, that it could be an 13 indication to become more inclusive of the types of 14 neoplasms that are indicated for these agents. 15 [Slide] 16 On the other hand, let's think about the 17 epidermal growth factor receptor where, in adults, 18 de novo glioblastomas tend to be amplified; 19 secondary glioblastomas do not. Are they different 20 tumors? And, how do you define an indication for 21 something that has activity on the epidermal growth 22 factor receptor or its downstream pathway, and what 23 neoplasms should you extend these drugs to or limit 24 them to? 25 [Slide] 226 1 Finally, I think that others today have 2 emphasized that it is important to look at the 3 entire pathway. When I first started to read about 4 the genetics of neoplasms I was always a little bit 5 discouraged when I would learn that, well, 20 6 percent of these tumors have this alteration and 5 7 percent of these tumors have another alteration, 8 but as we learn more about the intracellular 9 signaling mechanisms and how pathways can come 10 together, and we put together the alterations 11 within pathways we will get up to numbers like 60 12 percent and 70 percent and 80 percent of neoplasms 13 involve a particular pathway. Then, the rational 14 biologic approach would be to find the bottleneck 15 in that pathway and see if there are ways to 16 inhibit or activate that. 17 [Slide] 18 Finally, I will just tone myself down a 19 little bit and express a few cautions that I 20 considered that while I was putting together my 21 thoughts on this presentation. The central nervous 22 system is very different than the other parts of 23 the body in that it is encased in our hard skills, 24 and the necrosis and swelling that are associated 25 with rapid and efficient cell killing may have 227 1 truly adverse effects within the confines of the 2 central nervous system. 3 Environmental signals that may affect the 4 behavior of neoplastic cells may change during the 5 development. Specific targeted therapies will work 6 only if the inhibited pathway is intact in the 7 particular tumor being treated. 8 I just read a paper in Science regarding 9 the treatment of CML with STI571, and apparently 10 there is a population of populations who, after 11 responding to the therapy, become refractory and it 12 was identified that these patients have acquired a 13 mutation that makes the cells resistant to this 14 particular gene. They further proved that the 15 activity was still important in the malignant 16 behavior of this particular neoplasm. So, I think 17 in all of our discussions we have to remember that 18 neoplasms are constantly changing, constantly 19 evolving processes that may always be one step 20 ahead of us. 21 Then, finally, therapies that target 22 specific functions, such as proliferation, 23 migration, may actually adversely affect the normal 24 developing cells within the nervous system and that 25 changes rapidly, especially in early childhood, and 228 1 may actually be reasons to invoke the waiver in 2 this. With that, I would like to thank you. 3 DR. SANTANA: I would like to invite 4 Howard to come to the podium. 5 Perspectives on CNS Malignancies: Clinical Aspects 6 DR. FINE: I want to thank the organizers 7 who asked me to speak here. After Henry did his 8 usual nice job and Susan spoke about the science, 9 which is always one of my favorite topics, the 10 question is what can I say here? Probably not 11 much. 12 [Slide] 13 But what Steve suggested I talk to the 14 group about -- obviously, there are some world 15 renowned oncologists around the table but many of 16 you are not so involved in neuro-oncology and brain 17 tumors. So, he thought it would be useful for me 18 to just go over some of the basic clinical aspects 19 as far as how these patients do, the natural 20 history of their disease clinically speaking, how 21 we approach them, how we treat them and some 22 general outcomes that we expect from these tumors. 23 So, I thought I would do that. So, I don't think I 24 need this as an introduction. Suffice it to say 25 that these are an important group of tumors both in 229 1 the adult and the pediatric population, and 2 increasingly more an important group of tumors than 3 I think was ever appreciated. Certainly, I can 4 tell you that at the National Cancer Institute, on 5 a national level, this group of tumors is 6 increasingly being recognized as a very important 7 target for the next decade. 8 Along with the problem of these tumors 9 causing a significant amount of cancer mortality is 10 the morbidity that both adults, and in particular 11 the children, suffer st these tumors, not just from 12 the tumors themselves but from the treatments that 13 we use to treat them. I think whenever we talk 14 about brain tumors in either the pediatric or the 15 adult population, we have to think about toxicity 16 in a very different way than we do for systemic 17 tumors because the toxicity is almost permanent and 18 it is always a balancing act in trying to decide 19 whether a few months of increased life is really 20 worth significantly decreased quality of life. 21 [Slide] 22 I think when we talk about the pediatric 23 role, at least when I think about it, I think of a 24 couple of questions. Number one, are the tumor 25 types the same? And,; is a specific tumor type the 230 1 same in a child compared to an adult? I think 2 there are several ways that we can answer that, and 3 we have already addressed those ways in the other 4 tumor types. 5 There are obviously the biologic criteria, 6 and Susan and Henry have both kind of addressed 7 that, both as far as standard pathology is 8 concerned, as well as molecular diagnostics. But 9 the other way to address that is the clinically 10 behavior of the tumor, both as far as the natural 11 history of the tumor and how the tumor responds to 12 therapy. As I said, that is what I will try to 13 address over the next five or ten minutes here. 14 [Slide] 15 Again, we have seen this slide before, or 16 variations of this slide, relative to the first 17 question I asked, are the tumors the same? Well, 18 the tumors are the same except their distribution 19 is highly different between adults and children, 20 with actually by far the most common adult brain 21 tumor being metastatic tumor, something we actually 22 forget about sometimes, with high grade gliomas 23 being by far the most common problem after that. 24 With pediatric tumors we are really dealing with 25 embryonal tumors and then low grade gliomas as 231 1 opposed to the high grade gliomas. 2 I am sure you don't want to hear me go 3 through the natural history and treatments of all 4 the 75 different subtypes, or whatever the most 5 recent WHO categorization tells us the subtypes of 6 CNS tumors are, I thought probably the most 7 important -- and I asked Steve who agreed -- the 8 most important tumor to go over is gliomas. The 9 reason I say that is that although gliomas are not 10 the most common pediatric brain tumor, the fact of 11 the matter is, and we can and should open this up 12 for discussion after this talk but most of the 13 other brain tumors that we see in children are 14 hardly represented at all in adults. So, for this 15 discussion of the Pediatric Rule, it is unlikely 16 that a drug company is going to design a drug for 17 cranial pharyngiomas in adults where we are going 18 to have to worry applying the Pediatric Rule. 19 So, to keep this on a practical side, and 20 we can change that if you want but to keep it on a 21 practical side, the reality is if drug companies 22 are going to develop a drug at all for tumors, and 23 that is another issue but the few times they do, it 24 is going to be for gliomas because that is the 25 disease in adults and that is where I think we need 232 1 to address the issue of the Pediatric Rule, at 2 least in my personal opinion. 3 [Slide] 4 So, the first thing -- and you can quote 5 me on this; the reference is down below. It is my 6 anticipation this will be a truism that goes on for 7 years. 8 DR. SANTANA: It won't be dinosaurs 9 anymore or rainbows; it will be something else! 10 DR. FINE: But I think this is important. 11 A glioma is not a glioma; it is a heterogeneous 12 group of diseases and, as a matter of fact, it is a 13 heterogeneous disease even within a patient. So, 14 you know, Henry showed some data and Susan showed 15 some data that say that some of the molecular 16 alterations in the pediatric high grade gliomas do 17 not exactly correlate with those of the adult 18 patients and it is important to understand that 19 within the adult patients the genetic alterations 20 are hugely variable. Whether that reflects the 21 fact that they are many, many different 22 subcategories at a genetic expression profile level 23 of gliomas, whether that reflects the fact that 24 these tumors, as opposed to leukemias for instance 25 or even pediatric sarcomas, genetically messed up 233 1 tumors -- these tumors are highly aneuploid and 2 what genetic alterations are really important for 3 the pathogenesis of these diseases is not yet 4 clear. So, I think we have to be very careful 5 about over-reading the genetics that we find in 6 these tumors for now until we really understand who 7 the important players are. That, again, gets back 8 to what I keep talking about today, validation of 9 molecular targets. 10 [Slide] 11 So, let's first talk about the two major 12 categories using standard pathology criteria of 13 gliomas, those being low grade gliomas -- generally 14 if we talk about a four-tier scale like the WHO, 15 grade 1 and 2 gliomas, and high grade gliomas, 16 grades 3 and 4, variously known as anaplastic 17 astrocytomas and glioblastomas. 18 To contrast the natural history of low 19 grade gliomas and, please, with Roger and Henry and 20 Larry, world renowned pediatric neuro-oncologists 21 here, feel free to correct anything you see on the 22 slide but generally speaking, the natural history 23 in adults -- generally these tumors are limited to 24 astrocytic or oligodendroglioma histologic subtypes 25 or mixed histologic subtypes. While in children we 234 1 get multiple subtypes, and we have already heard 2 about that from the pilocytic astrocytoma to the 3 ependymal tumors to mixed neural glial types of 4 subtypes. So, that is one way that they are 5 different. 6 Certainly, in adult these are slowly 7 progressive and infiltrative tumors and that is 8 generally true for low grade tumors in children but 9 not always. Some of these tumors appear to be 10 self-contained. Certainly the pilocytic tumors 11 are, and they can be cured if they can be safely 12 surgically resected, something we really don't find 13 on the adult side. So, I think that is a key 14 difference. 15 Another very important biologic difference 16 is that most patients or almost all adults with low 17 grade tumors die of their tumors. These are not 18 benign tumors, and the way the majority of patients 19 die of low grade tumors is that they transform to 20 high grade tumors, at least about 60-80 percent of 21 them. That number, although it is very difficult 22 to come by, in the pediatric population is much 23 smaller. So, that reflects an important biologic 24 difference, at least in my mind, between these two 25 different subtypes. 235 1 Again, I think this is also reflected in 2 the survival. Again, why I never like to use and 3 would never use the word "benign" tumor for a low 4 grade glioma in an adult is that the ten-year 5 survival rate is well less than 30 percent, and 6 since most adults who get low grade gliomas tend to 7 be younger adults, that is not a benign disease. 8 Also, it should be noted that there appears to be 9 no survival difference depending on anatomic 10 location of the tumor. 11 These numbers and these facts contrast 12 with what we generally see in pediatric low grade 13 gliomas where the ten-year survival is probably 14 well over 50 or 60 percent, and that survival, as 15 Roger went over with me very clearly last night, is 16 very much dependent on location of the tumor. 17 Whether that reflects the surgical resectability of 18 the tumor or whether that reflects something about 19 the natural history and biology of the tumor I 20 think remains unclear at this point. 21 [Slide] 22 As far as how do we approach adults and 23 children with low grade gliomas, well, I think for 24 both these tumors if they can be surgically 25 resected, it is considered optimal. Certainly, 236 1 more so in adults. When we can't resect them 2 fully, or even if we can, usually that is not 3 enough and, as a matter of fact, it is almost never 4 enough with the exception of maybe truly low grade 5 oligodendrogliomas. Therefore, radiation therapy 6 is commonly used. There still is a big question 7 about the timing of radiation therapy -- radiate me 8 now or radiate me later, meaning at the time of 9 tumor progression. That remains an unknown issue. 10 Although long-term toxicity of radiation 11 to adults remains a problem that we talk about, it 12 isn't one of the major, major issues as it is, as 13 we will talk about, in children. There is a 14 question, increasingly so, of the use of focal 15 radiotherapy for low grade gliomas. Chemotherapy 16 has no proven benefit in the treatment of low grade 17 gliomas. There is increasing evidence to suggest 18 that maybe low grade oligodendrogliomas, 19 particularly with the 1P, 19Q marker, may have 20 sensitivity to alkylating agents, and maybe even 21 mixed gliomas may have some activity though, again, 22 I think that remains to be seen as far as how 23 common that is. 24 As far as children are concerned, again, 25 if we can fully resect most of these tumors, 237 1 certainly tumors like pilocytics, that is 2 considered optimal treatment. We are very 3 hesitant, because of the toxicity associated with 4 radiation, to use radiation and it is often, as 5 opposed to second-line therapy, a last choice. One 6 of the reasons it is our last choice is because, 7 indeed, chemotherapy can be quite effective in 8 these tumors, as opposed to adults, with 9 carboplatinum or platinum-based regimens, having 10 the potential to give quite high response rates and 11 control these tumors for a number of years. 12 So, I think there are significant 13 differences in the natural history of low grade 14 gliomas in adults and children. Whether that 15 should affect the Pediatric Rule is something that 16 I am going to throw open to the committee. 17 [Slide] 18 Let's talk about high grade tumors. Most 19 commonly in adults they are supratentorial as 20 opposed to in children where we are dealing with 21 basically almost an equal split of infratentorial 22 versus supratentorial. Both these tumors, however, 23 whether they be in adults or children, are bad 24 tumors. They are infiltrative. They are rapidly 25 progressive. They are destructive. They have high 238 1 degrees of angiogenesis. They disrupt the 2 blood-brain barrier and the prognosis is poor. 3 The prognostic variables that we know for 4 high grade gliomas over the years, shown by 5 multiple studies, many done by Victor Levin who is 6 here today, include very powerful predictors such 7 as age, grade, performance status of patients and 8 the postoperative radiographic residual tumor. 9 That is not to say the extent of resection. The 10 only thing that has been shown is that when you 11 measure radiographically the amount of tumor left 12 after surgery, that is a predictor of survival. 13 Surgeons like to translate this to say, oh, that 14 means we should take more out and whether that is 15 true or not is not necessarily the case. 16 Prognosis for children with high grade 17 gliomas also is clearly grave, meaning an 18 anaplastic astrocytoma versus a glioblastoma is a 19 very clear predictor. It appears that 20 postoperative radiographic tumor extent is also a 21 prognostic variable. Performance status is harder 22 to judge in children, as you all know well, and age 23 as far as small children versus teenagers is 24 something that I think is also less clear. 25 [Slide] 239 1 When we talk about treatment of high grade 2 gliomas, surgery is uniformly, I think it is fair 3 to say, considered important at least as far as 4 surgery for determining a diagnosis. I think as of 5 the year 2001, we want a histologic diagnosis on 6 almost everyone. Probably two exceptions to this 7 are patients with infiltrating brain stem lesions 8 where radiographically it can almost be nothing 9 else, and morbidity of biopsy of this area makes 10 the risk versus benefit ratio against doing the 11 surgery. Then, there is a cohort of patients who 12 have prototypic radiographic criteria of 13 glioblastoma who are basically morbid from their 14 tumors, for whom we know the treatment isn't going 15 to do anything for them and some of those patients' 16 families elect not to have biopsies. 17 Generally speaking, although it remains 18 controversial, for most of the major brain tumors 19 it is generally thought, when possible, maximal 20 debulking surgery is advantageous for high grade 21 gliomas, mainly for the purposes of diminishing the 22 mass effect from these large tumors, for the 23 purposes of decreasing steroid requirement over the 24 next several months. It also decreases the 25 potential sampling bias because, as we have talked, 240 1 these are highly heterogeneous tumors from one area 2 to another. Although, again, the trial hasn't been 3 and will never be done, that being a randomized 4 trial of biopsy versus surgery, I think most people 5 believe that surgery probably extends survival at 6 least to some extent, though probably not hugely. 7 Larry Kun is here who has irradiated more 8 children with brain tumors probably than anyone 9 else in the world. I would like to hear his 10 comments but, generally speaking, radiation is 11 still the gold standard for high grade gliomas in 12 both adults and children. 13 Involved field radiation therapy is now 14 standard as opposed to whole brain radiation, 15 thereby potentially decreasing or definitely 16 decreasing the neurocognitive toxicities of 17 radiation. Generally we are talking about 18 something in the range of 5940 or 6000 centigrade 19 spread out over 30-33 fractions. Different dose 20 and fractionation schemes have been looked at 21 continuously through the RTOG and other 22 organizations. They continue to be looked at but 23 to this point there has been no dose or 24 fractionation scheme that has clearly been shown to 25 be superior over the standard regimen that I just 241 1 spoke of before. There is a question of the use of 2 high dose focal radiation techniques, like 3 radiosurgery, a gamma knife and so forth though its 4 role remains to be defined. 5 Then, again, toxicity as far as the acute 6 toxicity of radiation, meaning over the first few 7 months, is generally one related to radiation 8 necrosis. The real toxicity we are concerned 9 about, particularly in children, of course, are the 10 long-term, well-documented neurocognitive 11 dysfunctions that appear to be dose and extent of 12 CNS related, as well as the age at which the 13 patient was radiated at. 14 [Slide] 15 How about chemotherapy? Well, I think of 16 chemotherapy in two roles, first as part of the 17 initial treatment or adjuvant treatment -- I don't 18 really like to use the term "adjuvant" because at 19 least on the adult side when we think of adjuvant 20 we think of breast cancer when the tumor has been 21 fully removed. These tumors are never fully 22 removed but at least as far as up-front treatment, 23 what is the role of chemotherapy? It is 24 controversial. There have been multiple randomized 25 trials. The results are mixed. The reasons that 242 1 the results are mixed, in my opinion, is that most 2 of these trials consist of patients that are hugely 3 heterogeneous in their prognostic factors as well 4 as their tumor types, and most of the trials have 5 been underpowered to detect subgroup analysis 6 difference. 7 We have performed a meta-analysis. There 8 has now been another meta-analysis that has looked 9 at the use of adjuvant chemotherapy. We and the 10 other group have shown that there appears to be a 11 survival advantage for the use of chemotherapy in 12 adults in patients with anaplastic astrocytomas, 13 with the best regimen appearing to be a regimen 14 developed by Victor, PCV, though there is some new 15 retrospective data from RTOG and UCSF that suggests 16 that single agent nitrosourea may be as good as PCV 17 in adjuvant treatment, and now there is the new 18 drug, just approved by the FDA about a year ago, 19 tenozolamide. Its role as up-front treatment is 20 being explored at a number of centers. 21 The question of the role of chemotherapy 22 for the more common glioblastoma remains 23 controversial. Our meta-analysis suggested that 24 there was a very minimal benefit. The benefit that 25 did exist appeared to have benefit in the patients 243 1 with the best prognostic factors, which is only 2 about 10-20 percent of all patients. So, the 3 majority of patients did not appear to benefit. 4 Whether patients get chemotherapy up front or not 5 remains a controversial area and is very physician 6 dependent, I think it is fair to say, in this 7 country. 8 Children with glioblastoma appear to have 9 somewhat of a survival advantage when they use 10 chemotherapy, though it is less clear that children 11 with anaplastic gliomas benefit all that much when 12 up-front chemotherapy is given. 13 [Slide] 14 When we look at chemotherapy for recurrent 15 gliomas, there have been few agents with documented 16 objective responses. Tenozolmide, as I mentioned 17 before, is the most recent of those and, outside of 18 that, the FDA, not counting Gliadel, I don't think 19 has approved a drug for glioma in 30 years, since 20 BCNU, and I think there is a reason for that and it 21 is not a political reason; it is a biology reason. 22 There are a few agents with proven 23 improvements in quality of life, and there are few 24 agents, maybe zero, with documented improved 25 survival with, again, the exception possibly of the 244 1 Gliadel wafer and that benefit, if it exists, is 2 marginal. 3 [Slide] 4 Basically, the treatment outcome for low 5 grade gliomas in adults is quite poor. In children 6 it can be good with the exceptions of the subtypes 7 we talked about. For adults the treatment of high 8 grade gliomas is horrible and it is absolutely no 9 better in children. 10 [Slide] 11 So, points to consider for discussion -- I 12 think a couple of things. Number one, clinical 13 differences in natural history of high grade 14 gliomas between adults and children appear to be 15 trivial, in my opinion. Potentially promising 16 agents for which there are drugs now being tested 17 in the adults include drugs that are targeting the 18 EGFR, PDGF pathways, PI3 kinase, the AKT, 19 angiogenic targets such as VEGF or its tyrosine 20 kinase high affinity receptor, FLK, and certainly 21 the P16/RB E2F pathway all are promising targets 22 that are being looked at in adults, and I see no 23 reason why children with high grade gliomas 24 shouldn't be given the opportunity to explore these 25 promising new drugs. 245 1 I do have to say the caveat, which I put 2 on the bottom of this slide, which I mentioned 3 earlier today. I think it is worth considering 4 what do we do if drug X that targets, for instance, 5 the variable deleted EGFR which is so common in 6 adult gliomas but is not found in pediatric gliomas 7 is being developed for adult gliomas? Do we invoke 8 the Pediatric Rule there? So, again, this drug is 9 being developed for high grade gliomas but there is 10 a specific target that we don't actually find on 11 the high grade gliomas in children. What do we do 12 with that drug? 13 [Slide] 14 As final points to consider, low grade 15 gliomas in children do appear to constitute a 16 heterogeneous group of diseases, many of which 17 appear to be different than adult low grade gliomas 18 both in their natural history and in the response 19 to therapy. So, what do we do here? Should they 20 be treated the same? As I also mentioned, should a 21 drug with modest benefit in survival, if one is 22 identified for adults for instance, but with 23 significant long-term neurotoxicity be considered 24 similarly in the pediatric population, given the 25 fact that we expect the child to more likely live a 246 1 lot longer than the adult? I think that is 2 something to consider as far as the Pediatric Rule. 3 Then, finally, the one thing I haven't 4 talked about and a major issue as far as 5 neuro-oncology in the population are brain stem 6 gliomas. These tumors appear to have unique 7 radiographic and clinical correlates. Although 8 pathologically these tumors appear to be similar to 9 supratentorial gliomas, they do appear to behave 10 differently. Should they be treated differently? 11 I actually don't have a firm answer about that and 12 I think that is worth some discussion. So, thanks. 13 DR. SANTANA: Thank you, Howard. Dr. 14 Burger, are you still on the phone? I guess not. 15 DR. FINE: That is usually what happens 16 when I talk. 17 [Laughter] 18 Discussion 19 DR. SANTANA: I just wanted to see if he 20 was still connected to see if he had any comments 21 on the two presentations. I want to get back to 22 one of the last issues that Howard challenged us to 23 try to answer to start the discussion because it 24 came up earlier this morning too. And I would like 25 to hear some feedback from various members of the 247 1 committee. That is, if a sponsor is coming forth 2 with the example you gave, drug or biologic X that 3 targets a specific receptor, for example, the case 4 he gave, but in pediatrics we have the same 5 histologic disease but the receptor is not 6 expressed, would the rule be invoked in that 7 scenario? I would like to follow up on that as a 8 point of discussion. Anybody want to comment on 9 it? Victor? 10 DR. LEVIN: I think it is a non-issue. 11 The real question is, is it a target in either case 12 and there are other EGF receptor kinase inhibitors; 13 there are antibodies. There are all sorts of 14 different approaches that one can validate that 15 that is a logical target for a lower grade 16 astrocytic tumor. So, I was perplexed by the 17 question because, to me, it was not an issue. 18 DR. FINE: That was just an example. 19 Clearly there are going to be -- not clearly, there 20 are likely to be things identified on adult gliomas 21 that are validated to be targets that aren't at 22 least obviously there, or may not obviously be 23 there in pediatrics. So, forget about how you feel 24 about the variable EGFR receptor but use it 25 hypothetically as a target that exists on a high 248 1 grade glioma in adult that doesn't exist in a high 2 grade glioma in pediatrics. The question is what 3 do you do with that as far as the Pediatric Rule is 4 concerned? 5 DR. LEVIN: It is the same issue. If it 6 doesn't exist, then maybe it is not as important a 7 target or, in the adult maybe it is not even a 8 target, just an abnormality that is seen. Just 9 because you see an abnormality it doesn't mean it 10 is a target. 11 DR. FINE: That still gets back to the 12 validation. You are arguing that all validated 13 targets in adult tumors will be found in pediatric 14 tumors. 15 DR. LEVIN: No, I would say that all 16 validated targets in the spectrum of astrocytoma 17 should be validated targets in the spectrum of 18 astrocytoma no matter what age, maybe excluding 19 under one, but within reasonable limits they are 20 going to be similar. 21 DR. FINE: So, that reflects your bias 22 that these tumors are exactly the same. 23 DR. LEVIN: I think these tumors are more 24 similar than different -- 25 DR. FINE: I agree. 249 1 DR. LEVIN: -- and I am not quite sure 2 that the reason that we don't see response -- that 3 biologically as patients get older the response 4 deteriorates isn't more a reflection of how little 5 we have to offer and it may basically reflect the 6 fact that we are using toxins and older patient 7 deals with DNA damage much differently than a young 8 person. I mean, there are a lot of different 9 reasons for failure of our therapy besides the 10 difference in tumor generating targets. 11 DR. KUN: And, both in pediatrics and 12 adults these tumors are very heterogeneous, as you 13 know, and the difficulty with trying to make a 14 blanket statement, particularly for the high grade 15 gliomas in pediatrics, is that there are subsets 16 that seem to track more akin to adult tumors 17 biologically and others that don't. So, I don't 18 think you can make that as a blanket statement. 19 DR. SANTANA: Amar? 20 DR. GAJJAR: Another practical point is 21 validating targets in pediatric oncology is going 22 to be very difficult. I mean, to subject a child 23 who is on one of these target derived therapies to 24 biopsy to validate your target is going to be much 25 more difficult than an adult going to repeat 250 1 surgical resections. I mean, you can have targets 2 which are not within the neural system but they are 3 never going to hold up to the same level to the 4 actual tumor cells. So, I think that is something 5 that we have to keep in mind. 6 DR. FINE: Right, but the question that 7 was posed by Victor's was, let's say, this receptor 8 was a validated target in adults but doesn't exist 9 in the pediatric tumor, what do you do with that? 10 And, part of the issue gets to our experience with 11 the RTIs, for instance, where we think we are so 12 smart and that we know that this is the only target 13 and, in fact, it may not be. One of the reasons 14 that this drug X that targets this receptor is 15 causing regression in xenografts may have something 16 to do with its intended target but may have other 17 effects, and do we want to give the pediatric 18 population the ability to experience those other 19 effects if we are not as smart as we think we are? 20 DR. GAJJAR: I think absolutely yes. The 21 answer is a resounding yes. I think what we have 22 learned from these therapies is that they are not 23 as specific as they were designed. I think, you 24 know, the metronomic dosing schedule with ordinary 25 chemotherapy is now supposed to be anti-angiogenic 251 1 and we don't know the mechanisms. In diseases 2 where the outcome is so poor I would not hold back 3 a child from deriving a benefit because we were not 4 smart enough to know the exact mechanism. I mean, 5 the common end target may be the same but they 6 could work through different receptors. 7 DR. FINE: That was the basis for my 8 invoking the question. 9 DR. SANTANA: I tend to agree -- I am not 10 going to call him Victor, I am going to call him 11 Dr. Levin so we can differentiate between the two 12 Victors. I agree with you. I think scientifically 13 if the rationale doesn't exist in the pediatric 14 counterpart you have no scientific basis to test 15 the indication. So, if you are telling me that a 16 glioma in adults expresses X receptor and somebody 17 develops a biologic to treat that whether the 18 Pediatric Rule should be invoked, and there is no 19 scientific rationale to suggest that that receptor 20 also exists in the gliomas why should we invoke the 21 rule for a pediatric population when that specific 22 target doesn't exist? 23 DR. PACKER: Except, you are going on the 24 assumption that all of these targets have been 25 looked at carefully in pediatrics -- 252 1 DR. SANTANA: Yes. 2 DR. PACKER: -- given the heterogeneity of 3 these tumors, the small sample size and the small 4 numbers of patients, and you are going to be saying 5 that we only will use biologic agents that have 6 been already proven to have that target available 7 in pediatrics, when you have just said yourself 8 that you don't even know if it is the right target 9 how it is being used. 10 DR. SANTANA: No, Roger. You are correct. 11 I made the assumption that there was enough 12 pediatric information to know that that receptor 13 was not -- 14 DR. PACKER: I think that is not a fair 15 assumption in pediatric malignant or, for that 16 matter, low grade glial tumor biology. Because I 17 don't think that is going to be up and running -- 18 we don't have the cell lines for pediatric glio 19 tumors; we don't have a lot of biologic data to 20 hold that whole group of children away from these 21 drugs if there is a good rationale -- and I would 22 exclude the child under one possibly, but for 23 anybody above that age, if there is a strong 24 rationale to go ahead with it in adult trials I 25 would suggest there should be a strong rationale to 253 1 go ahead with pediatric trials until you show me a 2 series that has looked exhaustively at enough 3 pediatric glial tumors to know that that pathway is 4 not intact. 5 DR. LEVIN: I think we are arguing about 6 things that we shouldn't be arguing about because 7 the real issue is that we don't really have 8 substantially better tools to deal with the target 9 identification in adult tumors. And the goal 10 really will have to be on a separate level to 11 create systems for studying material from human 12 tumors without having to rely completely on cell 13 culture, which changes the genetics as well as the 14 phenotype, and in animal models. So, we have a 15 long way to go but there is nothing that will stop 16 us, I believe, once we have the tools to use on any 17 tumor from any age patient. 18 DR. SANTANA: I guess the analogy, Roger, 19 is an analogy that was used earlier this morning 20 with APL. If you have APL that does not carry the 21 classic translocation involving the receptor would 22 you subject that pediatric patient to treatment 23 with retinoic acid? 24 DR. PACKER: It immediately goes back to 25 Victor's comment. If we have a way to clearly know 254 1 that that is the case the answer is no. My problem 2 is that the level of science that we have now 3 cannot answer that question for pediatric brain 4 tumors, specifically pediatric gliomas, and until 5 we have that level of science I would suggest the 6 rule should be invoked. 7 DR. SANTANA: Henry? 8 DR. FRIEDMAN: I agree with Roger totally, 9 but Howard has made the point we are going to have 10 to address. The practicality is that the Pediatric 11 Rule will only help us in pediatric neuro-oncology 12 for gliomas. We are going to get no help from the 13 rule in virtually all the other tumors we see 14 because there is no chance in hell that we are 15 going to have an adult trial done in any of those 16 other histologies, adult meningioblastoma for 17 example. Therefore, the only way we will be able 18 to get help from the application of the Pediatric 19 Rule would be if a target is identified in another 20 histology which then has a counterpart in pediatric 21 neuro-oncology. There again, with everything you 22 said, Howard, I agree, and Victor, with target 23 identification we are going to have to be able to 24 apply the rule in a non-histology specific fashion 25 where we are going after a specific molecular 255 1 target and know that that target has at least some 2 prevalence in pediatric tumors, otherwise it will 3 never help us in anything but glioma. 4 DR. POMEROY: I would add definitely to 5 that the danger of just going on histology alone is 6 you will never answer the question. You will never 7 know, unless you somehow study these tumors and 8 develop a mechanism to understand the molecular 9 basis we will never have a rational basis for 10 treatment. We will just be shooting in the dark 11 and using the same histology-based criteria that we 12 have always had. 13 DR. SANTANA: Mike? 14 DR. LINK: If we developed a targeted 15 specific therapy and we were mandating that a drug 16 company applies it to a group of tumors where we 17 have shown that the target doesn't exist, I mean, 18 you would look like a dope, wouldn't you? 19 DR. FINE: Roger's point I think is the 20 important point, which is again one of the reasons 21 I brought this question up. The problem is we 22 don't know so often in pediatric tumors and we are 23 talking about how we are going to apply a rule this 24 year. I mean, hopefully, five years from now or 25 ten years from now we will know the answer, or 256 1 hopefully less than that we will know. But faced 2 with drug X today that is in clinical trial for 3 adults, the way you defend it is -- and, again, 4 that is what this committee is here for, to try to 5 help decide, but if you say that high grade 6 glioblastoma in adult is the same as a glioblastoma 7 child and that the EGFR is -- I am just saying if 8 it is, if it is shown to be a validated target in 9 the pathogenesis of adult glioblastoma then, by 10 definition, it must be a validated target for 11 pediatric GBM if you are saying that GBMs are the 12 same across and so by extrapolation. 13 But ultimately you are right, once we have 14 200 childhood GBMs for which that receptor is 15 looked at, if it turns out it is not there, then I 16 think everyone in this room would agree there would 17 be no reason to use that drug. The question is, 18 given the lack of that knowledge, what do we do 19 when faced with drug X? 20 DR. MEYERS: But I think we are also 21 making a presupposition that our target validation 22 has been a hundred percent effective. Are you 23 prepared to tell me that we know with this kind of 24 pathway identification that these so-called 25 targeted therapies work exclusively in the tumors 257 1 which have the target of interest? I mean, we have 2 heard two examples, good examples. HER2 is 3 expressed in a high percentage of breast cancer 4 patients and only a small percentage of breast 5 cancer patients respond to Trastuzumab. The ras 6 inhibitors appear to work but probably not at all 7 through that mechanism. 8 I think we are assuming a greater degree 9 of knowledge and certainty than that to which we 10 are entitled. I think I would say if a drug is 11 appropriate to be tested in the gliomas of adults, 12 it is appropriate that it be tested in pediatric 13 gliomas. And it is not a question of targeting. 14 We are just not there yet in terms of the certainty 15 that the target is what we think it is and that the 16 validation of the target exists in adults, much 17 less in children. 18 DR. SANTANA: So, you are suggesting that 19 part of the purpose of the conduct of the trial is 20 to precisely not only test the therapy but test the 21 validation of the therapy. 22 DR. LEVIN: But let's put ourselves in the 23 real situation that we want to get access to drug 24 for medulloblastoma. Okay? Now, we know that 25 there are a variety of large groups of signaling 258 1 pathways. Say, ras, sarc, pkc are general 2 pathways. Okay? And, some pharmaceutical company 3 develops and inhibitor of one of the paths that 4 works extraordinarily well in one of the 5 adenocarcinomas but the people who study 6 medulloblastoma know that if they can inhibit this 7 pathway by a variety of different means it also has 8 a positive effect on survival. Now the situation 9 is would the FDA, under this rule, allow the 10 pediatric specialty group to go to the 11 pharmaceutical company and basically demand or 12 expect to be able to get access to that drug? That 13 is what the pediatric population needs, but the 14 question is, is that a valid legal pursuit within 15 the FDA? And, that is what I would suggest might 16 be our future as we move forward with better 17 signaling molecules. It will cover pathways. We 18 will know whether those pathways are important or 19 not. And, within some of those pathways we will be 20 able to pick families of compounds that we think 21 are more likely than not to be better for brain 22 tumors than they would be for adenocarcinoma but we 23 will have choices. 24 DR. MEYERS: I absolutely agree with you 25 but I think that that is what we should be striving 259 1 to get to, but in order to go to a sponsor and 2 compel them to extend a compound to an unrelated 3 histology based on a pathway, I think they would 4 say, well, let's first prove that it is effective 5 in the primary indication and uniquely effective in 6 those tumors which depend on that pathway which 7 have modifications of that pathway. And, I don't 8 think we have that quite yet. 9 DR. SANTANA: Larry? 10 DR. KUN: Yes, I think there are two 11 different issues here. First of all, if there is 12 an agent that shows clinical efficacy in a cohort 13 of patients with adult malignant gliomas, for 14 instance, then I would hate to see that precluded 15 for any reason from trial in pediatric malignant 16 gliomas. I don't think anybody around the table 17 would really disagree with that. 18 I think the second point is a harder one 19 to know. I mean, if an agent is specifically 20 developed for a target unrelated to a tumor system, 21 then at what point -- and this could in CNS or it 22 could be in ALL, at what point do we go and say 23 this drug should be available for pediatric trials? 24 Given the fact that trials are the standard for 25 therapy, so to speak, in pediatrics, you would like 260 1 to say that if there is a biological reason to 2 study the drug and the preclinical model suggests 3 that there is efficacy, then that should be 4 available for the pediatric trial. 5 DR. SANTANA: Roger? 6 DR. PACKER: It is a leap of faith, but if 7 this rule is going to be of help for 8 medulloblastoma there is going to have to be some 9 leap to say that if a drug has been found to be 10 very effective in adult malignant gliomas, and we 11 should live so long to find that drug -- 12 [Laughter] 13 -- that it should be applicable to other 14 pediatric brain tumors. I think you could make a 15 cogent argument that they share enough pathways. 16 We have not really been in that position that 17 often. temazolomide is probably the best example 18 of that and the drug company did not hold the drug 19 back on that basis. I would ask the question a 20 little bit differently because we are not going to 21 be able to answer the first one, how do we roll 22 this back to lower grade pediatric tumors, glial 23 tumors? How do we roll it back when we don't know 24 what those tumors have as far as biologic changes 25 by and large, especially in pediatrics but I don't 261 1 think we know that much in adults either? Yet, if 2 it is effective in adults with malignant gliomas 3 and it is of low toxicity, can we roll it back to 4 anaplastic and grade 2 tumors? My argument would 5 be a strong yes, but I don't have a strong biologic 6 basis to make that argument. 7 Similarly, if you are looking for reasons 8 to suggest a drug should be utilized, it also could 9 mechanism of action. If a drug is being developed 10 that benefits control of leptomeningeal disease in 11 another tumor type, then that drug, because it may 12 have a major effect on tumor spread or 13 dissemination or adhesion, should also be 14 considered strongly for those kind of pediatric 15 tumors where that is a major problem, such as 16 medulloblastoma. So, I think it is more than just 17 the genetic makeup of the tumor. 18 DR. SANTANA: Joe, did you have a comment? 19 I thought earlier you wanted to say something. 20 DR. GOOTENBERG: Actually I would like the 21 discussion to keep on going but at the end I want 22 to ask a clarifying question. So, if there is more 23 discussion to go, it should finish up. 24 DR. SANTANA: Dr. Burger, do you have any 25 comments or want to join the discussion? 262 1 DR. BURGER: Not really. I can talk but I 2 think this is a very complicated subject. If you 3 have any specific questions about the pathology I 4 would be glad to answer them. 5 DR. SANTANA: I just wanted to make sure 6 that you did not feel we are leaving you out of 7 this discussion. 8 DR. BURGER: No, I don't feel left out. 9 DR. SANTANA: Okay, good. Joe, do you 10 want to go ahead and address your issue? 11 DR. GOOTENBERG: From the standpoint of 12 biologics where I think a lot of this is going to 13 be played out, I think that is the arena for the 14 mechanism-specific indications that we might get, I 15 think we need to clarify that what we are talking 16 about here is the Pediatric Rule and that the 17 Pediatric Rule is, number one, license application 18 driven. It only comes in effect at that point. 19 Number two is indication driven, and what we are 20 talking about here is what we would consider the 21 same indications so that under the law we could 22 either mandate that studies are done or give some 23 form of waiver. 24 Already our feeling is that in biologics 25 in the future we are going to have indications that 263 1 combine the mechanism and the disease. This has 2 already happened. For example, APL was mentioned. 3 Retinoic acid is indicated for APL that has the 4 translocation, not for any other APL. So, if that 5 is found in pediatrics, no way would we begin it. 6 DR. HIRSCHFELD: Arsenic is a retinoic 7 acid. 8 DR. GOOTENBERG: Okay, arsenic. For 9 example, also you would look at monoclonal 10 antibodies and look at Herceptin indication most 11 likely -- I haven't looked at it recently -- is for 12 antigen-positive breast cancers. So, we think that 13 biologic indications in the future will be both 14 mechanism and disease specific, and the question is 15 whether we are going to focus on the mechanism and 16 say that studies should be done or not. 17 DR. SANTANA: But I thought I heard Paul 18 and Roger arguing the point that it should be both, 19 that because of the limitation of patient numbers, 20 in pediatrics in this particular scenario that you 21 are proposing, which I think is the more likely one 22 to be, that is, looking at both disease histology 23 and a mechanism, we are not at the point yet that 24 we have enough pediatric information for the 25 mechanism validation that I think if a sponsor 264 1 comes to you with a biologic looking at both 2 gliomas that express X, I think you should 3 seriously consider allowing -- this is the argument 4 that I hear from that side of the table -- that you 5 should allow pediatric patients to have access to 6 that drug without a full understanding whether 7 mechanism X is operative. 8 DR. GOOTENBERG: That is not how the rule 9 operates. We don't allow access to the drug. We 10 either mandate that studies be done or we waive and 11 say studies don't need to be done, and that is a 12 big jump, a big gap there. 13 DR. KUN: But I think what we are saying 14 is that that jump should be taken for the mechanism 15 or for the histology. 16 DR. PACKER: If you don't you will never 17 treat brain stem glioma on a study because we don't 18 have tissue on brain stem gliomas, yet the vast 19 majority of those patients will be dead within 9-18 20 months of diagnosis. You have to make that jump if 21 you are going to affect the field. If the mandate 22 is the only way to get the drug there, then I would 23 suggest you use the mandate. 24 DR. SANTANA: Donna? 25 DR. PRZEPIORKA: Just a request for a 265 1 clarification from Howard Fine, please, because 2 what it sounds like from that side of the room is 3 that a glioma is a glioma is a glioma -- 4 [Laughter] 5 -- similar to the sarcoma story and adults 6 and pediatric patients should be treated the same 7 way. Yet, I recall from your slides that adults 8 and pediatric patients are treated differently. 9 So, my question is are they treated differently 10 because the tumors are different or are they 11 treated differently because of tradition? 12 DR. FINE: Again, as I tried to explain, 13 high grade gliomas are not treated differently. 14 Low grade gliomas are treated differently. Because 15 a glioma is not a glioma is not a glioma, in our 16 ignorance we treat a glioma as a glioma as a glioma 17 within the adult population. Hence, we can 18 extrapolate and say since we do that with adults, 19 we can do that with children too because it may 20 very well be that the real subtypes of tumors that 21 we classify as gliomas may not go across age groups 22 but will go across genetics. But we are not there. 23 So, given our state of ignorance, the question is 24 should we then just treat them all the same? If 25 that is true, then we invoke the Pediatric Rule. 266 1 DR. GROSSMAN: I think the other 2 difference is if radiation therapy were as 3 neurotoxic to the adults as it were to the 4 children, we actually would treat everybody the 5 same. 6 DR. FINE: But, Skip, do you really think 7 that you can get a 70-80 percent response rate with 8 carboplatinum with your average low grade 9 astrocytoma in adults? 10 DR. GROSSMAN: No. There are differences 11 in terms of survival between adults and kids in 12 sarcomas and other diseases that we talk about too. 13 I am not saying that that makes them absolutely 14 identical, but I think if we had severe 15 neurotoxicity from brain irradiation in adults, we 16 would be pushing a lot more chemotherapy in the low 17 grade astrocytomas. 18 DR. FINE: Right, but I think it is still 19 an important point, especially the low grade, that 20 there must be something different about it, because 21 it is not that we can't get to those doses with 22 carboplatinum into a 25-year old but we don't see 23 the kinds of responses that Roger and others have 24 reported. 25 DR. LEVIN: One, we do see a lot of 267 1 irradiation toxicity so we do have a reason to push 2 chemo. Two, all low grade gliomas in childhood are 3 not infiltrated tumors. Most of the low grade 4 tumors in adults are infiltrated tumors. The third 5 thing is that I believe that the conversion of low 6 grade infiltrate of gliomas of childhood to adults 7 approaches 50-70 percent depending on year. In the 8 Gillis article it is basically 70 percent at 5 9 years because they are talking about 10 progression-free survival of astrocytoma being 0.7. 11 So, that being the case, there must be a conversion 12 rate of 30 percent in 5 years just from the Gillis 13 paper. 14 DR. KUN: Just because they fail doesn't 15 mean they convert. 16 DR. LEVIN: Yes, but my guess is they do 17 convert. 18 DR. KUN: Well, a percentage of them do 19 but it doesn't seem to be that high. 20 DR. LEVIN: For infiltrative low grade 21 gliomas. 22 DR. PACKER: If you look pathology 23 studies, I don't think that is correct but low 24 grade infiltrating tumors in pediatrics are not 25 benign processes whether we call them benign 268 1 tumors. Again, we get caught up in how we label 2 these things but those are tumors that require 3 treatment and they are tumors that often are not 4 treatable with radiation because of the extent of 5 the disease, and we need alternatives without 6 biologic data to support what we are going to 7 utilize, and we are stuck with empiric approaches. 8 DR. ELIAS: Yes, I just wanted to get back 9 to the issue of the burden of proof. If one uses 10 histology, I think the burden of proof is in a 11 sense invoking the Pediatric Rule because we have 12 the natural history of the tumor, the biologic 13 behavior, the years of experience with looking at 14 histology. I think when we are talking about 15 pathways we have a different burden, one of which 16 is that we know that very few of our pathways are 17 clear, single, straight line pathways. They all 18 have multiple effects. Many of the drugs that 19 target against one thing clearly have effects on 20 other targets. 21 So, in a sense if we had the issue of 22 medulloblastoma and let's say it shared a pathway 23 with lung cancer, the issue is what would it allow 24 us to invoke? Clearly, not just the fact that the 25 pathway was shared when we clearly have to be able 269 1 to demonstrate in a certain sense not just that it 2 is present but that it is fundamentally important 3 in both tumors, do you need animal models? Do you 4 need clinical data? What level of proof do you 5 need to show that that pathways is, in fact, 6 important in medulloblastoma in order to invoke the 7 Pediatric Rule? 8 DR. HIRSCHFELD: The answer isn't in yet 9 because that is why we are having these discussions 10 to try to evolve what approach to take. Clearly, 11 the modalities in terms of burden of evidence you 12 discussed are all the relevant modalities. It is, 13 in a way, a variation on the figure that we are 14 often asked by industry sponsors, what percent 15 response rate do we need in order to get approval? 16 And, we don't know. We never fixed that number. 17 But I think that when there is some level 18 of consensus in the scientific community that this 19 is the accepted mechanism, then I think it would 20 become relatively apparent. We need to have a 21 formal ruling on it. 22 DR. PAZDUR: Basically it is concurrence 23 of the medical community. So, the issue here is 24 that it is a widely held scientific medical belief. 25 The Pediatric Rule can't be invoked for hypothesis 270 1 generating, basically, it is to take something that 2 is already established and apply basically a 3 diagnosis or a principle. 4 Questions to the Committee 5 DR. SANTANA: I am going to go ahead and 6 try to tackle the questions so we can finish on 7 time. 8 I would suggest that for question A, what 9 general principles could be used to relate CNS 10 malignancies in adults to CNS malignancies in 11 children, that we follow the model that we proposed 12 this morning for sarcomas because I think there are 13 more similarities in adult and pediatric brain 14 tumors than there are with the prior discussion 15 earlier this afternoon. So, I would invoke that we 16 consider histology as a primary -- not the only but 17 as a primary determinant and, in addition, special 18 considerations to molecular characterization and, 19 in addition, something that we have kind of not 20 completely discussed but I want to throw in, with 21 some special attention to issues of safety, 22 particularly with neurocognitive. I know that that 23 is not how the indications are done but ultimately 24 the labeling has to address that. 25 So, I think in this particular group of 271 1 diseases, the brain tumors, I would propose that 2 histology and molecular characterization be the 3 guiding principles but with some special attention 4 to issues of safety as it relates to labeling, and 5 if they don't exist, you know, the sponsors have to 6 say they don't exist. But we should encourage them 7 to look for those when these trials are done so 8 that the labels accurately reflect that particular 9 segment of this population. Larry? 10 DR. KUN: But am I incorrect? Isn't the 11 labeling a secondary event? 12 DR. SANTANA: Yes. 13 DR. KUN: What you are trying to do here 14 is establish the precedent that the drug would be 15 available for study -- 16 DR. SANTANA: Right. 17 DR. KUN: -- and you won't know the impact 18 upon subsequent neurocognitive function, except to 19 be confident that it is a part of the study where 20 appropriate. 21 DR. SANTANA: Right, I just wanted to make 22 people sensitive to that issue, not that it is an 23 issue of the primary indication, Larry. 24 DR. PACKER: But wouldn't that be more of 25 an issue of clinical trial development, of how you 272 1 do the trials in pediatrics, rather than getting 2 the drug to pediatrics? Then, you said you had 3 another meeting coming up on clinical trials. As 4 you move it to pediatrics there have to be some 5 specific safeguards brought in. 6 The one thing I did want to add, and I 7 don't know if it is covered by talking about 8 pathways, is again some statement if also the drug 9 is aimed at a specific pattern of disease spread 10 that would be particularly useful in pediatrics, 11 i.e., leptomeningeal spread. That would be another 12 indication potentially if you were developing an 13 intrathecal drug for carcinomatous meningitis. If 14 that drug showed significant efficacy, to try to 15 make that drug available for pediatric tumors that 16 have leptomeningeal spread. I don't know how to 17 put that in wording but I wonder if that shouldn't 18 be also in the back of people's minds as they put 19 this together. 20 DR. SANTANA: Richard or Steve, did you 21 get that message? Good. 22 DR. HIRSCHFELD: Right, I would fold that 23 into what we call the natural history 24 characterization. 25 DR. POMEROY: I would only add that as far 273 1 as the lack of knowledge in pediatric brain tumors, 2 a number of us feel passionately that we want to 3 fill in that gap and build that up as part of the 4 criteria that we ultimately will use in extending 5 studies to the children. 6 DR. SANTANA: Any further advice regarding 7 issue A to the agency? 8 [No response] 9 For question B, which of the following 10 adult diseases has a pediatric counterpart and what 11 is the basis? I think, if the committee will allow 12 me, I would venture to say that if not all, for 13 many of these I think there are a similar disease 14 correlates and I don't think we need to discuss 15 those further. 16 Then the question that I always have 17 trouble with, which is the issue of the exception 18 examples that keeps coming back -- 19 DR. HIRSCHFELD: This is the last time you 20 will see this question, and specifically that is 21 why we invited Dr. Perlman to see if there were any 22 ways -- again, it is just an attempt to be 23 comprehensive and complete. 24 DR. PERLMAN: Your question with regard to 25 germ cell tumors and their different 274 1 classifications, regarding question C, I don't see 2 any risk or any problem with a different 3 classification of a germ cell tumor as anything 4 else. With regard to whether or not there is a 5 pediatric counterpart of germ cell tumors, I think 6 regardless of the CNS or gonadal origin, and if you 7 are talking about malignant germ cell tumors, there 8 are two biologically separate categories, those 9 that arise in prepubertal or, actually usually 10 infants, and those that arise in postpubertal 11 patients. Biologically, if you are confining 12 yourself to those two categories, either of those 13 two categories are biologically equivalent and, 14 therefore, with regard to the CNS germ cell tumors, 15 the number of infantile malignant CNS germ cell 16 tumors are so extraordinarily rare I am not sure it 17 needs to be addressed with this question. 18 DR. SANTANA: Any other comments regarding 19 that? If not, I am going to try to finish on time 20 and I will invite Dr. Meyers and Dr. Levin in 21 succession to give us some summary comments. 22 Peter, we are going to have some summary comments 23 by Dr. Meyers and Levin. You are welcome to stay 24 on board if you wish. 25 DR. BURGER: Okay, thanks. 275 1 DR. SANTANA: Thank you, Peter. 2 Summary Comments 3 DR. MEYERS: Thank you very much. I am 4 going to start just be reminding all of us of the 5 reason that we came here today. The purpose of the 6 Pediatric Rule is to ensure that we make available 7 to children, and specifically today to children 8 with cancer, the newest drugs in a rapid and timely 9 fashion so that we can learn their value in the 10 treatment of children. 11 The FDAMA initiative which has been very 12 successful and very effective in bringing a number 13 of drugs to pediatric trial is not relevant. It 14 doesn't do that early in the development of drugs, 15 and what we are trying to do is get drugs in early 16 development into appropriate pediatric trials. 17 So, I think that the meeting that you are 18 going to have, which will follow this meeting, to 19 address clinical trial design is really crucial in 20 this whole process because the point I was trying 21 to make earlier and the point that David Poplack 22 referred to in the development of ATRA and other 23 drugs for APML is that for a lot of these drugs we 24 need to find some way to get out of the paradigm 25 that you have to complete the adult trials before 276 1 we can initiate trials in children. 2 I think this is especially important in 3 looking at biological compounds, and in biological 4 compounds it is going to be unusual that we are 5 going to seek to achieve a maximum tolerated dose 6 in the same way that we have done for traditional 7 cytotoxic chemotherapy. We are going to be looking 8 for evidence of biologic activity which will often 9 be seen long before we see severe toxicity, similar 10 to that which we are all accustomed to in our 11 patients with cytotoxic chemotherapy. For that 12 reason, I think it is legitimate to challenge the 13 classic paradigm that one cannot initiate Phase I 14 trials in pediatrics until adult Phase I trials are 15 completed or nearly completed. 16 Someone this morning said we shouldn't use 17 drugs until we have an understanding of how they 18 work, like vincristine. I disagree with that 19 statement. I think there is quite a little room 20 for empiricism in oncology and, as much as I am an 21 advocate of learning about pathways and their role 22 in malignancies and identifying targets to address 23 those pathways, I think we are far from being smart 24 enough to say with certainty that a given pathway 25 is central to a disease, and our targets are not 277 1 always we think they are. 2 This morning we led off with sarcomas. I 3 think that was a wise decision because it allowed 4 us to come to some consensus early on before we 5 tackled the more contentious histologies that were 6 under discussion today. I would suggest that we 7 came to a fairly unanimous conclusion that the 8 sarcomas need to be addressed in the same way in 9 children and adults, and that there really is no 10 reason to use an artificial divide between 11 pediatrics and internal medicine when it comes to 12 the sarcomas. 13 I think when we started to look at the 14 neuroendocrine tumors, specifically the 15 neuroblastoma versus the small cell lung cancer 16 question, we saw some extremely intriguing data 17 and, to me, very educational data but I am not sure 18 that we reached a consensus that any drug which was 19 automatically valuable in small cell lung cancer 20 should invoke the Pediatric Rule for neuroblastoma, 21 and I think we came to a similar consensus in brain 22 tumors. 23 I think the other discussion we initiated 24 here today and we did not complete was what, in 25 fact, will be the basis for the indication 278 1 invocation, and will it be histology alone? Will 2 it be histology and molecular pathology? Will it 3 be some form of targeted pathway? I think the 4 group continues to believe that histology is 5 certainly still the first indication but that 6 increasingly we will be looking at molecular 7 pathology and pathway identification to invoke the 8 rule. 9 I think the final point that I would make 10 that I don't think we thought about completely 11 today is that I think our biggest problem is 12 ultimately going to be one of prioritization. 13 Malcolm reminds us appropriately that our ability 14 to carry out trials in pediatrics is ultimately 15 limited by the willingness of patients to 16 participate and the number of patients who are 17 appropriate to participate, and he has told you 18 quite accurately if we could accomplish trials very 19 four to five years I would be pleased. I think it 20 has been a little less than every four to five 21 years in some of our sarcomas, but we are talking 22 here about earlier trials, smaller trials, trials 23 in patients who have had progressive disease or who 24 have presented with high risk disease and even in 25 that population we are dealing with very small 279 1 numbers. I think it is our responsibility, from 2 the academic community, to make sure that we 3 prioritize the choice of drugs which we wish to 4 pursue, whether the rule is invoked or not, to 5 ensure that we are bringing to the children with 6 malignancies the best that we have to offer. 7 I think that prioritization will be based 8 in part upon availability, in part upon some of the 9 initiatives that were started yesterday at NCI to 10 develop some preclinical screening tools, and in 11 part upon risk/benefit ratios which will be 12 identified at some point in the development of the 13 drugs in adults or in preclinical testing. 14 So, I would say that I have found today's 15 discussion immensely helpful to me and I am very 16 grateful to have been allowed to participate. 17 Thank you. 18 DR. LEVIN: This is my fist participation 19 in some kind of an activity like this so I didn't 20 really know how to prepare my comments, but since I 21 am not a medical oncologist or pediatric oncologist 22 I focused on brain tumors, which I have been doing 23 for the last 28 years. 24 I will focus my comments primarily on 25 brain tumors but will generalize a little. There 280 1 is no question that at least within brain tumors 2 and outside of brain tumors there is some 3 inexactitude and difficulty in making the correct 4 diagnosis and some insecurity bout that. Within 5 adults and children there are going to be defined 6 differences both at a molecular and genetic level, 7 and there are going to be time-dependent 8 differences probably in terms of biologic behavior 9 that we incompletely understand now based on the 10 molecular and genetic understanding we have today, 11 but maybe tomorrow we will understand more fully 12 what those patterns are, why biologic changes in 13 the behavior of the tumor and survival occur. But 14 today we can accept the fact that we don't know 15 everything. 16 Given the similarities that were so nicely 17 put forth by Henry Friedman, we can feel confident 18 that within the sphere of gliomas, nerve sheath 19 tumors, meningeal tumors, germ cell tumors, primary 20 CNS lymphomas and sellar tumors that we can go 21 forth in concert with pediatrics. 22 I think the issue from my perspective is 23 for each individual tumor what is the way to move 24 forward the fastest to get the treatment to the 25 child? Clearly, the fastest way to get a treatment 281 1 for neuroblastoma to children is to do it in adults 2 where you can accrue patients for Phase II studies 3 in three months. It goes forward with the 4 anaplastic tumors as well. 5 So, I think the issue probably shouldn't 6 be so much age as it is getting the study done and 7 validation that against this disease this is a 8 valid treatment. Then maybe lessening the 9 requirements in pediatrics to just proving that it 10 is safe and that the PK supports the dose that is 11 being used, and to focus less on the initial 12 efficacy study trying to rediscover the wheel, but 13 trying to get the therapy into the patients as fast 14 as possible. 15 When you deal with primitive 16 neuroendocrine tumors the world is topsy-turvy 17 because there is no adult correlated. There, I 18 think it is going to have to be individual 19 cleverness, really seriously looking at signaling 20 pathways. People say they would like to do 21 empiricism, but empiricism has gotten us very 22 little distance in the treatment of glial tumors 23 and in the treatment of medulloblastoma. The 24 number of different types of treatments that have 25 really come forward is very small. Basically, they 282 1 are the same that have been used in general for the 2 past decade or longer. So, that really does not 3 hold for primary brain tumors. For primary brain 4 tumors we really are going to have to create more 5 knowledge and attract either the development of new 6 drugs or to get the companies and the inventors of 7 these drugs to allow us to get access to them 8 sooner so we can study them in animals, so we can 9 make a stronger justification for using them in 10 people more quickly. 11 I really don't think that there is an easy 12 way around the solution for finding a therapy for 13 uncommon tumors. I think you have to do it on an 14 individual basis and you have to provide sufficient 15 evidence that can justify its use in that disease. 16 I think random empiricism in this day and age is 17 probably not cost effective. There are going to be 18 too many options coming forward with respect to 19 drugs. It is very easy to make drugs today, much, 20 much more easy than it was years and years ago. 21 The biggest problem today is the targets. 22 So, in that process the companies are going to come 23 forward with large numbers of inhibitors of 24 specific targets, and I think the pediatric field 25 could be overrun by the empiricism and trying to 283 1 combine them. So, I think trying to, at the same 2 time, create a knowledge base will turn out to be 3 the most time effective way of getting treatment to 4 the clinic fastest. 5 I think that that basically summarizes my 6 thoughts, at least from a brain tumor perspective. 7 I am having a hard time understanding how invoking 8 this would really help at this stage. 9 DR. SANTANA: I want to thank Victor and 10 Paul for their summary statements. I want to ask 11 if Steve or Richard have any concluding remarks 12 before I make a final statement. 13 DR. HIRSCHFELD: I would l like to thank 14 all the members of the committee and the speakers 15 who put in the extra effort. I would like to thank 16 the members of our Division, particularly the 17 Director, Dr. Pazdur, and my pediatric oncology 18 colleagues, Drs. Al Shapiro and Ramsey Dagger. 19 And, I would like to thank Victor Santana for once 20 again leading an outstanding panel discussion. 21 DR. SANTANA: Thank you. Susan wants to 22 make a final comment and Jerry wants to make a 23 final comment, and I am going to take the 24 chairman's prerogative and allow them to do that. 25 Susan, please? 284 1 DR. WEINER: Thank you. Just one final 2 question I think for Dr. Pazdur and Dr. Hirschfeld, 3 there has been a lot of healthy and exciting 4 disagreement in this room today, including 5 disagreement from the final summary statements 6 about whether, for example, the adult paradigm 7 should continue or not continue in pediatrics, or 8 whether or not we should forego empiricism for 9 targeted therapies or vice versa. I guess because 10 of that disagreement and because of the anxiety 11 that inevitably incurs in patients and families, I 12 would like to hear something about how those kinds 13 of disagreements in the community will be resolved, 14 and what the interface will be with the cooperative 15 groups and the community in general. I think that 16 that would really put us in a position of going out 17 in the world and saying we are certain that this is 18 going to be a sound and rational procedure. 19 DR. PAZDUR: I think the answer to your 20 question, Susan, is time. One of the reasons I 21 think you have found a lot of disagreement here is 22 that the scientific underpinnings of most of the 23 questions that we are trying to answer are still in 24 their relative infancy. Everybody would like to 25 have targeted therapies. It makes sense. However, 285 1 oncology has been one discipline of empiricism 2 which I think all of us we like to see come to an 3 end and have a more rational development of drugs. 4 But I think that is going to take time and the 5 disagreement that I think you saw here among many 6 of the people represents an absence of data rather 7 than an abundance of data. I think as we develop 8 more targeted therapies and look closer into this 9 field, hopefully, we will have a greater database 10 to come to some consensus. 11 DR. HIRSCHFELD: Could I just add that 12 this will be an ongoing discussion. Today was 13 perhaps the beginning but it certainly doesn't 14 represent the end of this dialogue. 15 DR. WEINER: But there will be some formal 16 structure, some entity that will continue to look 17 at the questions that plague pediatric oncology 18 about access to drugs and about what is to be 19 tested, given the bulging pipeline? 20 DR. PAZDUR: Yes, this subcommittee will 21 continue. Obviously, this is not just three 22 meetings and then we are going to call it quits 23 here. So, yes, this is an ongoing commitment that 24 the Division has to pediatrics. In addition, 25 obviously when we do have pediatric questions, as 286 1 with adult questions about malignancies, we bring 2 in pediatricians that are on this committee to 3 answer questions that we have. But, yes, this is 4 an ongoing commitment that we have. 5 DR. SANTANA: Yes, and I think a follow-up 6 to that is I hope that this dialogue is not two-way 7 but it includes the cooperative groups very 8 seriously in this discussion, CTAP. Sponsors, 9 obviously, are an important point. So, I was glad 10 to see that a number of sponsors showed up today 11 and that Malcolm was here and that other 12 representatives in other roles of leadership in the 13 cooperative group were also here because I think it 14 is not only a dialogue between the FDA and the 15 sponsors; it is a dialogue I think, Susan, that 16 involves other people and I think, either through 17 this structure of additional structures, we need to 18 keep that going. Jerry? 19 DR. FINKLESTEIN: Sixteen months ago -- 20 not long ago -- I had the opportunity to co-chair a 21 meeting held at the American Academy of Pediatrics 22 downtown office in Washington. There were seven 23 groups attending, many of whom are here today. The 24 FDA was there; the public was there; Susan was 25 there; leaders in pediatric oncology were there; 287 1 members of PhARMA were there; pharmacologists were 2 there. NCI was represented by a number of people, 3 including Malcolm. Leaders of the American Academy 4 of Pediatrics were there, and for one of the 5 sessions there were staff represented from people 6 from Congress. 7 The goal of the meeting was to see what 8 could be done by having all these groups sit around 9 the table to look at drugs and therapies for 10 children with cancer and bring them earlier to the 11 child who is suffering this very devastating 12 disease. Now, this is the third meeting of an 13 FDA-created committee. I have to tell you that at 14 that meeting the FDA went into a separate little 15 meeting -- I remember it -- behind me, Richard, 16 Steven, Dianne Murphy and Mac Lumpkin went into a 17 room, closed the door as we were all struggling 18 with this; came out. Mack grabbed the blackboard 19 and said we can help. Obviously, they looked at 20 their mandate and they realized that they could 21 come to the table and accept the challenge. 22 Now, I am probably the senior pediatric 23 oncologist in this room, and for decades, in my 24 mind, it was always "we" and "they." When they 25 grabbed that blackboard I realized it was "we" and 288 1 "we" because there is no question in my mind that 2 they have stepped to the plate. 3 Susan, there is no question in my mind 4 that they are going to continue and I would like to 5 congratulate Richard -- incidentally, Richard is a 6 medical oncologist who thinks like a pediatrician 7 so I have to doubly congratulate Richard and I 8 certainly congratulate Steven for grabbing the 9 balls and keeping it going, and I look forward to 10 further deliberations of this group and I thank you 11 on behalf of my patients. 12 DR. SANTANA: Thank you. I think we are 13 adjourned and I think we have done our task that 14 was assigned. Have a good day. 15 [Whereupon, at 3:40 p.m., the proceedings 16 were recessed.] 17 - - -