DEPARTMENT OF HEALTH AND HUMAN SERVICES FOOD AND DRUG ADMINISTRATION CENTER FOR DRUG EVALUATION AND RESEACH INFORMAL DISCUSSION BETWEEN THE INSTITUTE FOR CLINICAL PET AND FDA INFORMAL DISCUSSION OF LITERATURE REVIEWS OF SAFETY AND EFFICACY DATA FOR PET: 15O-WATER AND 18F-FLUORODOPA Friday, July 28, 2000 9:00 a.m. Goshen Room Holiday Inn Gaithersburg 2 Montgomery Village Avenue Gaithersburg, Maryland 20878 A G E N D A Opening Remarks and Welcome Ms. Jane Axelrad, FDA 4 Opening Remarks Dr. Jorge Barrio, ICP 8 Presentation of 15O-Water Literature Review Dr. Alfredo Sancho, FDA 14 Dr. Patricia Love, FDA 25 Discussion of 15O-Water Literature Review Presentation of 18F-Fluorodopa Literature Review Dr. Adebayo Laniyonu, FDA 93 Dr. Alfredo Sancho, FDA 102 Dr. Ramesh Raman, FDA 108 Discussion of 18F-Fluorodopa Literature Review Summary Discussion and Closing Remarks Ms. Jane Axelrad, FDA 172 PARTICIPANTS Jane Axelrad, J.D. Associate Director for Policy Florence Houn, M.D., MPH Director, Office of Drug Evaluation III Adebayo Laniyonu, Ph.D. Pharmacology/Toxicology Reviewer Division of Medical Imaging and Radiopharmaceutical Drug Products Robert K. Leedham, Jr., R.Ph. Chief, Project Management Staff Division of Radiopharmaceutical Drug Products Patricia Love, M.D. Director, Division of Medical Imaging and Radiopharmaceutical Drug Products Ramesh Raman, M.D. Medical Officer, Division of Medical Imaging and Radiopharmaceutical Drug Products Nakissa Sadreih, Ph.D. Pharmacology/Toxicology Team Leader Division of Medical Imaging and Radiopharmaceutical Drug Products Alfredo Sancho, Ph.D. Clinical Pharmacology and Biopharmaceutics Reviewer Division of Pharmaceutical Evaluation II Jorge Barrio, Ph.D., Professor Department of Molecular and Medical Pharmacology University of California - Los Angeles David Brookes, M.D., Hammersmith Hospital, UK Peter S. Conti, MD, Ph.D. Associate Professor of Radiology, Clinical Pharmacy, and Biomedical Engineering University of Southern California Ron Finn, Ph.D. Memorial Sloane Kettering Cancer Center New York John Hoffman, M.D. National Cancer Institute Jennifer Keppler Executive Director Institute for Clinical PET P R O C E E D I N G S MS. AXELRAD: I would like to welcome everyone to discuss the approval procedures for Positron Emission Tomography products. When we last met I think it was in February, we talked about the guidance documents that we had issued with regard to the approval procedures. Today we are going to turn our attention once again to the safety and efficacy of a couple of PET products that are already in use 15O-Water and Fluorodopa. I thought that we might begin by going around the table and having everyone at the table introduce themselves. Then we have a number of FDA staff who have worked on the reviews of these compounds or who have been involved in our working group on PET and I would like to ask them to introduce themselves. Then I'll just talk a little bit about logistics and we'll start. I am Jane Axelrad. I'm the Associate Director for the Policy in the Center for Drug Evaluation and Research at FDA and the Chair of the PET Steering Committee in the Center. DR. HOUN: I'm Florence Houn, Office Director for Drug Evaluation III. DR. LOVE: Patricia Love, Division Director, Medical Imaging, Radiopharmaceutical Drug Products. DR. SANCHO: Alfredo Sancho, Clinical Pharmacologist. MR. LEEDHAM: R. K. Leedham, Division of Medical Imaging, Radiopharmaceutical Drug Products. DR. CONTI: I'm Peter Conti, I'm Director of the PET Center at U.S.C. and current President of the Institute for Clinical PET and I also represent the Society of Nuclear Medicines Government Affairs Committee. DR. BROOKES: I'm David Brookes, Chairman of Clinical Neurology at Hammersmith Hospital in London and head of the PET neuroscience program there. DR. BARRIO: I'm Jorge Barrio, Professor at UCLA and Chair of this Committee. MS. KEPPLER: Jennifer Keppler, Institute for Clinical PET. MS. AXELRAD: If somebody from FDA wants to start and they all have to use the mike up there. MR. JONES: My name is Eric Jones. I am clinical team leader, I'm in the Division of Medical Imaging Radiopharmaceutical Drug Products. MR. CHEN: I'm Jin Chen from Medical Imaging, FDA. MR. ARNSTEIN: Nelson Arnstein, Medical Reviewer in the Medical Imaging Division at the FDA. MR. WILLIAMS: Russell Williams, Medical Imaging, IT. MR. MOORE: James Moore, Project Manager, Division of Medical Imaging and Radiopharmaceutical Drug Products. MR. LEE: Roger Lee, Medical Officer in the Division of Medical Imaging. MR. MELAGRANA: John Melagrana, Pharm/Tox, Division of Medical Imaging. DR. RAMAN: Ramish Raman, Medical Officer, Division of Medical Imaging and Radiopharmaceutical Drug Products. MS. LOEWKE: Sally Loewke, Clinical Team Leader, Division of Medical Imaging. MR. FARKAS: I'm Ray Farkas the nuclear pharmacist in the Division of Medical Imaging and Radiopharmaceutical Drug Products. DR. LANIYONU: Adebayo Laniyonu, Pharmacologist. MR. LEE: David Lee, Team Leader, Clinical Pharmacology, Biopharmaceutics. MR. SOBHAN: Mahboob Sobhan, statistician on these projects. MR. ZOLMAN: I'm Joe Zolman, Medical Officer. MS. CHO: Kaye Cho, Chief Project Manager. DR. SADRIEH: Nakissa Sadreih, I'm the Pharm/Tox Team leader. MS. AXELRAD: Okay. Thank you. Many of the people who have introduced themselves, some of them will be speaking today about their reviews. But many of the other people have been working on PET for the last two and a half years behind the scenes. And I think it's good for you to get to see their faces and I would like to take this opportunity to thank everybody for all the effort that they've put in. It's been an enormous effort and I think we're making a lot of progress and I hope to make more today when we talk about these two drugs. Just a couple of things about logistics. I have to say this because sometime during the morning they'll bug me that we can only have one person the microphone at any given time. So you have to remember before you talk to turn on your microphone and then when you finish talking to turn it off because otherwise we get interference. This is a working meeting between the FDA staff and the Institute For Clinical PET. However, it is a public meeting. We have an agenda, there are agendas in the back of the room. You can follow along. What we will be doing is having a dialogue. We will be making presentations and then discussing the individual topics. At the end of each topic we will be turning to the audience and giving anyone in the audience a chance to comment if they want to make any remarks about the presentations. I think that's all I need to cover. So with that, I'm going to turn it over to Dr. Love. To Dr. Houn. Oh, to Dr. Barrio, sorry. DR. BARRIO: Thank you, Jane. I'm going to be very brief. In addition to this presentation I would like to welcome Dr. Brookes who has been kind enough to come from London for one or two days. Also Dr. Brookes has an enormous expertise in PET and Fluorodopa, he's a neurologist, and could provide to all of us a medical prospective of the value of the -- Also John Hoffman, Associate Director of the ICN is here. John worked at UCLA in the middle '80s and provided -- actually he was the one who wrote the first IND for Fluorodopa in those years, 15 years ago. And, of course, Ron Finn who has been in the field for more years than Ron would like to remember. A welcome to all of them and thank you for giving us the opportunity to share your expertise with us. This is, as Jane indicated, a continuation of the very intense effort to review the clinical indications of five radiopharmaceuticals. I would like to say that the PET community is excited and grateful to the FDA efforts for what has happened so far. The FDA clinical indications on ammonia and sodium fluoride was also approved -- well, they were approved back in March, I believe, I don't remember the day exactly, but the announcement was made in the Federal Register and right now this is a continuation of that effort in this package of five radiopharmaceuticals that we considered to be or have a very long history in the PET field. I think that's very much what I wanted to say. Since we will start discussing this issue Fluorodopa -- we are going to start discussing water first. I would like to also have the opportunity to say that Dr. Herscovitch is with us and as you remember Peter was in the previous meeting with MIDAC, and of course, he had worked in the fields for many, many years and he will help us certainly answer many of these questions. Thank you. DR. LOVE: Good morning. Thank you very much. I am very glad to see all of you here as well and we do look forward to an exciting discussion. I'm glad so many experts who have been working in this area of water are with us today. We've talked a little bit about the fact that this is now an update after the MIDAC presentation. First perhaps what we will do for a while is just talk a little bit about where we were with MIDAC after the end of MIDAC and then move into the details of where we are right now. This is an ongoing process and as the you know at the end of MIDAC there was some preliminary evidence that had been established, but the Committee had asked us to move forward. So let's look at some of those issues first. Initially water was identified as a natural constituent of the blood and it was clear that we were looking at indications at that time that looked at cerebral blood flow and the Clinical uses that reflected the sites of blood flow, and also at that time we were focusing on injected 15O-Water only, not on water as it was converted from any other product. The reason for that was at that time it was felt that those were different drugs and drug products and we were trying to focus just on one particular drug. At the Medical Imaging Advisory Committee pharmacology information, clinical pharmacokinetics and efficacy information were presented. And specifically for efficacy we were relying on 82 identified articles and I had identified two articles that were considered key out of the overall database and those included articles that looked at sickle cell anemia at that was from work from Dr. Conti's now but in Dr. Powers and also a grub article that looked ischemic disease. Also there were some preliminary discussions about brain mapping as a possible indication. MIDAC's discussion was quite lively. Many of you were there in the audience and remember that there was quite a confounded discussion from time to time about what was the actual database that was being considered, where we able to move forward to approval or not. Some of the committee certainly thought that there was enough information and other members of the committee thought there was not. But by the end of the day it did seem to be clear that the active ingredient in radiation safety risk of water seem to be reasonable but based on the literature of that were presented it was thought that the concept, the proof of the concept that water measured cerebral brought blood flow was not established on the literature basis that was used; and also efficacy was not sufficiently demonstrated to move forward with the indications that were proposed at that time. The basic reasons collectively that were discussed was the fact that the sickle cell article looked at not only cerebral blood flow from 0 15 water but it also looked at FDG uses as well and the database that was presented did not confirm the prospective use. By the way, I know that I'm speaking a little bit in jargon, this is the sickle cell article in pediatric patients that looked at 0 15 water and FDG combined use to identify sites of the brain that might have been at risk for cognitive worsening and at high risk in these pediatric patients; and they were looking at IQ quotients and the like to try to look at long-term follow-up of information. It was felt that there was not enough definitive collaboration of the long-term perspective use of the information in this pediatric study. And for the ischemic disease article, the grub article was accepted as a strong article, but the end points relied not just on cerebral blood flow but also on oxygen extraction fraction and other measures. So there was concern about the independent standing up these articles and the committee recommended that the agency go back and expand its literature search to identify other articles to strengthen the package for the indications for 0 15 water. There was also a lot of discussion about how one might establish a functional indication. How would you move directly to a broad indication for water, say it measured cerebral blood flow or maybe it even measures blood flow in any area of the body, how might one do that and move away from specific disease indications or a specific management indication. So there was some dialogue about that. And again in looking over the transcript it seemed that one of the key issues was that we would need to clearly identify that there was a -- or confirm the fundamental ability of water to measure the parameter; in this case cerebral blood flow, and also establish what was termed a strong link to the clinical utility of the management in a relevant clinical setting. That meaning perhaps an adequate and well-controlled trial with patients that might be obtaining a cerebral blood flow measurement. There were other discussions that occurred also during the dialogue about waters approval and one was the fact that brain mapping might be a unique indication for PET particularly brain mapping in a neurosurgical setting perhaps. Then again, as I mentioned, it was questioned why are we looking at cerebral blood flow only, why are we not considering blood flow in other organs? So since the MIDAC meeting, we have expanded our literature search to look at a number of these different areas to try to find additional information on the proof of concept also specifically looking for clinical studies. And because there were already preliminary data in the ischemic studies in the sickle cell studies I certainly focused attention on these areas but also expanded into brain mapping, looking at surgical decisions where there are other disorders and we will come to those as we talk today and also looking at the broad and potentially broad indication in a number of other organs. Obviously then our presentations will cover these particular areas and I will stop here for a moment, and Dr. Alfredo Sancho is going to go through the clinical pharmacology information first. DR. SANCHO: I am going to be presenting the clinical pharmacology and biopharmaceutical perspective on how to establish the concept of cerebral blood flow for measurement. My presentation will be relatively brief. It is broken up into five sections. The expanded literature base, this summary of the first set of literature we looked at the previous time which is the characteristics of water and the dosimetry information, the second set of literature we looked at essentially will focus on the pharmacokinetic and methodology aspects and I will conclude my presentation was some overall preliminary assessments. The second literature search -- well, the first literature search was the criteria was provided to us, but for the second literature search we essentially followed the following criteria. We looked for water and we looked for the year 2000 and, in my particular case only humans of both genders. This is a summary of the data of the literature search and the first set you can notice that we looked at 82 articles out of the selected five articles. There was one for the dosimetry, for methodology, the second set of articles were 98 articles none of them redundant with the first set out of those 98 I selected 38 as being for my references. They were broken up into pharmacokinetic four, 14 in methodology, and 20 in others. Those others just means that there are articles that did not fit any of those subcategories. I did not put any numbers for textbooks because they were just used for background information, understanding of concepts, and so forth. You can see the total number of 180 articles we looked at, and out of those 43 I selected and are broken down into their subcategories. The literature review criteria when I looked at each one of these articles I essentially was looking for key things, particularly the protocol description and the details it, the clarity of the objectives, the data source; what I mean by that is if it was human and if it was static and dynamic PET imaging which I will get into later, and explain that in detail. Also I looked at particularly what assumptions and correction factors were included into the modeling and how well were those assumptions and corrections described and justified or validated, and finally the compartmental modeling approach and statistical analysis to a certain extent. These are the five articles I mentioned I used for reference for the first set of articles or first wave if you want to call it that way. The key article of this set was Powers from 1988, dosimetry, pediatrics. I will elaborate further hence the asterisk next too it. Before I get to the second set of articles again to give background information of what we were able to obtain from the first set; the water characteristics of 0 15 are that the essentially it's a body, inert body constituent, has no pharmacological side effects except for the radioactivity obviously, radiolabeled. The active ingredient toxicological risk is acceptable and the physical half-life of about 122 seconds. Again this is a paper from Powers dosimetry in infants, no year-end was fouled -- I'm sorry -- no radioactivity was found in the urine due particularly to the short physical half-life of the product. The critical organs were calculated to be the ones that had the blood flow profusion and the water volume of those particular tissues. The dosimetry information from that article again from Powers of 1988 is the following. These are just some of the organs they are not all of the obviously they are just the highest organs with the highest absorbed it does. For adult dosimetry went to Radek 1993 publication from them; again these are the organs with their respective absorbed dose. For the second literature set these are the articles. Again, obviously Kety-Schmidt, 45, 48 being the historical pivotal article all of those articles such as Gambier from '87 and Vidden from '87 again are articles that I will elaborate further and Raichle from '83 I will elaborate and mention throughout my presentation. These are all first authors the next are the same set. There are some good articles that gave us again, for general information and background compilation information up to that date are these. Okay. The pharmacokinetics of water. First of all there is no plasma for pharmacokinetics per se because again of its limitations half-life is for half-life or 122 seconds and it and the Kety-Schmidt model is applied for this product because of its principles behind efficability for highly diffusible products. The fixed principle is an integral part of the Kety-Schmidt model. Now Kety-Schmidt essentially says that it is a one-compartment model. This was validated for the brain tissue by Raichle in '83 which demonstrated excellent correlation between autoradiography and PET for measuring of cerebral blood flow. Ostergaad in 1998 also demonstrated a good correlation between MRI and PET, again also measuring cerebral blood flow. Various methods are used to the Kety-Schmidt to measure blood flow. I will elaborate on that it a few minutes. But overall the Kety-Schmidt model has volume assumptions. For water again I need to emphasize as related to highly diffusible products and we are all cognitive of these. We have seen them in the literature as far back as, like I said, 48 and 45 by Kety-Schmidt himself. As I was saying ever since Kety-Schmidt, 45 and 48, there have been technological advances obviously with pet and therefore how these technological advances have affected the application of the Kety-Schmidt is what I am going to discuss now. Particularly if there are many adaptations but I am particularly going to focus on a compares between the dynamic and static approach of data acquisition or images, the data processing, the effect of time activity curve, and the scan length, and dispersion and delay effect on the model in itself. As I said the first example or first topic I am going to talk about is technological adaptation is that of the comparison of static and dynamic approach. The static approach essentially was the first proposed by Herscovitch and Raichle in 83 respectively. The static approach essentially you ignore of the first 40 seconds after administration and then you acquire single image for 60 seconds and you have one parameter estimation which is the blood flow itself. The volume distribution is a fixed value; throughout all patients the same value. For the dynamic approach it is slightly different. You start acquired images as soon as you have administered the product and you continue to do so for approximately 200 seconds so you have a multitude of multiple slides by positions. There are several parameters to be calculated for that particular subject. Obviously the cerebral blood flow, the effects of delay of the product, and dispersion into the tissue you are imaging. Again, this is more redundant because we are all aware of these articles. I'm just pointing them out for you. Here is a simple chart in which we -- from the article from Quarrels in 1993 in which we are comparing the two dynamic methods of measuring zero blood flow. He calls it autoradiographic method of measuring cerebral blood flow. It is the same for all intents and purposes as the static approach. Some of the issues that were brought up in the article mentioned for that the sensitive -- or the static approach or method is sensitive to the length of the data acquisition which also was brought up by Raichle in 1983 and the static method is also sensitive to volume distribution and variations between the tissues themselves or subregions of whatever tissue they are measuring, in this case the brain. Again that was brought up by Herscovitch in 1983. In contrast the dynamic approach and I don't to call it "model" but the dynamic approach says that it can correct for such deficiencies or difficulties not deficiencies but difficulties in the static by having longer acquisition times and by estimating the volume of distribution for that particular subject. Again initially proposed by Huang 1982. Both of these methods or approaches need to still be corrected the author concluded for the delay and dispersion effect. Again, going back to the chart or the histograph this is -- or broken up line is the line of truth or identity; the solid line is the linear regression of the data as compared and you can see that overall the dynamic approach has an overestimation against the static or autoradiographic approach. Some of the -- not challenges -- that's a little bit of two strong of a word, but some of the issues that were brought up in this article or authors that mentioned this article was that the diffusion effect of water into the brain may not be instantaneous. Again that was not initially proposed but it was proposed by Herscovitch in 1987; that the diffusion effect may have -- may limit or have an effect on the measurement of cerebral blood flow mentioned by Raichle in 1983; the effect of delay, in other words, how fast the tracer gets through the brain itself, was also mentioned by Carson in 1986; and the dispersion and variations within the tissue itself of one subregion compared to another was mentioned by Kanno in 1987. Again, these are just issues that were brought up within the article and discussed. I am not going to delve too much into the data processing itself and the statistical approaches. There are several statistical approaches, Dr. Love will elaborate on these and the effect of them. But here are just some articles and their respective years mentioned; Bittar 1990; Andreason 1996; and the respective issues or the topics that we will be referring to. The next example of technological adaptation is that of the time activity curve, and other words, the loading effect. In this particular study, a study from Gambier from '87, were four human subjects. He compared essentially -- you can see here in this chart -- the gray matter, the white matter and the total brain matter and averaged, if you what to call it that way, of both gray white. The vertical bars obviously are the standard deviations for each one of the time points. He concluded that the one compartment model for measuring the zero blood flow in the particular region may not be a very strong candidates for one compartment modeling. The data fitting may not be as tight as they would like. He suggested that the model needed to have some biological concepts input; in other words from the differences there are between the gray matter and the white matter. And again you could notice the difference between the zero blood flow and both gray matter of the top line the bottom line being the white matter. He suggested that a two-compartment model may be more precise and he also hinted that the cerebral blood flow measurements in all three regions may be affected as well by the time the data precision time itself. Another article is that of van den Hoff in 1993. In this particular case he is talking about the scan length and when should they be done, and comparing again gray matter versus white matter. In this particular study there were five subjects and you can see in the histogram or chart that they gray matter very quickly comes and the noise -- the noise in the measurements very quickly calms down and after approximately 50 or 60 seconds you can get a very nice position that will be representative throughout the entire study. While for the white matter it is very noisy in the first few hundred and so seconds and the author suggested that scans should be done of at least 150 seconds to be able to average out or diffuse the effect of that initial noise we have in the measurement of blood flow. Finally this is an article for all -- again the same van den Hoff of 1993 in which here he talks about the dispersion effect and he essentially did a cerebral blood flow measurement small region of interest, ROI, you can see that this is the one that he was able to measure using this methodology; while that of the true input function which is the one that he was able to correct for both blood dispersion and the delayed effect and demonstrated numerically that there is an effect in both of these issues or errors in the calculations. Again these are just technological advances that the literature has brought to our attention, like I said at the beginning of my presentation as technology has advanced we have been able to fine tune it to our needs or to advances. But over all my preliminary statements are that while they can be used to measure cerebral blood flow, again that is following the Kety-Schmidt model and its assumptions. In other words if you accept these assumptions and you apply your Kety-Schmidt you can use it for cerebral blood flow in small region of interest. But not to sound very repetitious but due to the technological advances with methodology as with technology cerebral blood flow measurements have led to adaptations of the Kety-Schmidt to have the goal for these adaptations has been to improve the position particularly the cerebral blood flow in very small regions of interest, not the total brain, but in small regions in comparing one small region against the other. Some points again that came out from the literature itself that we are posing as questions and some interest areas are what are the effects of the various adaptations that we were able to find? Again these are not all of the adaptations that they were able to find; there are many more and in tweaking if you what to call it that way of the provisional model Kety-Schmidt, what is the effect on the clinical utility of water to measure cerebral blood flow or mapping a very small region of interest, what adaptations of all these many that we found is the most appropriate for clinical use or for specific indication. It is unclear; again most of these adaptations obviously are based on what is the original question what does the author want to get from the study. And finally which adaptations are comparable? There was no study that really or article that was able to say these adaptations have these effects and so forth and so on or compare one data set from another author. I will now be followed by Dr. Love who will discuss the clinical aspects of this. DR. LOVE: Since Alfredo does have a couple of questions would you like to pause for a moment to respond to these or would you like to progress? We can do it either way. DR. CONTI: I think we might want to ask a couple of questions maybe this is a good break. Actually I have one question just from a point of view of literature review. I was just curious if you had looked at the multicenter Japanese trial that compared the various O15 techniques for measuring flow with xenon gas and IMP microsphere techniques that essentially show that all these are relative or essentially equivalent in terms of their ability to assess cerebral blood flow? DR. LOVE: Right. You are speaking of the want that was just faxed to us the other day? DR. SANCHO: Yes, that was just recently faxed to us, so I have not had time to look at it. DR. LOVE: I briefly looked at that and I think he will have to turn your mike off. Thanks. I briefly looked at that article that we received, I guess it was Wednesday, and it does seem to have some promise. I was going to mention it a little bit in my portion that I was about to get to. There are some interesting aspects on how it is comparing water to -- actually I guess three different measurements, one with microspheres, one with two other different radiopharmaceuticals and the fact that it is a joint comparison in that study to different standards of truth across different centers; and in a reasonably large number of patients I think the total again was about 76 patients. Certainly I think would give us perhaps more information. I think since we just received it on Wednesday we have not had a chance to thoroughly reviewed that. Plus as I understand, you were going to "you" meaning someone in ICP was going to see if you could get the remainder of the database that might have been used for the Japanese approval process. Our literature searchers were in English so we did not have all of that; although that particular article was in English, but we did not find it with our search. So we will continue to take a look at that. DR. CONTI: Just for a point of clarification the 0 15 water itself has not been approved in Japan. What has been approved is the use of oxygen 15 labeled gases and part of the approval includes the use of those gases to measure cerebral blood flow; and in that multicenter trial two of the institutions used to that technique for their measurement of the cerebral blood flow steady-state technique and the remaining institutions used correctly 0 15 water. DR. LOVE: We did get some names of some pinnacle articles that are in Japanese and we are trying to get copies of those that would be happy to get those translated. DR. BROOKES: I think Dr. Sancho's point about which application or adaptation is most appropriate is a very important one; because if you wish to know absolute blood flow in a particular area then using a multicompartment model with delay and dispersion corrections and all these other approaches is the way to go. On the other hand if you want to make serial clinical observations and see what effect a pathology or intervention is having you want the most robust and reproducible approach and if you have multiple compartments and multiple exponential corrections you invariably also have higher variants on your measurements. So I think one needs to consider this, and the exact need when you decide which model in fact is the most appropriate in practice. DR. SANCHO: Correct. Hence my question, what is the most appropriate adaptation -- based again on what was the original question what the author is going to do or the researcher is going to have. We welcome these articles and these suggestions, by the way. DR. CONTI: I think in part the answer to that lies with the question being asked by the clinician in that particular clinical setting. You may elect to choose a number of different types of approaches depending upon, as Dr. Brookes said, which point of the study you what to emphasize, and what the specific question is. In fact, you may not want a quantitative assessment of cerebral blood flow at all, you may want relative profusion or regional changes that are done in either a visual or semiquantitative analytical assessment approach, just as we took with the sickle cell data. So I think it depends in large measure on the clinical question and I really would not think that we would what to be tied to a specific adaptation in this type of indication. DR. LOVE: One of the questions we had as we were going through the database is whether or not some of these differences might affect our ability to cross compare and use the database collectively. We certainly recognize that there would be different software approaches that would be needed for different indications and so what we are thinking of is prospectively if we're moving to a label, let's say for this product, are some of these issues things that would need to be addressed in labeling if it was broad indication, if this was a limited indication would we need to rely particularly on certain types of -- on articles that only use certain types of software and software modifications or could we collectively take a wide variety of approaches that might be described in each of the individual articles and synthesize them down to one approach? That is part of what is behind this question. DR. CONTI: I think part of the answer we have dealt with in the FDG. We write the label as, for example, or this may be applicable or this may be, you know, in the sense that it is discretionary as to which particular approach one chooses; and you might want to cite some examples based on the different adaptations that are present. But not restrict us to using a specific adaptation for either all purposes -- so I think if we look at the way we did the FDG, we may be able to find some guidance as to how to proceed. DR. LOVE: I am interested in one of your comments, Dr. Hammerschmidt, about the fact that multiple compartments may be useful in some settings, and -- Dr. Brookes -- sorry, excuse me, anyway, forgive me, please -- multiple compartments may be appropriate in some situations and other more robust models might be useful in other settings; would it be possible perhaps to develop a list of certain types of approaches that might fall into one category or another. I am thinking along the lines of the kinds of approaches that we took with the manufacturing process, let's say of FDG, where we looked at different things and came down to a few uniform approaches; it is it possible to develop that type of a data set? DR. BROOKES: I would have thought in practice, for instance, if one wanted to see whether revascularizing the brain if you have a blocked artery changes flow constructively you would probably use both a complicated and a much more simple autoradiographic approach anyway; in practice you would get much more precise and probably rigorous answers with the compartmental approach. On the other hand, if you have a group of patients and you are following them serially it may be that the autoradiographic approach which is less precise and tends to underestimate flow would still give you much more clear-cut results in that the variance on each measure will be lower. So I am not sure you would necessarily want to restrict people to one or other application, but the likelihood is people would find one was the appropriate and the other less appropriate. DR. LOVE: Many of the clinical articles are quoting the individual authors' adaptation to a software. So maybe they have referenced many of these fundamental key articles that are in the literature but then their own publication describes their own adaptation. Are there any suggestions on how we might look at some of these different on-site local software approaches to make sure that they are all sufficiently appropriate for our purposes at this point? DR. CONTI: I think you are probably going to get into an area that you do not want to get into. Because the technology that we have today and the software that we have today may or may not be applicable 5 years from now to the type of assessments and analytical approaches that we take. So making a recommendation, by the time it is published it may be replaced by something that comes out in the literature. I think this is again in the design of the label I think you want to be as broad as possible with the maximum flexibility to allow the investigators and the clinicians to choose the appropriate technology that is available at that time it order to apply the radiotracer to that clinical setting. Again, the question is really the driving force as opposed to the type of analytical technique one uses. I don't think we want to bog down the indication with that type of information. I'm just wondering if this is a good time to take some comments from the audience since we have some experts in the area who are familiar with the theory and the actual applications would anyone like to make some comments? DR. HERSCOVITCH: Peter Herscovitch NIH. I guess I would like to make two general comments. The first thing is that when you measure -- when you want to measure blood flow with 0 15 water in absolute terms, ML is the blood flow per minute per hundred grams of tissue, you need and arterial time activity curve and a model. I think David is right that there are differences between the model, but by and large all are relatively robust, all approaches, in terms of providing these absolute measurements. But I think the more important question for the clinical use of 0 15 water is regardless of which variation of the model software is it proposed to use 0 15 water clinically to quantitate blood flow or will an alternative approach be used, and that is to inject 0 15 water and obtain a brief image and obtain a profusion map which is proportional to blood flow much like one gets with HMPAO or ECD. I think is just important to separate those two issues; definitely for the functional brain mapping approaches that might be used preneurosurgically one would not need at all arterial time activity curve in lots of other applications, for example, cerebral ischemia, stroke, and so forth, it is a matter of discussion as to when we will be quantitating blood flow with an arterial blood curve which is moderately complex and requires a certain amount of expertise and set up in any facility or when we will be just looking at a map of tissue profusion. I will say that in much the same way that 0 15 water does a superb job when carefully implemented for quantitating blood flow it also does an excellent job in terms of giving you a map of cerebral profusion especially at lower blood flows which are often relevant in super vascular disease where the diffusion or limitation of water is less of an issue. But it is just important to distinguish again between absolute measurements of blood flows which require an arterial time activity curve and obtaining a cerebral profusion map which does not. DR. LOVE: Thank you. DR. CONTI: Just to add one point to that, perhaps in the indication of the labeling procedure we may want to look at something to the nature of in certain circumstances quantization may be applicable using one or more techniques that have been documented in the literature or something to this effect were you have been a choice among the various approaches whether just nonquantitative mapping versus an actual arterial study -- blood sampling study as well. DR. LOVE: Thank you. Any other questions from the audience? [No response.] DR. LOVE: Thank you. Okay. Then we have just completed the clinical pharmacokinetics section. And just recapping then we are going to move from where we are now thinking a little bit more about the questions from the first review cycle. The detailed discussions that I will present or information that I will present will address the brain mapping and Alzheimer's which was not covered in the first discussion; and then some discussion on functional indications in general. And just reminding then, one goal from MIDAC was to establish the fundamental cerebral blood flow measurement. We do think we have that with the information that was presented plus the questions that we were just discussing, and now just moving to the link. Just as an aside, this kind of discussion is also indicated in the draft guidance for medical imaging and under functional indications where we're talking about different approaches one might take to validate a cerebral blood flow measurement; and then there is a lot of discussion in the guidance that talks about what might need to do, what considerations may need to be addressed when you are trying to think about approaches to clinical utility in the spectrum of disease. You all have seen this slide before, these are the different indications that are listed in the draft guidance. This is the one that is published in 1998. The current one actually I am told as of yesterday is now up on the Web for public viewing and it will be available in the Federal Register notice on Monday. So we will be thankful for that one. DR. AXELRAD: Wait. Just one correction. I do not think it is on the Web. It is on display at the Federal Register and it should be -- and the notice of availability should be published Monday, so it should be on the Web on Monday. DR. LOVE: Okay. Thank you, Jane. But at any rate it still has the four indication categories listed, structure, functional, physiological, or biochemical; disease or pathology detection; and then diagnostic or therapeutic management. This functional indication certainly is one that has been identified -- one of the major sets of indications that might be applied for PET products. And when you think of the functional indication one of the questions then is how do you establish this in a setting of disease or pathology or diagnostic management. The guidance talks a bit about the fact that the functional indication is something that is broad. It has implications regardless of the disease or its clinical setting that one might be using so that you don't always have to think about it and is not just for one kind of disease. But often when it is studied the information that is provided in the data set may come from one or more disease settings. And the guidance talks about selecting a setting that might be common, or that might look at a spectrum of disease maybe if it is cardiac disease, very mild disease, or very severe disease, or if you are looking at some other measure, again trying to find a variety of different patients so that you're sure that in the information that is being provided that we can cover all bases and apply that information to everyone. So that then leads us also to considerations about what else might we be looking for particularly if this is being derived from a literature database. The literature information comes not only from the draft guidance for imaging but also for the guidance on establishing evidence of effectiveness and it has a section on the use of literature. So it this setting since we do not have one or more specific studies that might have been developed specifically for a drug approval purpose we then look at the literature, we look for multiple studies, different investigators, perhaps different clinical settings that might be relevant to the overall proposed use or uses that might be under consideration. We would specifically look at the methodology section to make sure that it clearly indicates that we have the prospective design. That's a way of trying to avoid potential bias, also we look for full accounting of all patients that are identified in the design approach. We look to see if the end points are relevant to the overall setting and the questions being asked both specific to that particular article as well as the general application to the overall indication. We look for truth standards. The standards of truth may vary certainly with different articles. Sometimes the standard of truth may be another very objective measurement, sometimes the standard of truth is a clinical outcome. If you think about the FDG with the indication for a myocardial viability those patients had surgical procedures and then an outcome that was assessed. That helped to provide part of the database for the ammonia indication for flow there was -- often many of the articles had coronary angiography as a standard of truth. Also, many of the articles will have controls, other controls, other imaging modality controls and we are often asked about that; well, why not just a control instead of a control plus a truth standard or maybe a truth standard in absence of a control. The reason why we are looking for truth standards in addition to a control is sometimes you can have two things compared to each other, they agree, but one or both of them are actually failing. So it is a failed study and there are many examples of this in the literature. The truth standard essentially provides a way of internally validating that one study because it provides something from which the others can be distinguished. Often you will find the control is not 100 percent equal to the truth, so if both controls are equal to each other but maybe neither one is that close to the truth you will not know that until you have the truth standard there. So it is a way out validating this study itself. Then, of course, we look for image blinding. There is a lot of discussion in the guidance on how to do different types of blinding and how to serially unblind and certainly any of those approaches would be reasonable in many different situations. We also look for details of the analytical plan prospectively stated. The sample size that might be justified and appropriate for the plan; and then finally we look to see if the primary results that are derived from the study are specifically related to the prospectively stated analytical plan. Again, a way of trying to minimize bias in a particular study. So, then as you recall the first review cycle we used the same criteria and we identified, as I mentioned earlier, one key article on ischemia and one on sickle cell. I talked a bit about the fact that our question remaining from the advisory committee was due. We have some confirming information. Also because the brain mapping indication may very well be unique to PET the review has also looked specifically at brain mapping information. Again the advisory committee also gave us some articles themselves to look at, one particular set was from Alzheimer's disease; we looked at that and searched for others in that area that we have also looked at other areas of the body all to see whether or not this information may affect the overall kind of indication that might be developed. Specifically, again, looking at English literature primarily from 1980 to 2000, but there are a few articles that fell outside of that and as mentioned we had articles directly provided by the committee and also by ICP. This is a composite slide which lists the number of articles that were identified in the first review cycle. Those in the second; these are not duplicates and then the combined number of articles. Both cycles produced a large number of methodology articles, this is just an anomaly that these numbers are identical here. A number of case reports, there were several that may have used water but not water by injection. There are several review or summary articles, a few that were identified that actually turned out to be animal preclinical studies; a large number of the second cycle were research studies that looked at either normals or patients to try to use water to explain various types of pathophysiologic mechanisms themselves to understand another disease. Of the cycle, there were three that identified -- that met the actual search criteria, that long list of prospectively designed studies that might have been blinded and had other particular characteristics. There were two that discussed statistical considerations themselves. I am going to focus on these three as well as some of the research article information. But specifically before I get to that as I mentioned the sickle cell article was the Powers and Conti article, I did specifically look for other information. I think I've said it pretty much, but we identified 50 other abstracts but they were actually two studies that seemed to be most specifically related to water. Most of the others in here looked actually at other drugs, not necessarily water, per se. The two studies that we did identify were two that were mentioned at the original advisory committee; they're preliminary. One is a study of six patients and another one is an earlier study upon which Dr. Conti's article was based. So we did not find any new studies to address the outstanding questions in sickle cell. Likewise, for the ischemia question that was specifically left at the advisory committee I was looking for articles that were relying primarily on cerebral blood flow and of those screened, again I did not find something there. And that is the main reason that the search has expanded to the brain mapping in the other areas of disease. For brain mapping, as everyone in the room knows, brain mapping has been used extensively in research for a number of years both for PET and non-PET products. One of the main areas of research has been to identify the eloquent cortex and identify certain areas of normal or abnormal function and this information certainly has assisted in understanding the physiology and pathophysiology -- I keep hitting this button, I think I need to put it down. It is not cooperating well with me. But it also has provided a lot of information on physiology, pathophysiology and response to medications. Also the brain mapping has been used in a number of articles to discuss information on image guided surgery. These six articles are for one group and it is the reason why I am identified them, because they have all used the same research technique, mapping technique, same software, and these are the authors who are listed. And they have a set of six articles that are listed on this slide that reviewed a number of basic functions in the brain and specifically they were episodic, novel, and practiced recall of complex narratives, short- and long-term verbal memory, facial recognition, and there were also statistical power considerations discussed in this article or set of articles. These articles looked at, depending upon the article, 13 to 33 patients. Some of these are the same patients, some are unique patients and we have contacted the authors to ask for specific information on the exact discrete unique sample size. They are normal, they all use the same types of repetitive testing methodology for identifying memory areas. They use the same type of image criteria, they use the same statistical analysis concerns. And I am presenting this primarily as an example of the type of information that comes from research. And all of these six articles have identified a consistent side of the brain for these discrete areas. When you take that type of fundamental basic research and then apply to surgical planning, the goal has been to identify those normally functioning areas that have been identified in basic research and identify them and use them in anatomic mapping. The reason that is interesting to us is that one of the things in the guidance document talks about when is the identification of normal an important thing. Most of the guidance document talks about abnormal disease and the like. But there are times when identifying a normal area is relevant and this certainly seems to be one of them. But so then identifying the eloquent cortex to protect it during surgery obviously we are all aware of the historic perspective of this and the fact that there has been historic anatomic mapping that may not be most appropriate. It is based on population based information and it is not necessarily related to an individual unique site. Intracortical stimulation has been used but of course it means that you have to have another surgical procedure, you have to be alert sufficiently to cooperate with the testing and so that leads to challenges. So there are several imaging modalities have been used, PET being one of them. Devices has approved software for coregistering PET and SPECT with MRI or CT scans but their device approval does not specifically relate to approvals for preoperative eloquent cortex identification. But you can co-register the anatomic markings with findings and different images. Just briefly, there were three articles that were identified on the second review cycle. They are listed here. One looked at PET and functional MRI, one with neurosurgical planning for pediatric brain surgery, and another one looked at cortical stimulation. These provided preliminary information but as you see these articles are all very small ranging from two out of four patients who had both PET and functional MRI in the Freed article up to seven patients in Brookheimer's article. The reason these articles were interesting is because certainly the first two actually described to question that the surgeon was asking. They looked at what was the lesion, where was it, and why would I need to know something about either the location of speech or motor function and then talked about the outcome. But they were very brief articles and did not described a lot about the actual discrete changes in the surgeon's decision with and without PET. Then the Brookheimer article compared to PET to electrical cortical stimulation and in four patients who had positive electrical cortical stimulation all four also had increases in cerebral blood flow to the same area and in what appeared to be three patients who had negative stimulation they also had either decreased or no change in cerebral blood flow and there was one false positive. These articles then are helpful because they describe the clinical thinking and the outcome and one of them correlated the results with the gold standard but they were not clinical trials, there wasn't any information upon whether they were prospectively designed, they weren't blinded, the statistical methodology was not stated, plus they were also very small and the like. But the value of those articles is that when you combine them with the articles that were identified in the first cycle that provided 18 patients and then 16 patient case reports that brings us to a total of 48 case reports, information specifically anecdotal types of information on patients and their responses and collectively the strengths and weaknesses are similar to those that were just mentioned. Then there are three articles by the Bittar group which looked at localization of function again comparing it to intracortical stimulation, one looking at again functional magnetic resonance at one that looked at the actual location of the lesion to see whether or not it had any effects. I just want a pause on those for just a moment. The first article looked at 20 patients for presurgical planning. They specifically described the PET-T statistics that they were using for image volumes, they defined it as the mean change in flow which was equal to the voxels divided by the average standard deviation across the pool. They used the intracortical stimulation as the standard of truth. They compared the site of the maximum T statistic information with the site of the positive intracortical stimulation. They defined as concordant if they were within one centimeter of each other within the same gyrus. They prospectively looked for sensitivity specificity, positive and negative predictive value and identified the values all of the capa statistics which correlated specific terminology; none through almost perfect. And in their results of those 20 patients the sensitivity was 85 percent, the confidence intervals are shown in brackets. The specificity was 86 percent, positive predictive value 96 -60 percent. For their CAPA correlation they looked at the actual T statistic and felt that if the T statistic was greater than 4.75 then there was 100 percent correlation with the intracortical stimulation. It the T statistic was less than 3.2 there was 100 percent correlation negatively and if this T statistic was in between those two values then the conclusion was that you needed to go the next step and do the intracortical stimulation. So even though this is a small sample size it did identify prospective statistically significant results, it did appear to be a prospective study; the image criteria and the statistical methods were quite well described. But what wasn't stated in the article was whether or not it was blinded, any discussion of the sample size justification wasn't there and whether this was exactly a prospective hypothesis or whether these findings on the T statistic correlations were just observational was not entirely clear in the article. Nevertheless it does provide specific information that is useful in our thinking. The other article for Bittar looked at functional MRI with PET; again, looking at the same type of methodology, the same types of test, the same type of image analysis information and the like, and they found that they could identify an average site identification difference of 7.9 millimeters with a one to 18 range that turned out to not be statistically significant. The authors talked a bit about the fact that functional MRI information for mapping and water information are somewhat different in looking at different aspects of the neuronal activity and blood flow so that that might account for some of their differences. The one negative particularly on this study if we were going to use it to provide some support for PET is that it really didn't correlate their outcomes with the clinical out comes. In other words, if a patient had had surgery what would have happened? And again the specific test was not mentioned. I do say here that functional MRI is not approved. And on one hand that is a weakness for us but all the other hand we can use collective clinical information at times as clinical standards when we are looking at overall evaluation for a drug product. Then finally the last Bittar article looked at whether the central or precentral lesions had any specific effect on the need for mapping. This one article looked at 51 patients, 32 with central lesions, 19 with noncentral lesions and the control was based on 16 patients contralateral side image information. This study noticed that morphologic changes in the actual configuration the brain occurred as one would expect with central lesions, and in this situation their suggestion then is that mapping for patients with these types of lesions again will be most beneficial. A question that certainly has been derived from looking at all of the information, both the research articles as well as the clinical articles is that the dose of water varies considerably. Sometimes it appears to vary simply on the basis of the number of tests that are needed for any individual patient and the number of sites that need to be mapped. So that would make sense that you would have to give more doses to look at different areas of the brain. On the other hand, the dose that was given initially also varied. These articles also are coming from different countries and there are different approaches in terms of the dosing that might be used. Some articles used three tests per task, some used more than that. So the actual maximum dose that is needed is not clear, but just calculating some of the doses could probably approach upward 200 to 300 millicuries if you just add it up from some of the different articles. I think we have talked a bit about the different algorithms after Dr. Sancho's discussion and that is what this comment is about, how do the different articles or different methodologies affect the use of brain mapping? And thank you for your comments on that. So then preliminarily for brain mapping there is preliminary information for the clinical utility in neurosurgical planning, some of the information is very supportive, we do have controlled studies, open label controlled studies that might be providing some information for other testing hypotheses. Shifting then to Alzheimer's disease. As mentioned earlier Alzheimer's was not discussed during the first review cycle. In the second cycle we identified one blinded controlled clinical trial and eight, open-label, known-disease observational comparison studies. The eight studies looked at a total of 104 Alzheimer's patients and 87 age controlled. Collectively all of them identified abnormalities associated with decreased regional cerebral blood flow in or near the temporal or parietal area. These articles also looked at other aspects of brain function of recruitment or plasticity of the brain in other areas of the brain to compensate for the abnormalities in the temporal parietal area. The open-label study is shown -- and was authored by individuals here, some of whom are in the audience, and looked at a blinded evaluation of PET in the diagnosis of probable Alzheimer's disease. This is apparently a prospective study that compared the results of images with the clinical diagnosis, the patients enrolled had probable Alzheimer's, Parkinson's and/or were normal controlled. The standard of truth was the Washington University probable Alzheimer's disease criteria which in their hands was correct in 62 out of 64 patients and gave a sensitivity of 96 percent apparently specifically against autopsy. The protocol used a blinded clinical diagnosis as well as a blinded dementia rating by one set of physicians. The had a separate set of neuroimagers who did blinded image evaluation of the PET images of the image criteria were specifically defined and there was a training session for the blinded readers. The read was specifically comprised of an independent read which was a committed read and then a consensus was to be obtained as necessary. The results of the study, it enrolled 39 patients, 13 probable Alzheimer's, 11 normal control and 15 Parkinson's disease patients. There was an independent reader agreement in 33 of 39 patients which gave 84 percent initially and the remaining six patients for which there was not initial agreement, two of the three readers agreed independently; so based on the description on the study my interpretation is that they moved forward and did not do an actual consensus read on the next six patients. This is just a brief summary of the type of information the diagnosis, the Alzheimer's rating score, Parkinson's or elderly, the number of subjects, the number that PET identified as normal, Alzheimer's or other disease. This is derived from the article and reorganized for purposes of the slide. The results of that then produced based on the protocol image criteria the sensitivity for PET was 38 percent and the specificity was 80 percent. The authors discussed the fact that this was an unexpected finding so they adjusted the image criteria to take all abnormalities of profusion, not just those that were thought to be particularly unique on the basis of the temporal parietal region alone, and that increased the sensitivity to 92 percent and the specificity to 85 percent. This is in comparison to their sensitivity for the clinical standard of truth of 96 percent. The authors of this article have a lengthy discussion about the results of the study and some of the points that they address are the fact that this is a prospective -- although it is a prospective study these are known population patients and they may not represent the patients that one might see in a prospective setting of patients who are coming in with clinical questions. So that they recommended the need for a prospective trial and recommended that such a trial should have a longitudinal clinical follow-up for confirmation of these findings. The strength of this article then certainly appears to be prospective. There is a standard of truth. The image criteria and the statistical criteria are stated, it is blinded and there is a balanced discussion about both the strengths and weaknesses of the study itself. The weaknesses are that it is of known disease population. The primary, the most robust analysis of the database relies on a post-hoc analysis and there is not much discussion about it actual sample sizes that were selected, different sample size certainly could provide different types of information. So then for Alzheimer's disease there are certainly potential data for the use of 0 15 water in the evaluation of Alzheimer's disease and thus far I was not able to identify a lot of information about what is the most appropriate clinical setting for this use. For example, is it most appropriate for this to be a screening modality? Is it most appropriate for this to be a diagnostic tool? Or, is it most appropriate for following patients and their disease progression? So those are questions for which I would like to hear more information and discussion. Other organs. We looked at 75 other sources for a variety of different areas, heart, lung, liver, spleen, skeletal, total muscle, and bone marrow, for all of these the information that I was able to identify appeared to be primarily very early looking at methodology, various hypotheses, looking to validate different approaches, formulas, and software at this point. Most of the articles were in normal volunteers and not necessarily in clinical settings. So taking this information then, and just thinking about the only other issue we haven't talked about and that is the safety question, the original review cycle did not identify any major toxicity concerns. The only question at this moment is how would we be addressing the issue of radiation limitations if this was being used for multiple testing for neurosurgical brain mapping. So we do needs more information on what are the most appropriate maximum doses in this setting and this may also affect pediatric dosimetry. Again depending upon the brain mapping questions. So our interim assessment at this time is that there are preliminary data to support a functional use in measuring cerebral blood flow to assist in evaluating areas of higher cortical function. The reason for this is, this is based on brain mapping information, the fact that Alzheimer's information is looking again at a higher cortical function of memory and pediatric data from the sickle cell article we are also looking at cortical function. So, based upon what is found, this seems to be the kind of indication that the information supports. As Dr. Conti asked earlier, what about the article that was provided from Japan. And as I mentioned I am planning to look at that in great detail. I do what to see the other articles. Also there were questions posed to us a couple of days ago -- for the audience's sake we had a T-con a couple of days ago -- where some of these things were discussed. There are other data that has come to our attention. Other comparisons with 0 15 water and other radiopharmaceuticals; we will take a look at that in the overall context and our goal is to develop an information package that we're going to be able to take back to the advisory committee to either address their specific questions that were posed to us of the end of the last meeting or to present an alternative package for their consideration. With that I will stop now and we can go into the discussion. DR. BARRIO: Any questions from the panel? Peter. DR. CONTI: If I can backtrack a little bit and maybe just to make a comment on this interim assessment point just before we leave that. Evaluating areas of higher cortical function, maybe I am reading too much into it, but it implies neuropsychiatric and actual brain function implications. I am a little bit hesitant to go in that direction. I would rather you be looking to evaluate, you know, brain structure as opposed to in terms of its profusion and flow as opposed to the sequela of alterations and do the pathology in terms of the implication of the word "function." So I think we need to maybe just clarify that so that we are not confusing the issue here. Because you also eliminate mass lesions and things like this from the evaluation process like tumors or aneurysms or actual sites of infarction, things of this nature. DR. LOVE: That question is one that I think we need to talk a lot about in terms of what is the indication, is it a functional structural indication or what have you. I think from the information that has been presented and reviewed, the data are not directly looking at structure, per se. It is a flow which we are considering more as a functional physiologic biochemical process with the "structure" being provided by the comparator imaging modalities, CT or MR. Now, are you speaking of structure meaning the structure that is present happens to have abnormal flow and is that how you are using structure in your context here? DR. CONTI: I think we are talking about the physiology of the entity within the brain that you are looking at whether it is the white matter or whether it is the gray matter or whether it is a mass lesion or whether it is the sight of infarction. We want to deal with the flow to that or profusion to those lesions of interest or regions of interest as opposed to -- maybe I'm reading too much into this, but I do not want the implication of higher cortical function to imply behavioral issues or things to that nature. That is the only point that I am making. DR. LOVE: Right. And I certainly appreciate that. I think that is a point well taken because the data that we have reviewed has not looked at psychiatric issues. I'm speaking specifically of the eloquent cortex primarily memory function, cognition, and the like. So we can look at language to clarify that. DR. CONTI: I had an answer to the question you were talking about with other organs just as an aside. Keeping in mind the well-known clinical question of how well a tumor is profused or what the blood flow to the kidneys are or to the heart, how do you see us dealing with this from FDA's perspective given the fact that these are questions that have been asked for longer than I have been in practice certainly and studied by multiple techniques? How do we deal with that when we now have a tracer such as water that can measure flow and profusion? DR. LOVE: I am not sure that I understand. DR. CONTI: I guess my question is, should we be looking at a broader indication if we're going to get into the function indication what we're talking about flow or profusion yet we have used -- we have posed these questions with other modalities at techniques for many, many years and we have a tracer that can measure this, why are we necessarily worried about whether there happens to be an 0 15 study in the kidneys or something to this nature? DR. LOVE: You are talking about the total body blood flow indication as opposed to blood flow to the brain. I think there this is sort of the overall question on function of how do you do that and there are several different questions that has come up with the guidance document discussion as well in terms of what is the relevance. If a product is doing exactly the same thing, let's say it is an ejection fraction, with an echo cardiography, let's say, we still would look for a study in a clinical setting where ejection fraction is being measured to make sure that this modality can be well characterized, we can look at its sensitivity and specificity to describe its parameters. But we are not necessarily expecting that the sponsor of the product proves the relevance of measuring ejection fraction. So I think in a setting -- when we are talking about a clinical setting, we are talking about simply taking that setting in which this information is going to be generally used and studying it in that model. If on the other hand you're looking at something new, maybe the quantitative value that is being used for 0 15 water is different, or you are providing information, let's say for neurosurgical planning that is not available with other modalities, then in that situation we would probably ask for something more specific and unique to that clinical setting so that we know and can put in the label whatever information might be relevant for a physician who is looking at that and trying to make assessments. And as technology changes they will know, okay this was the original database, something has changed, they would know how to make a modification and adjustments. We have often been asked, okay, so what do you do about flow everywhere? Or a total body indication for another modality? And I think we have not had this discussion, just giving an analogy from another imaging modality, what we have suggested in one situation, let's think back to the ionate contrast era many years ago; the suggestion was that for a total body indication one would look at the areas of greatest risk because at that point it was the procedural risk and the injection intra-arterially that was a major concern. So the organs that were selected were those at highest risk and you would look at that and move forward with an indication for a set of uses. For MRI we have been suggesting something different; looking at where is the device most limited, and looking at those areas and not necessarily looking at the others. So I think maybe we could talk about what areas might be most relevant to think about for flow, perhaps you can direct us to those areas that are being used most often and we can look to see if there is sufficient information. It does not mean that we often absolutely have to have something in every single area of flow, but we would need to consider what are the most representative areas where documentation might be useful. DR. CONTI: I guess my point here is only that in certain circumstances measuring flow may be appropriate, regardless of the modality or technique that is used, from a clinical point of view or from a quantitative point of view. We will frequently do quantitative FDG studies for the input functions and we require such information and we extract that information as part of our study in sort of the dynamic phase of the examination, but there may be circumstances were the question of flow per se is of interest as has been the case throughout the years knowing the vascularity of certain types of malignancies has been an important question in treatment planning as well as in diagnostics. Knowing the blood flow to organs transplants and things like this so these are all very common question that come up on a day-to-day basis and we measure that parameter. So the issue is, if you have a tracer that is useful in terms of being able to measure flow or profusion should we not be considering a broader implication? DR. HOUN: I think we would be interested in considering it. I guess when we were looking at the different organs, you know, we do in our literature review we were looking for some data to support its use in transplants and not that it was done it two patients but it was able to demonstrate flow and it was accurate either because of the comparative outcome whether was clinical or eventually, you know, flow measured by another standard, their needs to be some data so we can understand where flow in various different organs would be important to know about. I think where we came up short was that in the literature we sought more as research and hypothesis generating and not so much a comparative controlled trial showing that in fact it does meet a standard. Unless they are articles and if you are aware, you know, whether they are in English or in other languages we would be happy to review them. DR. BROOKES: Could I address some of the questions you raised towards the brain mapping side of things? DR. LOVE: Yes, you may. DR. BROOKES: Firstly you talked a little bit about FMR and PET and you made the point that the location of -- were often a few millimeters separated by the two techniques. This is not entirely surprising because as you say, they are looking at a rather different quantity. The FMR is looking at the oxygen and the venules draining activated areas. The PET is looking at the nutritional blood supply to the brain area directly with water. The other thing I think you have to realize is that the brain is a naturally moving organ and when you activate the brain there is local patterns of oscillation and movement as well; and FMR is also very sensitive to that, it picks up movement as well as oxygen changes. So, in fact, you are going to get different answers with these two techniques. I think we do not know for sure which is the better answer. I suppose what one would say is that PET is directly affecting the function of the tissue rather than it's venous drainage which is why I feel there is a strong indication for using the technique but this is a question clearly that is going to have to be clarified. The second thing you questioned was the dosimetry; there seemed to be a big variation in how much water or how much radiation was being used in these studies. And I think one of the things is that as our scanner technology and software has developed this has radically diminished. For instance, with current state of the art PET cameras where you are working with about a 3 millimeter reconstructed resolution and you are looking with the three dimensional rather than two-dimensional data acquisition and you can probably get away with 8 to 10 millicuries per flow measurement. So if you want to do 12 or 16 runs to define speech and motor areas around your AVM or tumor you should be able to do that nowadays with 100 to 150 millicuries of total radiation and possibly less which is only a few miliseverts in practice. And so the days of having to use 200 to 300 miliseverts I think are rapidly coming to an end with the latest scanner technologies. Finally, you talked about the various brain mapping software that are around and whether there was a gold standard in which one should be used. And I think the fundamental point -- and Peter may well have comments on this too -- is that they are all doing the same thing basically, they are all currently, anyway, taking the general linear statistical model and applying it to a large set of vaxules. They have slightly different ways of normalizing brain shape into a standard space and they have slightly different ways of smoothly and filtering some normalized blood flow proportionately to the global mean and others use an ancova. But fundamentally there is not a great deal of difference between these approaches, and the most critical difference is probably the statistical threshold you finally assigned for positive or negative findings. And that still again a very difficult thing to determine absolutely. People have done phantom studies and have decided that a threshold of around the P.001 is a reasonable protection against false positives. But in fact the actual number of independently functioning vaxules is critically dependent on what smoothing and filtering you apply and you can calculate this. There is no single gold standard answer as to the exact threshold that is appropriate in any given situation. DR. LOVE: Thank you. Dr. Herscovitch? DR. HERSCOVITCH: Peter Herscovitch. I guess there is an awful lot of things your presentation raised but maybe we will just stick to the issue of brain mapping now. I agree completely with what David said in terms of the statistical analysis but there are perhaps a couple of other issues which are, if anything, more important and maybe more difficult in the clinical arena. One is the issue of task selection and, for example, things that one might just say have a subject read and then recite out loud what the reading something like that -- some of the very complex tasks which activate large areas of brain and it is important to pick your task very carefully. And this is especially true in the more complex tasks such as language and memory, somatosensory and motor function mapping which are often critical in say brain tumor surgery are a lot easier but the task selection is very important. The other is, you brought up the issue of the accuracy of localization in terms of millimeters in comparison with FMR, and I agree with David, they measure different things and in fact at lease now I would say that PET is at least scientifically felt to be the gold standard. But the issue is not how well PET or FMRI localized a specific function only, it is more importantly how well one can register the functional imaging data with the neurosurgical data because even if PET or FMRI are a few millimeters off where one would get a brain stimulation response, the important thing is to identify the exact area on the brain when you are operating that you have identified with PET. And I think if one could get to a few millimeters that would be very good. So those are other important issues and using functional brain mapping in the neurosurgical arena I think will be important. But the overall technology involved is really quite critical and perhaps quite complex both tax selection, statistical image analysis, and then identifying the PET foci on the exposed human cortex intraoperatively. DR. LOVE: Thank you. DR. CONTI: Let me just add two comments about brain mapping also. As I said earlier, this is also a technique in evolution and so new approaches and more sophisticated approaches may come into play and so the indication needs to be flexible enough so that we allow for this evolution of technology to go on. One comment that was brought up on the neurosurgery planning was the lack of outcome data or actually more appropriately the lack of change in the position approach. Part of the problem I believe with these types of studies is that once -- unless one is completely blinded to whether the patient has this type of data, it becomes a little bit difficult for the neurosurgeon to ignore the information. So designing these types of trials poses an ethical dilemma for us given the fact that this functional imaging when it is available is almost directly utilized as part of the patient management. So when you read these articles there is probably little in them in terms of, well I decided to move my scalpel five millimeters to the right because of the neurosurgical -- not because of brain mapping approach. When in fact if they did not do that they would be a trouble. So I think there has been an assimilation of the information already into clinical practice when it is available that sort of limits the types of designs of the studies. DR. LOVE: Yes, that certainly was well appreciated. It would be a difficult ethical question if one were to think about a randomized trial in the most rigorous manner. A couple of the articles, I think particularly the Kaplan article, tried to describe the original question and the fact that the approach has changed. This was in the study of the small patient series of data. They actually try to describe the fact that because the lesion was located at a certain place, I think in one particular situation they discovered that the patient had dominant speech on the opposite side from the lesion, that was an unexpected finding. So they described what they were able to do. We thought a little bit about what kind of trial might be beneficial in this kind of a setting. One of them might be a cohort study where you take two different centers, one that uses PET routinely, and one that uses another approach and try to match the lesions and the locations. So it is really more of a case controlled study as opposed to a randomized study. That is one potential way to do it. Another might be to actually do a little bit more along the lines of what was done in the Bittar article where you were actually looking at this identified lesion in comparison to intracortical stimulation had just augment it a bit with the clinical outcome information. That articles' approach was very reasonable. It could just be a few other adjustments to that article to strengthen it, but its basic approach was reasonable and at the clinical outcome. DR. CONTI: On the Alzheimer's you had posed a question on the Alzheimer's disease components of this as to the clinical setting not to preempt discussions that we may have in the future on FDG with Alzheimer's but the issue gets back to the same issue that we faced with FDG for cancer as to whether we could use it as a screening tool or whether to define either a known at-risk population for the indication. I think in general the known or at-risk population is a safer bet when all is said and done. Now, the at-risk population could be defined a little more broadly, you know, in terms of the clinical or genetic presentation or risk or something to that nature where there is data to show that there is a possibility that they have the disease to just screen the general population. I do not believe that is what we are necessarily trying to pursue. DR. HOUN: I guess building on the article that did the post-talk analysis to look at the new criteria for defining Alzheimer's disease where the sensitivity picked up, you know, we would view that as hopefully needing a confirmatory in another -- with another group of investigators looking at that new definition can they repeat that higher sensitivity in a larger group of patients? So I think, you know, we have some preliminary supportive stuff but I am still thinking that we need some further data. And if it is it that clinical setting in terms of people at risk, you know, that is a very different -- that is a very important setting, a very clinically meaningful setting, and it would be very important to confirm the performance of this test in that population. DR. CONTI: Again, the weakness was identified that it was in a known clinical setting in part of the analysis here but I think that is actually the strength of the approach also. I mean, I understand the need to perhaps look at this across a broader population but the tests we are not really asking for the test to be applied in that type of setting, we are asking for the tests to be applied in the known or at-risk population. So, when we go ahead and do a trial as you're suggesting it would be tough I think to go outside of that risk population. DR. HOUN: I would say that like under phase two development trying to find out the sensitivity specificity of your tests, you know, people do the tests in the known and no normal population to try to work that out but then in phase three as a trial to show efficacy if the goal is to use it to help clinicians who are not sure of the diagnosis but perhaps that's your highly suspect population that that is the clinical setting. DR. LOVE: I think what part of the concerned that we have sometimes when we look at an indication like the use to diagnosis, monitor, or predict disease, let's say it Alzheimer's, raises a number of different questions, different kind of ethical questions, what if you are misdiagnosing, what are the consequences to that individual if the diagnosis is in error? Alzheimer's, as we alike know, is a difficult diagnosis to make and if you go through a number of different procedures to rule out other diseases and usually it is a diagnosis of exclusion in the end, so we need to figure out a way to be sure that we're looking at the right data set the right patient population. I agree with Dr. Houn's comments that in phase two one would try to take a known disease population. I think our real concern with this article is that in the known disease population using the prospective image criteria the results were not what one would have expected in a known disease population, so it was the post-hoc assessment, reassessing the image criteria that gave the results that were more positive. So that is the kind of situation where one really would need to look further to see what is the actual criteria that should be used. It is not so much that O 15 water does not measure the flow, it is which piece of information is going to be useful in what clinical setting; and it may be the probable disease patient. DR. BROOKES: I was just going to say with the big improvements in statistical parametric methods I do not -- I have not seem that particular trial you referred to, but one can now take a database of 50 or 60 normals, take individual patients and match them with the whole database on a voxel basis and one can show that there are clusters of voxels which are abnormal and this vastly increases the sensitivity now to pick up the early disease such as Alzheimer's disease compared to just drawing regions on a temporal lobe or a parietal lobe. I think, again, in these new developments a lot of these older trials are going to become redundant because the sensitivity of these techniques to find preclinical and early cases is going to immense. DR. BARRIO: I was given this article by Powers, I guess, who was discussed briefly the other day and you mentioned that water did not really have enough sensitivity, et cetera, et cetera, in this population. One of the comments that I made during the discussion a couple of days ago in the conference call was that while we are currently working this is a little bit of anecdotal evidence in the new marker for plaques and tangles in -- patients. One of the problems that we had was how to select patients to really demonstrate the principle. If we receive a patient from everywhere from the clinic outside the center of station, et cetera, et cetera, that we found for a population of 20, 25 patients that probably five or six were elderly patients, by the way, were Alzheimer's disease based only on the FDG scanning. And, in fact, that was confirmed by other evidence later on. Later information, I mean, more recent information seem to indicate that at the time of death only one or two clinical diagnosis of Alzheimer's disease is correct. Then before death you can imagine that that number has to be smaller than 50 percent of course. Then since Dr. Herscovitch is here we were talking about his work, I was wondering because the comment that Pat made was that the standard of truth was sufficiently, you know, they have the confidence or at least you have the confidence that you really have Alzheimer's patients and therefore you know the trick here. You are using -- you are testing water with that standard of truth and maybe that water may be a better standard of truth than they are. I don't know, I am asking the question because we are really facing this problem, and I do not know exactly what your experience is in that regard. DR. HERSCOVITCH: Thank you Jorge for asking me that insightful question. The gold standard of clinical research in Alzheimer's disease is a very difficult problem that totally transcends any issues of imaging modalities. I think it is a very relevant question. The clinical dementia rating skills that are used primarily as research tools have by and large been validated, and I do not at all have the figures in my mind at all, but they are definitely in the literature. I think many of them do a relatively good job when validated against long-term clinical course and ultimate pathology. I do not have the figures. But I think there is a more important philosophical question that each Jorge's points raise, and that is, if the gold standard in an imaging trial is a research clinical valuation, for example, and the CDR rating scale developed at Washington University and then if an imaging modality say has 85 to 90 percent sensitivity, specificity, and so forth in relation to that clinical standard in patients who are clinically -- well, you cannot say clinically definite, but not necessarily pathologically definite that's pretty good for the imaging modality, but I can in my office apply the clinical criteria upon which the imaging modality was evaluated and presumably get the 100 percent even if the clinical tools are not ultimately correct pathologically. If your gold standard is any sort of clinical evaluation then at best the imaging modality can only do as well as the clinical evaluation and therefore why do the imaging modality, I can do a history and physical and some neuropsychological tests in the office, sort of being a devil's advocate. I think there are perhaps more important questions -- well, just backtracking, obviously if your imaging modality does not do a good job in terms of sensitivity and specificity in patients with clinically definite Alzheimer's disease or CDRs of about 1.5 to 2 then you sort of go home. But if you show them that, then the next step is looking at patients with possible Alzheimer's disease and the clinical relevant situation is an elderly person comes in with some memory problems and there is a rather broad differential in the things that we always test for that we never fight like thyroid disease and B12 deficiency and so forth we rarely find, but issues such as depression senalnoncolia and other neurologic diseases and can any diagnostic test distinguish between those. So the clinically relevant population is perhaps patients with early dementia or early signs of cognitive decline and then what is the gold standard there, and the gold standard would have to be as that paper suggested, a long-term clinical follow-up first to see if they develop Alzheimer's by clinical criteria and secondly if they develop Alzheimer's by pathologic criteria. So I think that is be real critical issue. But if a test can do almost as well as a clinical evaluation then one might wonder why one has to use tests. The other issue is that a couple of other points with Alzheimer's disease is FDG might be better if one wanted to do the such a study then 0 15 water for a couple of reasons by and large in the clinical intends to get better, I think, image quality was FDG. Also FDG is probably more directly related to the underlying pathology disease that is loss of neurons and the need for substrate and the use of substrate to support neuron metabolism and blood flow but really that can be tested empirically. It is quite easy actually to do FDG and 0 15 water in a clinical trial in the same patients and then follow them. And the last point is with regard to that paper as David said, technology has improved not only in the method of image analysis but we should point out that that study was done and I participated in the patients; that patient study is done with a PET six tomograph in the mid-1980 which had an 18 mm full with half maximum resolution; now resolution is better. That may be good or bad but I think there is a risk in applying those results to scanners with current or much better resolution. DR. HOFFMAN: John Hoffman. I have actually a paper coming out in the next month or two in the Journal of Nuclear Medicine, the largest pathologic series of patients with FDG PET. One of the points I have to agree with Peter that was a mid-1980s paper, the technology has changed and the reason you could find one paper with water is that it is not what is done. FDG is used for Alzheimer's disease. Another issue is, even in the best of hands the clinical diagnosis of probable AD is only about 80 percent in a very good memory disorder clinic when they subsequently come to autopsy. So that is the best that you are ever going to get with clinical criteria as the gold standard. And even more profound is the fact that if you -- and there are series that substantiate this -- there are different pathologic criteria for Alzheimer's and in some series that agreement between pathologists is only 60 percent. So very problematic. The gold standard is causing a lot of problems and that is ultimately what one would probably have to use. DR. AXELRAD: Can you just tell us who you are with? DR. HOFFMAN: I am at the NCI but I was formerly a professor of neurology in the PET -- Emery. So I think FDG and UCLA has a tremendous amount of data now accumulating on various patients in various stages of memory disorder with Alzheimer's disease. I do not know if David has any other comments with dementia. DR. BROOKES: The only other thing I would say which is Parkinson's disease is an interesting condition because about a third of the patients have concomitant dementia and Alzheimer's and using FDG I have not tried water. You pick up preclinical cortical changes of Alzheimer's in about a third of these patients before they show any frank dementing problem. We are now following a population of these to see whether we have it right. Whether those are the third that will get the dementia. So if you are considering setting up trials to try and validate this approach they may be one model system. DR. CONTI: Just to bring us back to the focus on the profusion a little bit, the community would like to see this move on to the MIDAC assessment and I wanted to get the sense from you folks at this point where you see it and whether or not we can take a few moments to really take a hard look at the actual language of the indication and whether or not it is appropriate at this time. DR. LOVE: I guess that's part of the goal for our talking is what kind of package could we have to take back to them. And part of my concern is, as I was mentioning, there are certain kinds of articles and confirming information that were recommended that we have not yet been able to identify. So we certainly would like to know whether or not there are other data available that someone is aware of, other articles that could be used specifically. We certainly will look at the other information that you have identified. The conversation that we just had about Alzheimer's and whether or not this is an appropriate model for studying, maybe we could pause and just address that a little bit more in the context of functional indications versus disease specific types of indications, why are we looking at this and then come back to your overall question, Dr. Conti. That might be useful. I think certainly in doing the literature review I was looking at Alzheimer's as one of the models saying we need to find a strong clinical link in one or more different models where 0 15 water has been used as some type of clinical setting that would provide information to the advisory committee to indicate that there is a clear benefit. One thing I think we're hearing in this discussion today as well as the take on the other day is that there are times when you need a very quantitative assessment and there are times when a less quantitative by perhaps a ratio or a percentage type of assessment might be useful and other times when a more traditional qualitative assessment might be used. Maybe we need to think about different examples to present to the committee because one of the issues certainly discussed last time was, does this measure flow? And then how are we using these different measurements in a clinical setting? And finding information that might provide solid examples to the advisory committee on how it is used in each of those different settings might be beneficial. Certainly one of the committee meeting comments was the fact that the brain mapping was valuable. I would say that I would like to -- I don't want to throw that out necessarily as a specific indication because it is the kind of advancement that would be a major indication in many ways. So if there are other data for that, I think we would certainly want to see any other information that is available for that. DR. CONTI: I think we can take this maybe one step at a time. Is there a consensus that there is adequate validation of theory that this tracer can measure blood flow? We believe there is. DR. LOVE: I think that is what Dr. Sancho's slides conclude that it does. DR. CONTI: So that answers basically the first MIDAC concern. Is there sufficient literature and data to show the validation? And I think that that is put to rest. The second question is, are there ample clinical indications for the measurement of cerebral profusion given that we now have a tracer that has been validated to measure cerebral profusion and blood flow. And the issue is, has cerebral vascular disease been a question on clinician's minds for longer than any of us have practiced in this room? DR. HOUN: I think the question is that it is a validated flow marker and the question -- that's been put to rest. The question is, is there data to support its clinical use or a clinical indication. And where is that data? DR. CONTI: Right. DR. HOUN: If we do not find it in the literature, you know, how people have yet to write it up, or it's not -- I guess we are concerned that we need to provide people with literature data to support that inadequate and local controlled trials, this particular indication has been shown to be valid through the population that has studied it. DR. CONTI: Well, I mean, I guess I take issue with that. And I take issue with it because of the fact that the tracers that have been available in nuclear medicine and in studies that I have used to compare 0 15 water to, these are other approved tracers 0 15 water has been shown to be equivalent to those tracers. The literature is there. I am not sure I understand what the problem is here. I mean, do we have to do controlled comparative trials when we've got literature evidence that it is a validated technique; the technique is shown to produce the same results as inferior techniques, in fact. I am not sure what you are asking us to do. DR. HOUN: I think there are two sets of issues. One is where are we in relation to the advisory committee's recommendations? And the other is, how are you going to -- actually what kinds of studies would provide the kind of evidence that has been requested? The first advisory committee meeting the presentation identified the fact that cerebral blood flow per se is an indication that has been given to many other radiopharmaceuticals. Those radiopharmaceuticals however are not really quantitative approvals, they are really qualitative approvals. Many of those drugs are certainly being used in other ways clinically but the actual labeled indication is a qualitative assessment or measurement -- not measurement -- evaluation of blood flow. The settings for most of those were usually stroke and just in the sense that that was the type of setting that was used. But given that context, the data that were presented to the advisory committee were not uniformly accepted for the approved -- for the recommended indication for measurement of cerebral blood flow. So we have been asked to go back and identify clinical settings where it is useful. And that is why we are looking at the literature for specific clinical settings. The other question is then, well, why isn't a direct comparison to one of the other radiopharmaceuticals sufficient in and of itself? I have two responses to that. One is, you have identified that there are some other articles and data that are available for us to review and we will certainly look at them and there very well may be something in those other data that moves us past where we are right now. But with the information that we have identified the direct comparisons were usually in a more of a research type of setting. They were not really looking at what we were talking about a moment ago, the clinical setting where the patients are walking in for a questionable evaluation. They get that evaluation, you look at the data from both, and whatever gold standard might exist. Usually the articles are direct one-to-one comparisons and the gold standard is not there or it is not in the context of the controlled trial. So it is more of a trial design question. It is not so much that we are saying that it does not measure flow, but when we look for the database the administrative record that we would be able to develop and represent to the advisory committee we are looking for those links. Now, they may be there and we have looked at some articles and I am sure there are many others plus our approach was focused on specific sets of information. There maybe something else that is there that we have not seen, but if you have that, if you know of those articles then we would be very glad to look at them. DR. CONTI: I also would like to point out that I do not think when we went to the MIDAC originally, I believe we were, for whatever reason, requesting a quantitative approach; is that true? DR. LOVE: It was measurement and that is what the measurement would mean. DR. CONTI: I think we have the opportunity now to redefine the indication to remove that and put that as a potential caveat for elective application to the user. And in that context then we now need to have this discussion once again; do we have evidence that we can use as in a nonquantitative setting? DR. LOVE: The data certainly that we have been looking at has all been quantitative in one way or another. The data that has been reviewed has -- I am just saying, that is the data that we have looked at. We have not looked at data for qualitative assessment. DR. CONTI: You have. You have in fact; the sickle cell paper that I wrote is nonquantitative. It is one of your lead articles. DR. LOVE: It is nonquantitative but all of these are based on a quantitative assessment of the images, they are based on a quantitative use of information in the software and the analysis. All the algorithms are quantitative. So I am not saying that we cannot reassess the image, I mean, the indication, but they are all based on some quantitative assessment or reconstruction of the information as R.K. is saying here, but I think if we're going to move from quantitative to qualitative then we would be looking for specific information that would justify that should the indication be limited to qualitative. I think we need a broader discussion before we can make a final decision on that point. DR. CONTI: I think we need some definitions. DR. HERSCOVITCH: Yes, by quantitative blood flow I think we are meaning at measurements of blood flow in units of MLs per minute per hundred grams of tissue and that is what the field more usually means; not necessarily quantitating local tissue radioactivity. I think that is the major dichotomy; do you have an arterial line to quantitate blood flow, or do you use images of tissue radioactivity and those images can be used visually as in the Powers' Alzheimer's disease article, or those images can be used with actually very powerful quantitative statistical methods that Dr. Brookes referred to earlier. But still one is using images of radioactivity, not quantitative measurements of cerebral blood flow. And I think the literature strongly supports, if not proves, that the image of regional radioactivity that one gets by the intravenous administration of both water and then a brief, say, one-minute image is a very, very good representation of absolute cerebral blood flow and probably better than those images that are produced by what Dr. Conti called the inferior or less well developed methods such as HMPA0 or ECD. So I think it is -- I do not want to use the word "indisputable" but there is very strong evidence in the literature that you can get a profusion map of relative cerebral blood flow without an arterial line. And I think that issue should also be laid to rest at one should speak more to the clinical utility of 0 15 water and not whether it either measures blood flow in absolute or in relative terms. DR. LOVE: Right. I think we certainly agree we need to deal with the clinical utility. And I agree that we probably are looking at semantics here. When I am speaking a quantitative I am meaning not just the millimeters per minute per how much tissue, but also whether or not many of the articles we are looking at the actual ratios of information or percent change, and the like, or these measures where part of what was being used to validate the clinical endpoint, so my quantitative is larger than that. But I certainly recognize, yes, qualitative is there. My concern would be if we are going to go purely to a qualitative. I think I would have to look back and see how many of the articles that I am really using are looking primarily at some type of quantitative assessment versus how many are qualitative. That would have some bearing on whether the indication should be moving one direction or the other. The other I think is as we have talked if you are seeking an overall functional use and we recognize that this use varies from qualitative to a very precise measurement per minute, if we know this, then what is the rationale for ignoring one large set of data and only looking at something else. Not it might be that is just what has been confirmed and we can think about that, but it might also be that there is other information that are out there that could be brought to bear to try to answer the whole question at one time. DR. CONTI: I am going to read an indication that I wrote down here and I want you to consider this. The intravenous -- and this is paraphrased from the CeraTech indication with a little addition. "Intravenous injection of 0 15 water is used for the detection of abnormalities in regional cerebral blood flow or profusion and may be used to obtain absolute measurements of blood flow." DR. LEEDHAM: The indication you are paraphrasing, isn't it in the CeraTech indication doesn't it say "in stroke"? DR. CONTI: Yes. It goes on to describe several indications. It talks about stroke, Alzheimer's disease, epilepsy, TIA, migraine, tumors of the brain, we can certainly -- DR. LEEDHAM: Right. Our defined clinical settings. DR. CONTI: What we have done, for example, with FDG, is provide examples, and we can do the same here. Maybe we want to choose the examples the there is evidence for, but not restrict the use to the settings where we believe that there is no evidence for leave it opened ended. DR. HOUN: I think the FDG oncology indication says, in assisting in the evaluation of patients who either have a diagnosis of cancer or are highly suspect for. So that includes the clinical setting. It does not say it is going to diagnosis a specific type of lymphoma, breast cancer stage, et cetera. But I am wondering if we can pick up on Peter's question on clinical utility. Because if we do have a functional indication in terms of measuring cerebral blood flow, this part of to assist in evaluating what is the community looking for. Evaluating what? DR. CONTI: Cerebral vascular disease is -- and you can cite examples of what those are; they could be stroke, they could be transient ischemia, they can be in -- I mean whatever you what to have as cerebral vascular disease defined in a broad sense, it could be intercranial mass lesions such as aneurysms and tumors. DR. HOUN: I think it is the articles and will -- the reason why we ended up putting this higher cortical function is because we felt that the articles that met our criteria for having at least some degree of prospective evaluation of a hypotheses with endpoints that were controlled and had a defined image evaluation plan, meaning that they were of some level of scientific rigor, that those were in those patients with brain mapping, the sickle cell article, and it was not in patients with tumor or stroke, and so, how do we get around that. Otherwise we could open it up to anything that people theoretically feel a drug would be useful in, but where is the data to show it? And I think that is what we are looking for; and it may be out there, we may not have selected the articles. DR. CONTI: The stroke data is very extensive. It just happens to use different techniques. Blood flow as a measured parameter in the stroke literature is pervasive whether you use 0 15 water for the gas steady-state technique measure of flow is almost ubiquitous. And whether you are using it to make treatment decisions as to when to reprofuse or intervene or whether you are using it to do the diagnosis, I mean, I have got a stack right here of stroke literature. So I am not really sure I understand again what the -- if you are looking for intravenous injection of 0 15 water to diagnose stroke specifically I think you may be creating too narrow a focus. You may what to broader your focus to something like cerebral vascular disease and leave it opened ended so that people can apply it. DR. LOVE: Are the articles that you have there, are they looking -- you say they are looking at other techniques and I am assuming that you mean extraction fraction and other test beyond cerebral blood flow; or what do you mean by that? DR. CONTI: Well, many of those exposed oxygen extraction techniques rely on the parameter of measuring blood flow in terms of the calculations for oxygen metabolism in fact. DR. LOVE: I agree. But I am saying that was the issue that the advisory committee had. Because the original database in the original indication that was proposed had to do with cerebral vascular diseases, the original data set talked about Moyamoya disease, stroke, and several other settings and that was the data set that the committee felt needed to be strengthened. So we are looking for something specific. Now, if we can't do it on cerebral blood flow alone which is what their request was, and if it means that we have to look at the others and the other parameters looking at metabolism or oxygen extraction fraction then we would be expanding beyond 0 15 water. I am not saying that we cannot do it, but our goal was to find the data set to focus on 0 15 water by injection looking at cerebral blood flow with the indication for cerebral blood flow; but if we need to expand it, then maybe we need to talk about that. DR. CONTI: I do not think you need to expand it. The techniques have been proven and validated to be equivalent in terms of their ability to measure the parameter of interest. There is literature on it. I'm sorry that you do not have the multicenter trial for your review, but in that study there were five centers from Japan, two used the gas inhalation method and three use the 0 15 water, and the data are identical in terms of the measured parameter. And that has been shown in multiple publications. So I am not, again, trying to -- I do not want to expand the indication to use other tracers here because that introduces another level of discussion. But the point is, if you're interested in measuring cerebral blood flow there are both multiple PET as well as multiple other modality or other tracer techniques to measure this parameter and many of them have been compared across the literature. DR. LOVE: I think if you have the other literature as we were saying earlier, we would be very glad to take a look at it. Our challenge particularly with the stroke literature that was available to us was that this was the original indication that was proposed, and we are trying to find something to directly answer the questions. So if you have it or if we have some alternative information that we can look at and maybe modify the package then we can look at that. Thank you. DR. AXELRAD: Okay. We have been going for about two and a half hours. Do people want to take a 10-minute break and then start the FDOPA and then maybe break for lunch at around 12:30 or something like that? I would like to get some of the FDOPA done before we break for lunch. Is that all right for with everybody? [No response.] DR. AXELRAD: Okay. [Brief recess at 11:20 a.m.] DR. HOUN: Dr. Love is going to introduce FDOPA. For those of you who are sitting down and those of you who are not sitting down we are going to start anyway. DR. LOVE: Okay. Thank you. FDOPA as Dr. Barrio has mentioned earlier was one of the other PET drug imaging products that we agreed to consider as part of the initial assessment of commonly used drugs, PET drugs and in the process of establishing the procedures. Basically FDOPA's question that we were initially considering is that it is used in Parkinson's disease but we also wanted to take a broad spectrum approach and look at it from a prospective of a biomarker itself and see what other information might be available, and if there were any other disorders that might be relevant. So the information that you will be hearing for the remainder of the day will focus on that. We have a pharmacology presentation a clinical pharmacology clinical pharmacokinetics presentation and then the clinical efficacy assessment of the literature that will be presented. I guess given time, since we are going to have one hour, what we will do first is hear from Dr. Adebayo Laniyonu who is the clinical pharmacologist toxicologist on this particular product. Then Dr. Sancho will speak about the clinical pharmacokinetics if we have enough time within that hour. Then we will break for lunch and come back and do the clinical part after lunch. So this is Dr. Laniyonu. Dr. LANIYONU: Thank you. Good morning. My name is Adebayo Laniyonu and I am a pharmacologist with the Division of Medical Imaging and Radiopharmaceutical Drug Products. This morning I will be discussing the preliminary results of the research side of pharmacology, toxicology of 18 Fluoro-L-Dopa. During the course of a normal NDA review preclinical pharmacology toxicology usually for causes -- on the safety of potential efficacy of a new molecular entity. And since we are dealing with an imaging modality today I will be focusing my attention on the pharmacological business for the use of FDOPA and then I will be examining the toxicity in the profile of FDOPA. Those are the two key areas I will focus my attention on. In the search criteria that we used was that we looked at peer review publications written in English from 1989 and online database search only. We also solicited materials from the PET community and I will concentrate my focus on the preclinical animal data. This turned out 98 articles that form the core of my review and I have literally read all the -- From the perspective we find that more often than not preclinical data are usually incremental in their knowledge and approach so that I was able to look up the entire knowledge base. My presentation outline is via slide. First of all I will be considering the synthesis of metabolism of dopamine and Flurodopamine. I will then go on to discuss the similarities and the differences in the way the body handles FDOPA and L-dopa; I will then go on to the toxicity profile of fluoro-dopa, and then finally I will end up with my preliminary assessment of the literature. From a historical perspective it is interesting to note and that until about the 1950s dopamine was simply considered an intermediate in the biosynthesis of the -- Indians with norepinephrine and epinephrine. Today the role of dopamine as a neurotransmitter in its own right is virtually well established. This scheme shows how dopamine is synthesized in the body. To synthesize starts from the amino acid Ltyrocene which forms a drop to lead to form L-dopa. This term is the regular must have it the biosynthesis of dopamine, -- to form L-dopa, L-dopa is -- performed by the enzyme AA AAAD -- to form dopamine the dopamine performed is then transported by the body and the transporters seeking into the vertical do and off where it is going. Dopamine can be fully metabolized to dopamine -- and also by transferase. Also L-dopa itself is susceptible to the enzyme -- transferase that continue to fuel -- L-dopa. Dopamine can also be -- performed by the -- enzyme to form the hydroxphenylacetic acid. So also the -- can be acted upon by the -- transferase to form -- [indiscernible] to finally give you homovanillic acid. I would like to point your attention to this, what is called the metabolic -- L-dopa as well as dopamine -- coming back to them during the course of my discussion. This simply shows what happens or how the body handles F-dopa. The major significant difference between the schematic I showed you previously and this one is that this time -- that's -- bypass and Fluoro-dopa is acted upon by the -- of enzymes that's acted upon L-Dopa to form fluoro-dopamine -- to be stored as fluoro-dopamine. Fluoro-dopamine is also metabolized by the -- enzyme to form the fluoro- -- [indiscernible] -- which finally is converted to homovanillic acid. And this schematic is taken from one of the publications of Dr. Barrio and it simply shows what happens in -- fluorodopa can cause a blood brain barrier and taken up into the neuron -- fluoro-dopamine -- and then you will -- metabolize to form dopa, and finally you can list of dopamine by -- to release dopamine for -- Does the body had left dopamine FDOPA and L-dopa simply because the sensitivity to COM TE show specific differences enzyme inhibitors affect metabolized profile radionuclides. I would like to play Georgia 10 today is was metabolized through the during the course of my discussion. This schematic shows what happens or how the body handles FDOPA. The major significant difference between the schematic I show to previously and this want is that this time I'd no longer -- directed by bypass with nor dopa is acted upon by this down the do so in that mind and acted own dopamine to be restored as we dopamine. The second shows social done from Dr. Barrio and this schematic is taken from the what other publications of Dr. Barrio added simply shows what happens in the own originated your go above can cause about brain barrier and taken off into the nor Ron taken place to form for dopamine the a divided does the bus ago and then you will metabolized to form L-dopa and finally you could accompanying and by to release the dopa made last all the video that receptors of that the dopamine and do the individual click assets I presume that's the Willamette that metabolized drifted from hair because neurons suffice to say that you could actually metabolized in the floor dopa and about brain barrier. I was asked to focus my attention on this to through fluoro L-dopa fluorodopa and there is a in the preparation to actually be outside of to the last all of the to form -- L-dopa and also there can be done about essential to form fluorodopamine. Fluoro-dopamine cannot cause a blood that barrier so you end up two utilities fluoro-dopa as well as -- that can cause a brain barrier. What then happens is that the fluoro-dopa causes the brain barrier and is converted to fluoro-dopamine. In review further the fluoro-dopamine also carries the -- that leads to more specific noise and that can actually affect the image. Conceptual basis for using FDOPA as a biomarker what did similarities in the patterns of metabolized to F-dopa and L-dopa. F-dopamine and dopamine is stored in present not vested vessel. Toxicity profile of FDOPA no study examined that -- toxicity profile of FDOPA. No study quantified purities -- acceptance. Toxicity profile focus on the active ingredient alone. F-dopa and L-dopa metabolism is qualitatively similar. Safety database for L-dopa F-dopa administered well last does this is L-dopa the basic PC/TE requirements of the active ingredient appeared to be met. Pharmacology toxicology preliminary assessment 18 F-dopa, L-dopa can be used as a biomarker of the dopa -- The question then is what is -- metabolites or do -- metabolites affect image or results? And there were two general opinions in the literature. The first opinion and this is really the only elaboration that I got that is -- to these findings. It says that there is really no effects. That you really do not need to take into consideration the presence -- needs to give the -- and the two -- that I will be discussing today is the metabolite -- L-dopa and -- and they are -- from the same group -- cerebral metabolism of fluoro-dopa in primates. From the first paper, this shows the plasma metabolite profile without any of the enzyme -- neither did they use -- or the count inhibitor. And it shows the chromatograms of -- human blood taken 30 minutes after injection of 6 fluoro-dopa. I would like to point attention to the fact that the amount of -- 6 fluoro-dopa that is actually produced is very negligible. The major metabolite is the 6 fluoro-dopamine as well as the 6 fluoro-dopa -- and at this time -- you still have very little amount of 6 fluoro-dopa have or vanillic acid being produced. So this would tend to support the idea that really you do not really need to account for the presence of the -- into image acquisitions results. Again, the same group studied the -- metabolites of fluoro-dopa and compared it with the L-dopa. Again, the predominant result is the thought that F-dopamine is the major metabolite that is formed at 30 minutes and also by -- minutes it's about 50 percent of it -- is still fluoro-dopamine. The significant findings though is that to gain is virtually little amount of -- fluoro-dopa being produced at both -- 30 minutes and by -- 17 minutes is virtually known detectible. So this group would like to come to the conclusion that no specific interference by -- fluoro-dopa would be minimal. And that fluoro-dopa or domain is the medial metabolite that you needed to worry about. The second opinion though, and there are lots of -- in the literature will say that there is an effect of -- fluoro-dopa and that the rate of -- or L-dopa. [This reporter and transcriptionist did not understand Dr. Laniyonu. Suggest reading the slides.] DR. LOVE: Are there any questions for Dr. Laniyonu? [No response.] DR. LOVE: Then we will go ahead to the next presentation. DR. BARRIO: I would like to ask a question just for clarification. In one of the slides I guess it's number 12, entitled, "Does the body handle F-DOPA and L-DOPA similarly?" You indicated there at the end you mentioned radionuclides, I am not so sure what you mean by this. You mean F-DOPA labeled with F-18 and labeled with caterion 11 or -- DR. LANIYONU: Yes, now, sometimes in some of the studies that I came across in the literature when you are doing L-DOPA labeling, sometimes [indiscernible] and other times they compare with the [indiscernible] label. And sometimes the profile that you get might be different depending on the -- compare with the [indiscernible] but the majority -- one actually use the tritium, but some of the centers that I came across use the -- DR. BROOKES: A couple of comments. Firstly, Carbon 11, DOPA and 18-F-DOPA do not behave quite the same way. It seems that -- I mean, you may begin to talk about this, but Carbon 11 seems more treatment sensitive than 18-F DOPA so the results are not entirely equivalent. The second point was the Fernau study where you said that they did not feel metabolites were a problem. But they underestimated the problem with methylation because they did not use the decarboxylase blocker and so it nearly all went to dopamine instead of hanging around to be methylated. So their results are not reminiscent or representative of the clinical situation at all where DOPA is routinely given with a decarboxylase blocker. DR. LANIYONU: Yes, I agree with your assessment and that's the - simply used, simply examine the metabolites in the absence of - or the count inhibitor, simply to set the profile. And when they did that they came out with the accomplish on that. Really you have virtually little amount of the - FDOP being produced. Then in a clinical setting where you actually want to minimize the peripheral - to fluoro-dopamine in any way you need to give - the question then is what then happens in this new environment which we are not looking at the plasma profile - now give it fluorodopa - which probably - for now was not a problem became a problem in an ideal clinical situation and then they accepted our question - have been able to get is that, yes, the fluorodopa - is increased simply because one might presume that because one of the two - blocked and then you - but [indiscernible] a higher plasma level of the fluorodopa and more can cause the blood brain barrier - how do you correct for this. And some of that - clinical - you simply do a correction by looking at - in the region of the brain in which you do not expect dopamine to be taken into the - simply correct - or do you use a count inhibitor. And if you do use the count inhibitor - in a different clinical suggestion - to assess whether it is for screening and issues of - and I believe Dr. Raman will be going into more details about this. DR. SANCHO: It is always tough when you have to do a presentation before or after lunch. I will try to be brief so that we may go to lunch. The purpose of my presentation will be to focus the clinical pharmacology and biopharmaceutical aspects of F-18 dopa to assist the dopaminergic systems. Then it will be broken down into only the literature source of the of the pharmacokinetic information I was able to find, both classical and the region of interest pharmacokinetics, some dosimetry information, and finally of course my preliminary assessments. Literature search criteria, again, we queried ICPs and the general community for suggestions on articles and we also did our own search with criteria for the search key words are here, time span was the last ten years, 1990 to approximately June of this year. The language is English the subjects are here but both genders. Very similar as before. I was able to find 93 articles that met this criteria out of which I selected 15 articles and they were broken up into the subcategory of pharmacokinetics, dosimetry, and methodology and other; seven for pharmacokinetics, two at dosimetry and three in methodology, and again, like in my previous presentation other just means it did not fit any other criteria specifically. In textbooks there is no number but I did use textbooks for background information. When I looked at each one of these articles I looked for specific issues and particularly again of protocol description, how well it was elaborated if there were questions beforehand the author was posting, and how well was the author able to come back to those questions or objectives. This is just the articles that I selected, 1993, and I will be talking or elaborating further on a couple of these articles, particularly Cummings from 1997 and Brown from 1998 under dosimetry. Again these are first authors and in chronological order. The first part of my presentation is going to -- well, you have seen this cartoon before, but the first part of the presentation is going to focus all the plasma or classical pharmacokinetics and what occurs there. I will follow with the influx issues and finally what happens at the vesicle or region of interest or compartmental region. I'm redundant there a bit too. For plasma pharmacokinetics or classical pharmacokinetics we were able to find peak information that had only carbon dopa extensively, and AAAD -- blocker, yet that data was not extensive in detail and we were not able to find the information the key information using F-18 FDOPA without enzyme blockers whatsoever in humans. Keep in mind if it is in humans and we were not able to find much information when using two enzyme blockers in this particular case COMPT and Carbidopa. One of the articles again was by Cummings form 1993. He provides the following information in which he compared -- he had approximately 18 subjects and they were all pretreated with Carbidopa and he compared the rate all of formation of the rate of plasma clearance of the of the rate all of the metabolites and the parent compound. You can see the rate of clearance here. For the metabolites the main metabolites like my predecessor already explained, or should, is 30MFD and out of the 18 subjects it included the study of 61 included in the data; two subjects were excluded, and the rate of formation of the main metabolites at the clearance is right here. The 31 patients mentioned here were not his patients; they were cohort patients from other studies in which he was able to compile the data process it. Now, there was no clear description either of Cummings' articles saying what stage of the disease these 31 patients were. So, when we compare those 31 patients with the 16-subject data he presents, you can see the rate of formation and rate of clearance relatively the same. He was able to identify again, back in 1993 two of the metabolites which he at the time, I shouldn't say identified, I should say he found, because he wasn't able to identify it at the time, he just called X and Y. Again, the rate information for these two or the metabolites and the clearance is shown here. Noticeably you can see the differences between the main metabolite or the metabolite there. That took care very quickly of the plasma pharmacokinetics and of plasma now I'm going to talk a bot on what information I was able to combine or compile in the influx portion of this; the product itself FDOPA is where there's a main metabolite. Here we have again Cummings but this time in 1997 these were -- it is a compilation but the graphics, some of the number seem to be chopped off. Here he compiled or aggregated several studies mainly healthy subjects, and the last study for Darwin in '96 are Parkinson's disease patients; there was no description of what the extent of the disease or illness was at the time. You could see the number of the subjects including each one of the studies is relatively low. And you can see the rate of influx of both the parent column palette of F-18 L-dopa into the straight or main cortex. In the cerebellum obviously is the reference tissue used in these studies. We would expect not to see any values there. There is a study from 1995 which does provide values much higher than the rest of the other studies. Cummings speculates that for that particular study there might have been technological problems or the way the author processed the data and there might have been a slight difference over the rest of the authors. Here we have Cummings again in 1997; he presents three studies in which this time is the main metabolite 30MFD. These patients again all subjects were all treated with carbidopa and that the number of subjects or patients presented here there was no description again of what extent the disease was in. Again cerebellum reference tissue no numbers or values. This is a very nice chart or cartoon. It was from an article from Brown from 1998. This is a compilation of images from 11 healthy subjects who were all pretreated with carbidopa. You can very clearly see the downslope curve of the plasma concentrations of both. Now, this is radioactivity this is including the apparent compound as well as a metabolite and the plasma. You can also see here the brain values over time and -- sorry for the graphics again -- and the multicipital cortex. Now, though the two, the much higher ones are that of areas of interest between the pertain to the car date [ph.]. Bercoli demonstrates. Summary of plasma pharmacokinetics in using a carbidopa with AAAD blocker. The cerebellum is the region for reference, as you noticed there was no values in that tissue; the rate of formation and plasma clearance for F-18 FDOPA with carbidopa for E treatment is known. There's a lot of literature. I just finished some of it. Rate of blood brain influx of both the parent compound and its metabolite into both the straight and the cortex of healthy subjects and patients with Parkinson's disease seems to be similar. Again, you need to take this with a grain of salt because this is a comparison across studies. Se do not know the details of the studies, e do not know how strongly to compare or how reliably we can compare these numbers. So I made two points there, the cross-over studies and there was no validation really of the comparison. Finally, I'm going to now talk about what happens with both FDOPA and its metabolites -- once it enters the brain itself or crosses the blood brain barrier. Essentially we can make some assumptions that it is instantaneously -- F-18 is instantaneously metabolized to transform to F-dopamine. We can also assume that it is instantaneously -- F-dopamine is instantaneously up taken by the vesicles, but from that point on we have three models that are debated and compared and contrasted with each other in relation to what is the fate of F-18 F-dopamine, what happens to it once it gets into the vesicles? Essentially I'm going to present them to you, and what are the plausible effects of these three proposed PK models are the acquisition of time legs. DR. BROOKES: Could I just interrupt? DR. SANCHO: Yes. DR. BROOKES: That first point there is totally wrong. It is not assuming that at all. That is the rate limiting step. DR. SANCHO: Correct. Correct, that wouldn't be. But once it gets into the blood - not into - across the blood brain barrier. DR. BROOKES: The uptake, the transport is faster than the conversion. DR. SANCHO: Correct. No, the uptake - the conversion is faster than the uptake. DR. BROOKES: No, no, wrong. DR. SANCHO: Ramesh? DR. BROOKES: Quite the opposite. DR. RAMAN: Hi, Ramesh Raman, FDA. I think perhaps what Dr. Sancho is trying to say here is what is within the center compartment. That is, once F-DOPA crosses the blood brain barrier it is immediately and rapidly converted to F-dopamine is what I think he is trying to say. DR. SANCHO: We have references and we could perhaps share those with you and we could talk or discuss it later. That is the purpose of this meeting and to help clear up these issues. This just proves the point that there is a lot of literature and different opinions and we are glad you pointed that out. Again, the three models - come back to that point. Just from a pharmacokinetics perspective, again Cummings from 1997. Again, I already talked, but this is for the pharmacokineticists out there, this is a generic one, if you want to look at it at way, one huge compartment model which eventually gets subcompartmentalized. I already spoke about the rate information declarance, we already talked about the influx rate, and now I am going to talk about what happens again, like I said, to Fdopamine once it get metab - The three models essentially are the model one, that F-dopamine once it is taken into the brain it is irreversibly trapped. Model two essentially says that F-dopamine and its metabolite -- acid -- occupy the same compartment and are cleared from the same compartment together vesicles that would be. And the model three essentially says F-dopamine is not cleared from the brain until it is metabolized. Once it is metabolized the assets it has been subsequently cleared. Each model has its own little assumptions and so forth and that is what we're going to talk about now. The comparison of the data to compare these three models or contrast these three models the data can be either from blood samples or it can be through images themselves. The plasma slope intercept was proposed by Jett in '81 the tissue slope intercept -- in 1985. Now, if all three models make a clear statement that carbidopa needs to be administered prior, approximately 60 minutes before that F-18-dopa is administered. Once we can do the comparison and contrast of these three models with the assumptions -- not the assumptions or the utility of carbidopa and either set of data we can make the following observations in the model one best estimates the rate of F-dopa within acquisitions of less than 60 minutes. Model two, you increase your -- sorry, decrease your precision because you're adding another compartment. So, therefore, the author or the authors who use model two and propose it say that this increase in imprecision, if you want to look at it that way, it can be compensated by doing data acquisitions of longer than 60 minutes. So in other words, the more data you have, the more time points you have, you can better fit the data for the two-compartment model. Model three essentially says that you still need even more data. You would need to go beyond 60 minutes up to approximately 180 minutes to be able to compensate for the increase in imprecision because of the three-compartment model now. All models come up with the same conclusion that the deaminated tracers or biproduct - metabolites, I should say, eliminated from the brain with approximately a 1-2 half-life. Again, when I say "region of interest" it is not the plasma, it is not the influx rates, but it is what occurs when both the parent compound and the metabolite into the - across blood brain barrier. All three models were affected by the start time, the length of data acquisition and it was - but all three models, really, we could not find a lot of articles that could address the point of what happens to these models and the data and how well the data is set. When you have more than one enzyme blocker holding down whatsoever, keep in mind that I said earlier that all three models and I stated that carbidopa which is one of the enzyme blockers had to be administered never mentioned or addressed the issue of what happens when you administer both COMT and AAAD inhibitor or you have none whatsoever. And also there was no elaboration on how did the C stage -- affect these models and the fitting of the data. For dosimetry purpose. The mass dose from the safety perspective that is the mass dose of F-18 F-dopa as compared to L-dopa would be considered negligible for safety purposes. It really has no relation to dosimetry. The assumptions made in the dosimetry paper, this is from Brown from '98; is that the bladder volume, the voiding sequence, the hardware utilized and the pre-hydration will all affect the bladder -- absorbed dose. Which I am leading to say that the bladder is going to be your critical organ. Any enzymatic manipulation of the metabolism will produce changes in the biodistribution of dosimetry to data like my predecessor already explained and like I have already presented some data. When you add an enzyme inhibitor at some stage it will modify the rates. There was no pediatric information encountered. This is information again from Brown '98. He used Pamphlet No. 14, the ICRP 60 and he gave this patient -- not patient -- scratch that -- subjects 100 milligrams of carbidopa pre-treatment. Again, the bladder is your critical organ. These are the remaining subsequent organs and the sequence based on their radiation of absorbed dose. There was no information on brain. That's why it's like that. Preliminary assessments. 18F-FDOPA can be considered to be a biomarker for the dopamine -- system in humans. It also appears that the start time and length -- that should be start time -- and length of data acquisition, the stage of the -- that co-administered enzyme inhibitors may affect the pharmacokinetic model used to best fit the acquired data. Again, re-emphasizing, although there's three proposed models, there's a lot of information in all three, there is no data to what happens to it when there is more than one enzyme inhibitor given and there is no information or there was no elaboration on what the effect of the disease date had on the data -- fit it to either of the three models. And there are some points or areas that need to be further clarified; from my perspective I have questions like -- again, something redundant -- the effect of the -- enzyme inhibitors or lack of, the pharmacokinetics, both the classical pharmacokinetic plasma and the region of interest with this particular product. The effect of the recirculating. In other words, what happens when you have -- to the kinetics when you have a long study and you have this recirculating, both parent compound and metabolite. The effect of the single combination enzyme inhibitor for -- at this point from the dosimetry perspective. And the best -- acquisition and processing protocols for the study the time length -- that's referring to the time length, shall we do a 60 minute, 90 minute, 180 minute, acquisition; and finally, the same issues all of these above related to pediatric information. I was unable to find any robust information on this. Are there any questions, besides what Dr. Brookes pointed out? Thank you. DR. BROOKES: Maybe I can make some comments. Because, I mean, the first thing is that when we look at fluorodopa uptake over 90 minutes in humans or patients, the reason we are going out to 90 minutes is because there is a relatively slow decarboxylation rate. It's taken up your K-1 as you showed was quite rapid, and fluoro-dopa equilibrates throughout the brain within 20 minutes or so, but then there is a subsequent slower decarboxylation rate in the K-3 which is the best discriminator that is considerably lower than your K-1. And I think Albert Gheda [ph.] and others, Kuibara [ph.] have actually done decarboxylate enzyme measurements -- activity -- and correlated it with K-3 and shown there is a linear correlation. So I think what we are measuring in that first 90 minutes is actually primarily the decarboxylation state. Though Jorge tells me it may also affect some intraneuronal transport once it is inside the brain. I think that's very difficult for one to sort out in practice. But we are assuming that when we measure its uptake over 30 to 90 minutes it is primarily decarboxylation we are looking at. So the question then is, how do these various blockers affect that? And what is the right model if I get you correct. The first thing to say is there is quite a literature on how COMT inhibitors affect that K-3 value, or the other approach which is more conventionally used in clinical studies is to just do a combination influx constant which by and large reflects K-3, but also contains a transport component, and that's the graphical approach that you mentioned by Gheda and by Patlac. And we are assuming that whether you measure one or the other you are looking primarily at decarboxylation. And the COMPT blocker, as far as we can see, as you would expect, does not appear to affect the rate of decarboxylation. What it does do, clearly, is affect the background signal because the level of methylated dopa and the amount of dopa that you have available to be taken up and stored and decarboxylated rises considerably if you block. And if you give 200 milligrams of antacapone or 100 milligrams of tulcapone in the days when people used it, you effectively increased the amount of dopa to about 60 minutes from 20 percent of the original starting level to 50 percent of the original starting level. So you don't completely block methylation, but you probably more than halve it by giving these CNTN inhibitors. But what you don't do, if you use what's called a plasma input function, is apparently affect its decarboxylation rate. And we have data, although it's not published, to suggest that you don't get a central effect from antacapone, but you do from tulcapone, but that is not immediately germane to this. Because in order to see that you have to go our for three or four hours. As far as the correct modeling is concerned, it rather depends on what you want to know because if you want to work out absolute truth and follow every possible metabolite and how quickly it takes for all these metabolites to diffuse in and out of the brain and be excreted, then you need to go out for four or five hours and fit multi exponential codes to all this data with cleaned up plasma codes, et cetera, et cetera. But as with the flow situation, if you then try to measure K-3, K-7, K-9, you get huge variances, huge standard errors on individual rate constants because you're trying to fit to one input and one uptake or washout curve, seven or eight different rate constants. If you use the simple graphical approach which gives you just the influx constant, this combination constant of everything, you, of course, only really get a handle on decarboxylation, but on the other hand you can measure that very much more accurately and the standard error on that is probably less than 10 percent in most units. So it's back in a sense to the flow situation. If you want to look at different clinical groups of subjects, or if it want to follow their disease progression, then the most robust parameter is the simple graphical influx constant. If you are Albert Geller [ph.] and you want to know what is going on in the brain to all these metabolites, then clearly you have to do the much more rigorous and time consuming fitting exercise. I think there is a no single answer to your question. DR. BARRIO: -- and you have to correct me if I'm wrong, but in a clinical environment, of course, the graphical approach you seem to set -- ideal situation if you have to quantitate because you don't need any arterial sampling or you don't need anything that's very simple approach to get an assessment of the -- primary system. And, of course, answering your question here in regards to the dosimetry, that's a simple basis -- I mean, all the questions is, certainly using a combination of inhibitors or changing -- I mean, the inhibitor, certainly you would change the dosimetry. But, on the other hand, you will need to use less fluoro-dopa because now you have more available. Therefore, you can change your dose and therefore you would change your exposure to the clinical -- based on -- and that's not a problem -- essentially needs to be determined based on the condition of your experiment and nothing else really. DR. SANCHO: Correct. But, again, it goes to what is the best approach. DR. BARRIO: Right. DR. SANCHO: Like Dr. Brookes just said, what is the question. If we know what the question is, we know which enzyme blockers to use and we still would need information for the dosimetry -- or whatever. Again, I'm just posing the question that there were three models out there. DR. BARRIO: No, that's fine, that's fine. I think it's an appropriate question. The only thing I'm saying is that if different institutions or different hospitals or clinical settings use different procedures, then you will address the question based on -- and therefore you have different requirements, of course, nothing else. But there is also the same really. I mean, with some variation as described by Dr. Brookes. DR. BROOKES: I think there have now been multi-centered fluoro-dopa trials set up in the last two years people have been looking at disease progression in the presence of an agonist or in the presence of -- glytheonate blocker. And the similarity between influx constants across all the various centers is quite surprising and gratifying given that we're all using slightly different protocols to measure it, but really because the graphical approach linearizes the uptake, it's just a question of where you put your points along the straight line in practice. DR. SANCHO: I was going to say, there's a Dutch group that is putting out a paper or just put out a paper in which they validate the graphical approach and they provide us information on that. That's one of the articles I reviewed. DR. LOVE: Thank you. I think that this is very important, the kinds of issues that we are discussing, but they also relate in part to some issues that Dr. Raman wants to bring up. So maybe we could come back to this, this afternoon after clinical discussion we can look at the different assumptions and move forward then. Okay. Thank you. One hour? MS. AXELRAD: Let's take 45 minutes if we can do it. Let's try for 1:15. [Whereupon, at 12:30 p.m., the meeting was recessed to be reconvened this same day at 1:15 p.m.] A F T E R N O O N S E S S I O N [1:25 p.m.] MS. AXELRAD: Okay. Where did he go? There he is. DR. RAMAN: Good afternoon, everybody. Welcome back after the lunch. For those of you who do not know me, my name is Ramesh Raman. I am the Medical Officer with the Division of Medical Imaging and Radiopharmaceutical Products. The purpose of today's meeting here is to determine whether there is evidence in terms of safety and efficacy of 18F-DOPA. At this point let me pause for a second and here may be a very slight discrepancy in the terminologies, but overall what we are referring to would be perhaps the same product, F-DOPA is what I would like to call it as, and FDA approved dopamine. Nonetheless, to see if we have enough evidence based on the -- that's out there to move forward and to use F-DOPA perhaps as a PET imaging agent. The methods and the basis on which this is moving forward is similar to the already other approved PET agents. At this time I would like to acknowledge and thank several of my team members and also the NIH Library for helping me with an extensive search that was very, very helpful. I also want to thank the ICP for their time and the effort put in interaction and being here today. Also want to specifically thank all the authors of the reference articles that I have looked at. And if I don't recognize somebody here in the audience, it's just an oversight, but I'm acknowledging -- as well. I am kind of arbitrarily dividing my talk into three different sections. If one has to understand what F-DOPA is and what it does, I think it s very important to know what -- neuroanatomy is. My colleagues, both Dr. Laniyonu and Dr. Sancho have laid the foundation for me in saying that FDOPA can be used as a marker for the dopaminergic system. I will be revisiting this site from the clinical perspective. The second part of the talk will include having encountered this enormous amount of literature. There was abundant information that I gathered. I have summarized those as observations and along the -- identified some issues. And then I moved forward into the third part of my talk which would be how and in what way could I validate these issues and findings, and that is where the literature search and the results of the search will be discussed. I will finally be summarizing my talk in terms of how FDOPA stands as of today in terms of whether it can be used effectively, whether it is safe, and what were the findings of the search. Additionally, I will be concluding my remarks with perhaps what the final recommendations would be and hopefully would be providing a direction as to where we can go from here. With that, let me start off with the neuroanatomy part. When the issue of the dopaminergic system and the neuroanatomy is emphasized further for the fact that most of the studies have been done in patients with -- and particularly Parkinson's disease. And, in fact, this was one of the areas that the ICP asked us to look into. If one has to do that, it's a very complicated system. A lot of anatomical terms. But it is important to recognize that you can clearly categorize these into two broad groups, an anatomical group which consists of the basal ganglia, the corpus striatum and the lenticular nucleus in which you also have the substantial nigra and when one talks about it from the functional point of view, it is important to recognize the word "striatum" which consists of the caudate nucleus and the putamen. From an anatomical perspective these again can be subcategorized into the long-length systems, intermediate-length, and the ultrashort. This is based on the cell of the origin to where the neuroanatomy body is located to where it projects eventually. Of importance here is the nigrostriatal system which indicates that the subbody is located in a substantial nigra and it projects to this items which -- combination of the correct -- This is important because, again, most of the studies that have been done has involved studies of patients with Parkinson's disease and it is the belief that it is this system that is affected the most. The importance of understanding the other systems as well is that we also know that as the disease progresses, including diseases such as Parkinson's disease, as the severity progresses, these other systems are also perfected. Having known the anatomy and where to start, one can also correlate this with the function that these things do. Again, of implementation is the nigrostriatal system which have -- to the dorsal striatum and it is involved with movement initiation and execution. Likewise, these other regions have their own functions and it is again recognized that as disease progresses in many of these disorders, they do manifest with symptoms that can be localizable from a neurological point of view to these systems. This entire system is a very complicated system. One needs to understand the actual function and therefore one needs to understand the neurotransmitters that are involved. One can talk about this in the form of the circuitry. Here I am presenting a substantia nigra in the bottom of the slide and the heart box here as the striatum. The cortex is represented here. The blackened lines represent areas that are stimulated, the open lines are areas that are indicated. Let it suffice to say that it is a very complicated circuitry and it exists in two forms. The first one is being called the direct pathway where you have projection from this site from the substantia nigra to the striatum on the globus thalamus medialis which stimulates the thalamus which eventually stimulates the cortex. This is referred to as the direct pathway. and activation of the direct pathway leads to cortical stimulation. The opposite of which is the indirect pathway where it bypasses the medialis but goes into the subatomic nucleus which inhibits the thalamus and eventually inhibits the cortex. This is referred to as the indirect system and therefore activation of the indirect system causes cortical inhibition. It is important to recognize and understand this because when you see patients with movement disorders, which is a spectrum of diseases, once you would be able to perhaps focus and understand which pathway would be affected. This entire sub -- in neuroanatomy the power source or the gasoline for this gives the neurotransmitter dopamine. It is very intriguing, it is very fascinating that of the entire central nervous system neurotransmitters dopamine is only less than 1 percent, but in combination with serotonin. And what is further intriguing is that more than 80 percent of it rests in the basal ganglia. Therefore this is a system with an enormous amount of reserve. Also it is further intriguing that less than 10 percent of the striatal neurons are dopaminergic. I am revisiting the sites of the metabolism here just to further re-establish that the metabolism of FDOPA, if you have to say there is molecular basis for the use of FDOPA, it's metabolism is similar to LDOPA. The cartoon here has been seen earlier by my peers and colleagues here. Just to refocus that you have a blood brain barrier, we have two sets of enzymes, both are located peripherally and centrally. The FDOP metabolism is very qualitatively similar to LDOPA, similar enzymes both in the periphery and in the center. What is important to recognize is that in the absence of any of these inhibitors FDOPA is connoted by the peripheral AAAD enzyme, to F-dopamine which has known cardiovascular effects. If any, it causes hypertension and the like. But the important issue is, given the doses at which FDOPA is used as an imaging agent, it is minuscule compared to what we usually use LDOPA as a -- but nonetheless a safety important aspect. The other aspect is simultaneously the peripheral FDOPA is converted to its intermediate 3OMFDOPA by the peripheral COMPT insight. It is important to recognize that this intermediate can also cause the blood brain barrier and in fact, the initial imaging artifacts and the nonspecific activity that is seen in the striatum and in the other regions of the cortex of the brain is due to this intermediate. The so-called parrot FDOPA that is able to eventually cross the blood brain barrier is converted by the central AAAD enzyme to Fdopamine which is -- the FDA. It is this Fdopamine that is stored in the [indiscernible] vesicles and it is this that is imaged by PET. Therefore, having mentioned about the AAAD enzyme and if we are concerned from a safety perspective of its conversion to Fdopamine which has cardiovascular effects, the use of the AAD enzyme which some people refer to as carbidopa, is shown in this slide here. It is important to recognize that when one uses the AAD inhibitor it affects both efficacy and safety. From a safety perspective it is -- conversion to Fdopamine is prevented and therefore even the worrisome cardiovascular effects are curtailed. Additionally, also, there are studies to indicate that perhaps the amount of radiation to the bladder is diminished with the use of the AAD inhibitor. The other aspect also is because of its inhibition and prevention of its conversion of conversion of FDOPA to dopamine there is relatively an increase -- of FDOPA that can potentially cause a blood brain barrier. What I forgot to mention earlier which I should have, is the thickness of these arrows had a meaning, but they are not to scale. When one uses the AAAD inhibitor there is also a relatively overdrive shift from the peripheral COMT enzyme enough that you have more formation of the intermediate which again crosses the blood brain barrier and causes these initial imaging artifacts that we talked about. The issues of -- you know, whether to use the COMT inhibitor or have inhibitors are not -- were discussed earlier by Dr. Laniyonu. But, nonetheless, it is suffice to say that with the use of the AAAD inhibitor we have effects both on the safety and the efficacy. Having talked about the AAAD inhibitor it is just only reasonable to also talk about the COMT inhibitor. The primary focus of the COMT inhibitor is has more efficacy-related effects. When one blocks both the enzymes with the AAAD inhibitor as well as the COMT inhibitor, from the efficacy perspective there is clearly a much greater availability of the parent component of FDOPA to cross across the blood brain barrier enough that more of it can be converted to FDA and eventually more of it can sit in the present vesicles to be imaged. Additionally, perhaps the greatest benefit of the COMT inhibitor is that this overdrive that occurred in the absence of the COMT inhibitor is now shut down and therefore there is relatively decreased intermediate of -- FDOPA and therefore less of it crosses the blood brain barrier and causes less artifactual effects. Additionally there are information and literature that with the use of the -- in combination the use of the AAAD inhibitor and the COMT inhibitor and reflecting the effects of the COMT inhibitor there is approximately a 22 percent increase in specific striatal uptake. And the non-specific uptake of up to 44 percent that is caused by the theos decreased. Additionally also there are studies to indicate that the dosage of the FDOPA itself that one can -- or the availability of the parent FDOPA within the central nervous system can be increased to the tune of three times. Therefore, having talked about the anatomy, the neurology, and having understood, perhaps the metabolism of FDOPA which is similar to LDOPA, one can reasonably come to the conclusion and set the stage that FDOPA -- what is the molecular basis, can it -- can it be used as a dopaminergic system? Yes. But it is important to recognize here that FDOPA itself acts as a -- because it needs to be converted to FDA which is Fluorodopamine and it is the Fluorodopamine that is imaged. And secondly, given that the Parkinson's disease and the movement disorders are at the focus of interest and the dopamine is the transmitter, therefore, diseases in which the substantia nigra is affected can be used and we can use FDOPA as an endemic measure to assess this function. Also, despite the quantitative differences between LDOPA and FDOPA given their qualitative similarities one could reasonably consider FDOPA as a biochemical marker. There are additional studies which does rationalize and validate this in terms of histopathological correlations where the dopamine cell cones and its levels in degenerative neurological disorders have formulated with the FDA localization. It is also very important to recognize that FDOPA not only localizes itself within the striatum on dopaminergic system, it also localizes in the extra striatal and the non-dopaminergic sites. Having talked about and established that we could use FDOPA to review the dopaminergic system and at the focus of interest is if the disorders, the movement disorders which includes Parkinsonism and Parkinson's disease, this light here gives you all the syndromes that can be clinically classified within this disease and we know the presumed location of the lesions. Of highlight here on this slide is that there are several areas of lesions that can cause -- syndromes. And within the Parkinsonism, the incidence of the disease can be broken down into, in this slide here, it is again of importance to note that over 70 percent of Parkinsonism consists of idiopathic Parkinson's Disease and obviously the other disorders are a minority, but it is these diseases that you should be able to distinguish from Idiopathic Parkinson's Disease. And if the focus of the topic today -- the discussion is movement disorders of Parkinson's, therefore, the question is, can FDOPA do it. The overwhelming information in the literature that I talked about reviewed and came to some observations are right here in this slide. I am going to summarize this because this also formed the basis on the issues that have been identified. The majority of the studies as I indicated earlier were related to Parkinson's disease which included patients who are asymptomatic early and late. And this was studied by the 2D which is for two dimensional using the MTGA methodology which is meantime graph analysis methods. And there were differing patterns of the putamenal and the striatal that were looked at and published. Also, the more recent studies in the literature point to the use of the 3D which is the three dimensional camera with the statistical parametric mapping particularly in patients early in their disease. Also, the literature and the observations indicated that there were several studies which attempted to differentiate idiopathic Parkinson's disease from the others which I talked about and most of these studies used the 2D methodology. Finally, the more recent studies have used the COMP inhibitor as a premedication in addition to the AA inhibitor which have been used all along. I am summarizing the findings of my observations in a nutshell here. Given the focus that it is Parkinson's disease which has been studied the most and forms the largest amongst the movement disorders. We can argue what the two methods do in terms for the same disease. It is important to note that all patients who are studied either with the older methodology -- by "older" I mean, the 2D methodology or the newer one had similar rating schemes for staging. Here of importance is when one looks at early Parkinson's disease and when they were studied using the 2D MTG methodology, the FDOPA decreasing uptake was unilateral and opposite the site that was affected. When the same patients were studied with similar rating using the 3D, it localized bilaterally. Also, as the disease progressed from early to late, there was a bilateral localization using the 2D methodology whereas not only was it bilaterally striatal when you used the 3D, it was also extra striatal. The flip side of this is when one uses the 3D in the statistical parametric mapping methodology not only does it target the dopaminergic site, it also targets the other monaminergic and nonaminergic sites. Also, if there are certain differential patterns within the uptake using the 2D methodology which the decrease in the posterior containment is one of the observations that were made. Here I found in the literature that it is not only -- Parkinson's disease where you have a decrease -- the differential diagnosis is listed in this slide. It is quite extensive and some of these diseases can be easily mistaken for Parkinson's. Therefore, a decrease for -- uptake, per se, is not -- for early, early Parkinson's disease. Also, when one talks about the striatal uptake itself which I mentioned includes the containment and the -- it is not only seen in early or late idiopathic Parkinson's disease, it is also seen in this list as mentioned here. For ease I have highlighted the blue disorders to say that these, if any, are the more common ones that one can mistake for Parkinson's disease. Not only does FDOPA is decreases in the striatum, there are also diseases where it is increased -- activity. Of importance is the disorder of attention deficit hyperactivity disorder. This is becoming a newer disease that is being studied more in the recent past. Also, there are other agents that are being used to study this disorder. And important in this lies in the fact that this is a disease that occurs in the pediatric population and it is of concern. Additionally, within the 2D, FDG methodology if there is a correlation that one can come up with it is projected here in this slide. It is important to note that when one looked at qualitatively was that when the decrease was to the tune of 35 to 40 percent decrease in -- uptake. This correlated with 80 to 90 percent loss of striatal dopamine. In other words, when one -- and also it correlated when patients became symptomatic. In other words, the flip side of this is when patients were asymptomatic, the decrease was much lesser and it could be assessed only by quantitating methodology and not qualitatively. And as the disease progresses, where there is a greater decrease in FDOPA, this correlated with approximately a 60 to 80 percent loss. This slide here summarizes my observations and the issues so identified. So I was asking a question for myself. That's the presence or absence of the COMP and the AAAD inhibitor. Does it affect image and image interpretations? Does it have an effect on safety? What is the meaning of these different methods and the techniques that were used and the findings in terms of the 2D versus the 3D. What is the meaning of the extent of -- when you use the 2D it locates itself unilaterally, but when you use the 3D it locates bilaterally. What is the meaning of when you use the 3D technology it localizes itself not only to the striatum but also the extrastraital regions. Additional, also, within the 2D methodology you had clearly studies which looked at it quantitatively and qualitatively and the clinical meaning of this was something I was seeking. Also, as I mentioned earlier, there were not -- there were several diseases which show a similar pattern of decreasing uptake. In order to rationalize this and to come up to some sort of a conclusion, I asked these questions: How then can we use FDOPA and label it? And what would the clinical utility be? In order to do this I also want to revisit the site of, again, this document. Because if we are going to label it, it has to meet a claim. Dr. Love earlier this morning already spoke about these indications and it was my clinical intuitiveness that perhaps if any FDOPA will qualify as a biochemical marker. Also, based on the information from the studies of the use of FDOPA in perhaps preclinical Parkinson's and asymptomatic patients I was asking the question, is there a role of FDOPA to be used as a screening tool? I just want to re-emphasize here that typically when one considers a drug device or an agent as a screening tool, it is important that one needs to have very strong identified preclinical stage. This, I think, is lacking when you are talking about preclinical Parkinson's because they're asymptomatic. So you cannot identify the preclinical stage. Additionally, the screening device or the tool should clearly have a large impact on the mortality and the morbidity. The agents that have been used as screening tools have usually studied patients in large numbers and it should have effect on the treatment and the prognosis. And, finally, I have mentioned this already, so the final question was, would there be any clinical utility with the use of FDOPA? And to accomplish this, I conducted the literature search. This was the criteria that was used. It is justified to say that during the initial run we -- for ten years in English in humans and this went from -- 1999 to 1988, but the entire search lists, I added on several others to this which includes books of reference and also others published outside this frame. Of note here is that there were 37 articles that were identified by ICP; of which 20 I considered were to be equal to clinical trials and these were also on this list. From a different perspective this entire pool of in could be subcategorized into two, those studies that involved humans. And, again, a majority of them within -- of category involved patients with neurological disorders; however, the bulk of the articles in the literature fall into the first bulleted item. And these were articles of the nature of review and hypothesis generating that talked about the -- methodology and the like. Having identified those articles I reapplied the criteria that has been sent by our guidance document. Dr. Love already talked about this, this morning. I am not going to dwell on this too long except to indicate that I reapplied these criteria to see how many articles would fit into the search that I had. Of importance was also the issue of standard of truth. And some comments were made this morning particularly with Dr. Herscovitch in terms of the clinical diagnosis. All the articles that I have looked at, the clinical diagnosis was considered as the gold standard. Well, as of today, autopsy is the gold statement. But next to autopsy what we have today is the clinical diagnosis and that is one made by the movement disorder specialist. So all the articles that I looked at they used clinical scores and this is what they were comparing it to. The search findings are summarized in the statement here. I totally identified 472 reviewable materials. These includes not only the articles that were searched, it also included the articles, as I mentioned, the review articles and also articles outside the time frame. It also included the books. And when I applied the guidance criteria to the human studies and the articles of the clinical nature type, clinical trial type, I identified 21 articles that I could potentially used to see if there was any efficacy in it. The reference list has been made and it was made available and I would be happy to share what these articles are. The important point of it was that there was not a single article from these 21 or the others that met all the criteria. I am summarizing the findings of the 21 articles here in terms of strengths and weaknesses. All these articles, the 21 articles that I identified as potentially valuable for efficacy had very strong statistical methods for imaging. Also, the objectives are very clearly stated. The weaknesses, however, were that all the patients who were studied, when there were patients they were with no disease and when they were -- or they were normal volunteers or normal subjects. A comment was made earlier this morning on the strength and weaknesses of studying patients with no disease. It is very important that if you are going to study -- study of patients with no diseases, if the images are not blinded, it has no value. None of these articles had blinded reads. The number of articles that were read prospectively were a minority. There were only four where there was clearly a prospective design. There was lack of statistical robustness from clinical endpoints. Also there were clearly methodological differences between these articles. And of the majority of the articles as I mentioned included the 2D MTGA methodology. But then when you went to the region of analysis the division of interests that were used were of different characters and obviously they were subjective in nature. Of these 21 articles I finally further narrowed it down to three articles in terms of articles that were literally strong. And I am using the word "literally strong" they emphasized that one of these articles had all patients accounted for. There were the three prospective design studies and there were two where there were similar endpoints of the study. These are identified here by these numbers. And I would be, again, happy to -- the reference list gives you the numbers. The extent of the subject that were studied in terms of the numbers shows here on the slide. On the three studies used the AAD inhibitor as premedication. One study used both the AAAD inhibitor and the COMP inhibitor. The methodology and the technology that was used for one study was the 3D with the statistic of parametric mapping. But the same study also analyzed the data both for region of interest and using the SPM in the region of interest. There was two articles where the clinical endpoints were similar. And these were longitudinal studies where patients were followed for a period of years. From a different angle, the same articles -- the three articles can be talked about in terms of what were their advantages and what were their disadvantages? The advantages were, as I mentioned earlier, they were all prospectively studied, the objectives are very clearly stated and the statistics for imaging was very strong and there was full accounting of all patients. And the added advantage was two of the three used similar endpoints in terms of technology and methodology and patients were studied with similar disease criteria and two of the three had longitudinal follow-up. The disadvantages were, as mentioned earlier, none of them had blinded reads, there was a lack of robust statistical analysis for clinical outcome. And two of the three articles that studied similarly in terms of methodology technology also had -- follow-up were from the same author. And as mentioned earlier, all patients were with no diagnosis. And because, as I mentioned earlier, the no diagnosis is again is clearly a negative in terms of if the reads were not blinded. And one article of these three included data from another study that was acquired previously. To summarize, in terms of the efficacy at this point there was a lot of information and it was clearly information was helpful in establishing that FDOPA marks the dopaminergic system and may be used a biomarker. However, there was no single study that validated the other findings that I talked about earlier. Efficacy is one part of the talk, the second part of the talk is, is it safe to use FDOPA? Again, there was no single article that either monitored safety or analyzed data from a safety perspective. But, however, the safety of FDOPA may be inferred, as discussed earlier by my colleagues based on its qualitative similarities to LDOPA. We know we have been using LDOPA since 1964 as a therapeutic drug for Parkinson's disease. The doses at which FDOPA, the imaging agent is being used, is a minuscule dose and a fraction of it. Also, the known effects of LDOPA occurs only after prolonged use and also a dose of several times that is usually used as a dose for FDOPA. The other important aspect is, if one considers premedication, either with the AAAD inhibitor or with the COMT, there are no safety data that was really looked at as a single dose. Again, it is not a very big important safety issue because this would be a single dose, if any. Safety also includes radiation of dosimetry, this has been talked about by Dr. Sancho earlier. Of important is the fluorinated compound in particular if there is a push towards studying pediatric patients, it is important to make attributions to this. I'm deferring further comments on this because 18FDG was reviewed, sodium fluoride has already been approved. Therefore, my preliminary assessment at this time is FDOP is a functional biomarker of the dopaminergic and non-dopaminergic systems and may be used to measure alterations. However, the clinical utility this is unclear at this time. Perhaps controlled clinical trials in relevant setting because these have not been identified to perhaps be the focus. At this time I think the direction we need to move forward I am summarizing here in this slide. I will be very happy to review any new information. Also, it is important to focus and to resolve what these methodological and technological differences are and how can we apply this effectively in the future is an important part. And, finally, how, in what way will FDOPA be of added clinical benefit either in the diagnosis of patients, either as a screening tool, monitoring their disease or how to manage them is the question that needs to be answered. Thank you. DR. LOVE: Thank you, Dr. Raman. Are there any questions? DR. BARRIO: Thank you for the description, Dr. Raman. You raised an important number of questions. I don't think for what I understand you are concerned about the issues about where the dopaminergic system is being monitored by Fluorodopa in terms of the assessment of the integrity or the biochemical function; isn't that right? DR. RAMAN: I think -- I'm sorry -- DR. BARRIO: I don't think you have concerns from what I understand in your presentation about the fact that Fluorodopa using your words is a functional biomarker indicator of alterations into the dopaminergic system. DR. RAMAN: Yes. DR. BARRIO: You are? We are on the same page on that? DR. RAMAN: Yes. DR. BARRIO: Let me ask you about the extra dopaminergic areas. What kind of information you have in that area that make you state that this may confuse the information you may obtain? DR. RAMAN: It is not merely an issue of confusion, it is more a process of understanding the pathology and the diseases further. I could perhaps take the liberty of making reference to Dr. Brookes's paper along with Dr. Rachid that came out in 1994 that used the 3D and the statistical parametric mapping and that was one of the studies that I felt were stronger and clearly it has shown that -- for the same sets of patients, these are patients with early Parkinson's disease. The way it was studied it localized itself, not only to the striatal areas, but also to the extra striatal areas. DR. BARRIO: What do you mean by that? What areas are you talking about? DR. RAMAN: Well, the areas where the other -- the dopaminergic projections are, that is to the mid brain, the frontal lobes, the regions of the ponds and the like DR. BARRIO: David, probably you may like to answer that question from the clinical perspective. DR. BROOKES: Well, I think you've raised a number of very interesting points. If we talk first about the technology because you are querying why 3D and 2D were giving different answers. And it's a very simple reason. The 3D has much higher resolution and much higher sensitivity so that when you measure in 2D you are getting contamination from surrounding areas and so it is much harder to discriminate, say, the core date in early Parkinson's disease from normal or the apparently unaffected side in early Parkinson's disease from normal because there is a lot of overspill from signal and surrounding tissues. So when you use the 3D approach this is dramatically reduced and it becomes much easier to separate apparently unaffected sides actually being involved. The other thing is the SPM side of things. This is the statistical package that goes looking through the whole brain volume for where changes have occurred in your patients compared to a normal database. And the reason it is so much more sensitive is because it looks not just at the whole butamben and draws a ring around it, but it goes hunting all the picture elements inside the butamben and it says, if I look where the disease is worst inside the butamben, I can discriminate it even more sensitively than if I just draw a ring around the whole thing. So that is why there is this apparent difference in striatal findings between the two techniques. If you go back and smooth the data down to how it would have been the 2D approach, you get the same answer; there's an overlap again. As for the extrastriatal areas, I mean, dopamine fibers apart -- MS. AXELRAD: May I just ask a question? DR. BROOKES: Sure. MS. AXELRAD: Just from my very basic understanding of this, well, doesn't that mean that relying on studies that used the 2D methodology isn't very helpful because it turns out that when you smooth -- I mean, that they may not have been showing what you thought they were going to show and therefore that whole body of literature that was based on that type of analysis may be not very helpful to us where we are looking to the literature to make a finding here? DR. BROOKES: Well, I think some of the findings still stand. For instance they showed there was a very clear correlation across a group of patients with how severely rigid or slow they were and how much DOPA was taken up by the butamben and that still stands. It's just we share it much more sensitively now. We were able to pick up preclinical disease in relatives in families where there was a clear generation-to-generation transmission even with the 2D approach. It is now much easier to pick it up with the 3D. So I don't think you should throw out the whole lot. But I think one has to be aware that the sensitivity on the resolution was less. And so that we are capable now of detecting things much more sensitively with the latest scanning approaches. I forgot what I was going to say now. The question of what the extrastriatal signal is, I mean, you flagged up the fact that it may go in to other nerves, it's surprising how little that seems to happen. There is signal in the brain stem and it's conceivable it could be from the seratonagic nerves, but on the other hand we don't see a hot spot where the media raphe is which is where all of the seratonagic bodies are, whereas if we use other tracers which mark seratonagic nerves, that stands out. We don't see signals from the thalamus where there is a lot of noragenergic terminals, whereas we do say with some of these new cocaine tracers. So my impression is that probably there isn't that much non-dopaminergic signal in practice. But you are picking up areas, as you say, like the cingulate in the frontal areas where there is a greater concentration of nerves than state in the occipital area. This new technology now pulls that out; that's right. From the point of view of correlating it with pathology, I mean, you made a rather contentious statement, but it was the gold standard. In fact, there are now five different genes causing Parkinson's disease and it's not at all clear whether the louibody in fact is the Parkinson's pathology, though everyone tends to accept it as such. But there are a lot of things which call themselves Parkinson's disease and they problem don't have all the same calls or necessarily the same pathology. But I agree with you, it's the only practical way to work. If you -- there is very little pathological validation of Fluorodopa. I think only a handful of patients have come to pathology and Barry Snowe from Vancouver have got the biggest series. There is this correlation between cell counts and dopa uptake. Dopamine levels, as you quite rightly said, are much lower. But dopa is not measuring dopamine. What dopa is measuring the activity of dopa decarboxylase which is probably relative up regulated. So it's telling you about Dopamine terminal function rather than an absolute level of Dopamine itself. I don't know if I can go on and talk about clinical indications if you want now, or did you want to raise some other -- I mean, I think one of the problems is you're criticizing us for not having done the right trials. But then we never did these studies as designed to be trials to present to you. So in a sense it's a little unfair to criticize on that basis. MS. AXELRAD: I don't think we're really criticizing you for not having done the right trials. But we are trying to -- we have been asked to make a finding based on the literature and the trials as they're reported in the literature and we are simply trying to point out the difficulties we are having based on the literature that we reviewed in making the findings that we need to make. And I think the first thing we need to discuss is whether, you know, as we did with 0 15 water, did we look at the right studies, are there other studies that we should look at, are there things in the students that we looked at where we may have come away with an assumption that is questionable. And then if at the end of the day we sort of all are in agreement on what the studies that are out there and what they show, then the question is, what should we do next or what should be done next in terms of performing trials that would show what needs to be shown. But I think we should talk about the indication maybe before we get into that. DR. BROOKES: I quite agree. I think as someone who works as a clinician as well a researcher, the value I find of having techniques like Fluorodopa PET available where you have diagnostic difficulties, one of the major areas of difficulty is when you have a patient who is severely tremulous and you are really not sure whether this is a benign tremor or whether it's Parkinson's disease. I think that's probably one of the major sources of confusion. And those are one group of patients. We certainly find it valuable to have this technique available. Now, you can argue how rigor it is to say if the scan is abnormal you've got Parkinson's and if it's normal you've got tremor. I guess it's probably not that simple and this is the sort of trial that needs to be set up where we take a group of these patients, we randomize them, and clinicians have to say whether they think they do or don't have Parkinson's disease and then we have to do the scans, and then eventually we have to find out what they had. But that could take year. You know, patients with Parkinson's disease live for up to 30 years. So these are not easy trials to do in practice and it's not clear in the absence of a biological marker that we can take along side the scanning, how we get a quick answer to your question, it's still mainly based on clinical conception as to what is going on here. A few of these patients do have genetic markers, but they are the exception. Only probably less than 1 percent of them in practice have genetic markers. Other areas where scanning can be useful are in the same way that in Alzheimer's you find a patient who perhaps is having some memory difficulties and you are not sure whether it might be just they're a bit tired, or depressed, or they've really got Alzheimer's. With Parkinson's if you get someone who is slowing up a big and appears to be feeling a bit and appears to be feeling a bit stiffer, particularly on one side, a lot of people will say, oh, he's just getting old; but, you know, there may be a very real problem there. Now, you can give these people DOPA and just see what happens, but there is a big placebo effect when you do that kind of thing. And if you get it wrong, you are giving someone DOPA for a long period and there are patients who have had simple blind tremors and have been on clinical doses of lethe dopa for several years as a result of misdiagnoses and it's not clear whether that is harmful or not at the moment. There's a lot of debate. So that's another issue, and the third issue that was raised was these different forms of Parkinsonism. Can Fluorodopa PET differentiate them? Well, the answer is, it can with some and not with others. For instance the main differential problem is between this condition multiple systematrophy in Parkinson's and there I think Fluorodopa PET does not work well. It really cannot discriminate the two with any real sensitivity. But some of these other more oddball ones like progressive supernuclear palsy, cortical basal degeneration, it discriminates about 80 to 90 percent of those cases because there is a very uniform knockout of Dopamine fibers in those conditions. Whereas in Parkinson's the core date function is well preserved compared to the butamben. In there you do get a reasonable discriminator. So it is not ideal and usually we combine it with an FDG PET study, if that's the question we're asking. But I think there are definite values to this scan. It is not something I do routinely on all my patients, but there is a group of patients, and as you pointed out, there is something like a 20 percent error inaccuracy in diagnosing Parkinson's disease. About 20 percent of the patients turn out to have pathology other than louibodies and in that sort of 15 to 20 percent a number of these patients actually benefit from having a scan. Then there is the whole other indication which is another one which is whether it's useful for following disease progression and obviously in a research environment that is true. There are already two major multicenter European, U.S. studies running doing just that. And one was just reported with Pogalied, the -- was an SP study. Now, again, these are all patients where we think we know what's going on, so they haven't been, in a sense, I mean, they're not the population you really want trial data on. But it will give you a whole load more data on just how effectively this technique can be used to monitor disease progression in practice. DR. RAMAN: Thank you for the comments. That was a lot of very good comments. I agree with a lot of what you said. Let me take on thing at a time. In terms of the studies that involved when tremor is the concern and if one needs to distinguish essential tremor from the Parkinsonian tremor I have reviewed the literature where the studies were done. In terms of the essential tremor, I mean, there is this thing as the clinical diagnosis. Essential tremor usually bilateral, the coarseness of the tremor, or the fineness of the tremor, the symmetry of the tremor, and then there is this thing as the response to treatment. Whether it applies to Parkinson's patients or patients with essential tremor. If you have someone who has essential tremor who is not Parkinsonism, I think there would be no response to levodopa. And, in fact, the tremor may get worse, and that may be used as a good clinical working role to make that distinction. Additionally, also, if you have patients with Parkinson's disease and if they are treated with levodopa, per se, the response to the levodopa itself, there is an enormous amount of -- information onto how it correlates with the diagnosis in terms of Parkinson's disease. When you do go to the progressive supernuclear palsy and the cortical basal ganglia degeneration and the like, yes, I do agree that in these disorders the pathology is diversified. It is extensive, but then the clinical stage at which one would confuse Parkinson's disease with these disorders would be late-stage diseases. I don't think an early Parkinson's disease patient would be confused for a patient with such a progress of disease such as progressive supernuclear palsy or cortical basal ganglia degeneration. You did attribute that if one had to look at the entire proprietary perhaps FDG PET may answer these questions. Therefore, if you are really dealing with a clinical situation of a patient with Parkinson's and therefore you have to distinguish them from these, if you took lay Parkinson's patients you're going to see pathology all over. And FDOPA study may not be the only answer to it. DR. BROOKES: Okay. I mean, two things to comment on. Firstly, I get sent the phone number of tremor patients for scans because clinicians who are experts in movement disorders really can't make up their minds. The treatment response is not as clear cut as you might think. If you have, as you say, a patient who starts with a fine postural bilateral tremor, they've had it nearly all their life, you can be pretty convinced it's a benign essential tremor. That is true and it will respond to beta blockers usually. But there are later onset tremors which are less symmetrical. And, in fact, one of the problems is there are some families in which some of the patients or relatives get a tremor and some get pure Parkinson's disease. And it's a big question as to what is going on. So there is this gray group where they have tremors where you are really not quite sure. There is some degree of mild rigidity. They don't swing their arm, it's a bit asymmetric, you're just not sure. And although it's not a huge clinical practice, it's a significant one; similarly with Parkinson's disease you've suggested that if you give them Levodopa that gives you the answer. But in fact there has been two big clinical or pathology series now and about 5 percent have proven brain stem -- body disease cases never had a DOPA response. And about 20 percent only had a poor one at best. So, again, it's helpful, but it's not reliable. I am not trying to oversell this technique. In fact, I have no axe to grind at all, I should say. But it does seem to me there are occasions when it can be quite useful and to throw it out simply because -- you know, I think -- and as I say, I get a steady flow of patients where people just aren't quite sure and they feel the scan could be helpful. This is really all I'm saying. The cortical basal business you said it was something that is a late problem. I'm not sure I agree with that, because quite often one sees a patient who comes in where they've got stiffness in one arm and it's been going on about a year. But you think there is something funny going on, they haven't responded to DOPA as well as you would like and they seem to be a bit apraxic too. It's not just that they are slow, they have difficulty imitating as well as they should when you give them hand postures to imitate. So you're not quite sure. And, again, if you see a uniform loss of DOPA uptake, it tells you, this is likely to be atypical disease here with a much worse prognosis. DR. LOVE: Let me ask a question. You mentioned earlier studies that you have been involved with that had to do with genetic transmission. Is that the population you were just describing where you're trying to differentiate at-risk family members and how large a study -- DR. BROOKES: Yes, we've looked at about ten kindred in detail now to where there is clear multiple members affected. And you would predict that in the at-risk age range, about half of them would be gene carriers. We are finding about a quarter of these relatives are sharing what looks like preclinical disease with the bitumens affected, but not the core dates. And I think of those who are affected, about a quarter of those over five years have gone on and developed it. So we are now in the position in these kindred to start picking up preclinical disease and it looks as if we have a hit rate of about 25 percent overall or, I guess 50 percent of gene carriers in the at-risk age range. DR. LOVE: Listening to your conversation it seems as though the -- or correct me if I'm wrong, basically you are suggesting that the value of FDOPA may be in a smaller population, be it an orphan drug, what we might call an orphan drug population in this country, very small set of patients who may very well have clinical benefit from this particular image; is that what you are saying? DR. BROOKES: Well, I'm not quite sure what size it would be. When you consider that there are several hundred thousand sufferers, if we are only talking about 10 or 15 percent, you've still got the fairly sizeable population out there who could usefully have a scan to aide in diagnosis. So I'm not clear what you regard as "often". DR. LOVE: Without getting to the specific numbers, but I guess my question is, are you suggesting that this would not necessarily be a routine test for evaluation or diagnosis of Parkinson's, but that in where there are certain specific questions that need to be answered, then this test might be of value? DR. BROOKES: Well, I think that's true. But I think also you don't want to underestimate the patient lobby. I think a lot of patients have that when you first say to them, I think you have early Parkinson's disease, say, well, can you do a test and prove it? And that's a whole new indication. But from a jobbing expert's point of view, he wouldn't want to scan all these patients, I wouldn't have thought. DR. LOVE: Well, I guess one of the questions from Dr. Raman's presentation certainly would be the extent to which the information would be diagnostic, pathicneumonic, or put you into a disease category -- a syndrome category of Parkinson's and to what extent would it confuse the diagnosis if there were other disorders that might have abnormalities. Certainly you have indicated that there are some where there are other definitive image characteristics or clinical characteristics that would aid in the diagnosis or a response to treatment. After you exclude some of those, do you feel that you would be getting to a situation where you could actually diagnose the disease or the syndrome? That's just a general feeling there one way or the other? DR. BROOKES: Yeah, I think you could only say you have a Parkinsonian syndrome. I think you are quite right, you can't pronounce on the pathology with PET. It's not specific enough; that's quite right. DR. LOVE: You mentioned a moment ago about -- one, you mentioned that there were several -- there are data. Are these data that are published either from your institution or in your country that might -- DR. BROOKES: Which data are we talking about? DR. LOVE: Well, one that was the genetic transmission and also you were talking about the relationships or the ability to discriminate. DR. BROOKES: Yes. DR. LOVE: 2D, 3D, and how you might interpret the 3D differently, what advantage it would be. And I think that goes back to a question that Jane Axelrad had asked earlier. One of our challenges for us, internally is going the be to relate the 2D database to the 3D database so that the information doesn't have to be reproduced with the 3D. So we need some information that shows the translation from one technology to the next. DR. BROOKES: Okay. Well, the genetic stuff up until the 30-second relative came out in annals two or three years ago, Paccini's, I guess you have it, and there is also a twin database that came out more recently in annals from her. The 3D data came out last year in brain from our group. Now, I think most centers are now using this approach, but I don't know how wide a library of data is available in the states, but I think everyone pretty much is signing up to it. In the next year or so when some of these big drug trials are finished, the plan is to process them all with this technology and that will make a huge amount of data available, both sides of the Atlantic. But I guess relatively few people have published 3D SPM process data at this minute in time. MS. KEPPLER: One thing to remember, and my colleagues can correct me if I'm wrong, but I believe a lot of the scanner out there that are in use are still 2D scanners. So it would be difficult to throw away the 2D data on the basis that it might be how it's used. DR. LOVE: Right. I certainly don't mean throw it away, but our issue would be how would we label it? And particularly if the diagnostic -- let's say we are looking at a diagnostic indication and the image patterns are different on 2D so you would be approving -- the question we would have is, do you approve in say if you get a 2D camera image, this is your pattern, if it's a 3D image, this is your pattern, or is there something different? And obviously I think you certainly need to know what technology you're using and what assumptions you are making. But, from our perspective when we're moving from a data set to an indication and is it a broad indication, is it going to be one that says, evaluation of the dopaminergic system in patients with a wide variety of movement disorders, or is it going to be more to diagnose a syndrome of Parkinson's or Parkinsonian syndromes, or what? I think these are some of the kinds of areas where we need some clarity so we know our direction. DR. BROOKES: The first thing to say with regard to the 2D and 3D business is that the 2D is just less sensitive so you are more likely to get false negatives rather than run into any other problem. Most people who have state-of-the-art equipment can do the 3D, I just don't know how widespread these scanners are at the moment. From the point of view of movement disorders in general, there is not a whole heap of other areas outside of tremor and Parkinson's and the atypical variance where dopa scans is of huge value, to be honest. The psychiatric disorders you mentioned, there are mild increases, but I don't know that they are particularly diagnostic. These involuntary movement disorders, I don't know that Fluorodopa cans contributes a great deal to that to be honest. I think it is primarily in the Parkinson's area. DR. LOVE: I think from the Parkinson's perspective, again, we are looking at what is the utility and how would we be incorporating that into any labeled indication. And we have been -- we obviously have had dialogues about Parkinson's and Parkinsonian syndromes in general with persons who are looking at a lot of different aspects of this system and this issue of what is the clinical setting, is it a screening, is it a diagnosis, is it case finding in relatives who may be at risk? All of those are different clinical settings. It raised different issues, questions, populations and the like and whether or not you have a rare disorder that might be confused, or not, it affects the kind of use of the information, how useful is it to you to make a decision as well as what kinds of patients are coming in. And given the fact that there is a relatively low incidence or prevalence of the disease, doing long-term studies certainly has a challenge in and of itself. So that's the other reason I am particularly probing these questions to figure out what do we need to do next? DR. BROOKES: Two things. Firstly, it's probably not as low as you might think. It's about as common as Alzheimer's. So you're talking 2 percent of the elderly. There are a lot of cases out there and probably the same number again undiagnosed. From the point of view of screening, that is not realistic. It would be far too expensive and impractical. I think if there are people who have a clear dominant transmission and they are in the at-risk range and they want to know. Even then if you get a negative result you cannot say to them for sure they will never get the condition. So I have never advocated its use really as a screen. What we have been doing is looking in kindred where we are trying to look for the genotype and using this as a way of picking up affecteds to try and do the genetic screening. But I don't think it's something one would ever do on a population basis. DR. LOVE: Then it really comes down to a specific category of patients that appear to have the syndrome, but you are really doing an unusual differential diagnosis, perhaps in most situations clinically? DR. BROOKES: Well, that's one category, the other category is where there is diagnostic uncertainty. I think despite what you read in the textbooks, there are some tremors that you really aren't sure, and there are some patients with a rather stiff arm where you are not sure whether it's rheumatological or they really do have underlying Parkinson's. So there are a collection of patients out there when it can be quite useful. But, to be frank, I think patients have been screaming for some sort of doctor independent test, because there is nothing more frustrating when a doctor says, well, it's my impression you have early Parkinson's disease. And they say to you, well, can you prove it? And I think a lot of them will want a test if it's available. DR. LOVE: Dr. Houn had to step away for a moment, the question was the differential diagnosis with other diseases and the others that had a decrease -- DR. BROOKES: With other Parkinsonian syndromes. Well, our impression is that in all those patients where we are convinced that it is Parkinsonism, the PET scan is positive, I think. I can't recall having had a false negative. As to discriminating these very oddball variants, then usually one has to do an FDG scan along side the DOPA scan because these other variants target the striatal neurons themselves and not just the nigral neurons. And the only way you can be sure whether you're dealing with a striatal as opposed to a nigral degeneration is to look at both with the two techniques together. I don't think Fluorodopa PET is sensitive enough on its own unless it's one of these PSP-type case. If I see someone where the core date is as badly affected as the butamben, I'm very suspicious that it's not Parkinson's disease. And that is really the only clue you get. DR. LOVE: I think given the fact that you have identified a few other articles, we would certainly ask if we could have copies of those. I think that they may be useful. DR. BROOKES: If you let me know what you want, I would be happy to -- DR. LOVE: Yes, yes, and we can do that after -- you wanted to say something, go ahead. DR. RAMAN: Just to say that the two or three articles that Dr. Brookes was in the list of the articles of peer review particularly the ones with the kindred and the family of Parkinson's. DR. BARRIO: I would like to ask, actually, David a question. One of the things that happened rather recently, at least in UCLA and probably in our centers too, that this pallidotomies are being performed in patients, you know, these surgical procedures to improve the symptoms of Parkinson's disease. And surgeons before the surgery are sent patients, number one, to make sure dopaminergic; and number two, have the FDG that appears to be higher in the -- butamben in Parkinson's disease. And also -- toward FDG post-surgery that supposedly were brought down. It's about 10 percent because of the result of the surgery. Then the question for him is -- for David is whether in terms of the users of Fluorodopa you have any experience in that or know of that particular use in your country of Europe. DR. BROOKES: Well, this, I guess is referring to David Eitelberg's work, is it? That was FDG really, not FDOPA. What he found was the -- DR. BARRIO: No, no, I'm referring to surgeons, one in Arizona, somebody else, I believe, in California, this is a common -- a rather common surgery that's being performed in patients now. DR. BROOKES: Right? Pallidotomy. But my understanding was what David showed was if they had reduced striatal glucose metabolism, they did badly. So that he was advocating screening Parkinson's disease with FDG PET rather than FDOPA PET. And if they had low glucose you shouldn't do a pallidotomy on them and this is obviously a fairly small, select group of patients. But now that we are -- I guess the same might apply for the subthalamic stimulators as well. What he was basically saying is if you have a typical disease you do badly when you have surgery. But that was more an indication for FDG rather than FDOPA PET, as I understand it. DR. LOVE: Someone in the audience has a comment? AUDIENCE PARTICIPANT: I'm Thomas Shalley from North Shore University Hospital. Actually we did that study, the valid -- study, we used to do the patient with the FDG like Dr. Raman said, and then we did the Fluorodopa also. And this is how we selected the patient for this pallidectomy. And then we reevaluated them with the Fluorodopa and their -- again. We did the patients like 20 or 30 in our center. And quite a lot of them are very successful in that. DR. LOVE: Someone was asking if that information is publicly available also, please? DR. BROOKES: It's published. DR. LOVE: It's published. DR. BROOKES: Yes, it's published. DR. LOVE: Are there any other questions or comments for the presentation? AUDIENCE PARTICIPANT: It's Thomas Shalley again. Regarding the 2D and 3D, see, in the U.S. I don't think there are that many centers doing these studies, only specialized centers that are using these Fluorodopa studies. It is difficult to get and not everybody is -- so most of the centers they are using now for this Fluorodopa studies, they have changed to 3D already. Most of them. I don't know anybody who is doing with the 2D now for these Fluorodopa studies. So in the future there won't be much confusion regarding the 2D and 3D. DR. LOVE: Dr. Barrio, you had another question? DR. BARRIO: Yes, actually I wanted to go through some of the slides to make sure I understand some of the statements, if you allow me a couple of minutes to -- I mean, if we can go through this. I think we addressed the issue of the 3D and 2D, I think I would like to restated in terms of Dr. Brookes's statement that the extrastraital localization that has been referred to is -- from my perspective of the animal studies, animal models, of course, and got this perspective from human studies is actually very minimum. And also geographically, localizing the brain in the completely different areas, therefore, I mean, you can never be confused for what is going on in the -- butamben. In fact, using fluorometal titles that has a higher background thing to noise background we were thinking at first that when you, the disease progress or rather using animal models with more MPTP and have more degeneration that the brain stem -- will certainly go up indicating that there was some noraminergic or whatever, the outtake there, there was increasing with the progression of the cyst. But when we quantitated this clearly it remains the same. Actually this is a very constant number when you can see the butamben going down significantly to almost controlled levels. I mean, to -- levels, depending on the density of the cyst. Then even though -- you know, we don't know if this might be norminergic perhaps, but even though that's maybe the case, it really doesn't matter, it's really far away from the site of action and there's no confusion really on what is going on. The other question I have for Dr. Raman relates to something a little bit more trivial, and it's to this -- a slide number 42. The Fluorodopa literature review preliminary safety summary, it is fluorine related, I assume variation dosimetry concerns. I am not sure I understand this. Of concern is specifically pediatrician population and then refers to sodium fluoride and I know you mentioned that, but I lost track of what you were saying, therefore I would like to ask you just to make sure I understand what you mean. DR. RAMAN: It is just to kind of not to forget that the fluorine compound, you know, particularly its issues and concerns and safety in terms of deposit in bone, et cetera, et cetera, as it may be in the pediatric population. Again, I have said I deferred for the comments on this because, you know, this has already been reviewed and it has been approved when sodium fluoride was -- but just not to forget that aspect when the pediatric population would be studied, that was the point. DR. BARRIO: Yeah, the Fluorodopa does not fluorinate -- under any condition we know. Therefore, in that regard should not be any concern. It's like FDG. And in fact patients do not or should not contain any fluoride. It's the same situation we have with FDG. DR. LOVE: I think the other issue is, right now we are really focusing on the clinical and biopharm and peak pharm/tox information. We just haven't looked at the chemistry yet to see about free fluoride ion and the like, so that's why it's deferred. DR. BARRIO: Then the question for us in this junction is whether we could look at an indication based on David's assessments of the clinical situation and his enormous experience in that particular area and the work done by Dr. Charlie and -- many in Europe. Whether we can, you know, -- well, the question is, would you like to see more data? Would you like to review the data that Dr. Brookes could make available to you? DR. LOVE: Yes, I do think we want to see the other information that has certainly been mentioned today to get a sense of the kinds of studies and the situations in which the data were collected, and that would help us, I think, in deciding exactly what our next step would be. I think one of the comments that certainly Doctors Raman, Sancho, and Laniyonu have mentioned is that clearly it seems that FDOPA is marking the dopaminergic system and we are not really questioning that, per se. We are really looking at the clinical utility again, how are we going to label this, what kinds of information are available to move us from where we are to clarifying some of these questions. It certainly sounds like you have some information that might assist in that. And I -- go ahead -- MS. KEPPLER: I think I hard Dr. Brookes or Dr. Raman say that the articles that Dr. Brookes had mentioned were in the search that he's done. DR. RAMAN: Yes, there are two of those, two of the three that he mentioned. MS. KEPPLER: So there is still one remaining? DR. RAMAN: Yes. DR. LOVE: If we have seen them, then I think the question is that we would be looking at them to see if the context in which they are being considered may need to change. And -- DR. BROOKES: There is an article specifically on tremor. I'm not sure if it was in your database or not, but I did a study some years ago looking at tremor and Fluorodopa. DR. RAMAN: Perhaps I do have it, Dr. Brookes. I have seen 26 or 27 of your articles. DR. LOVE: But, at any rate, I think we will need to look at this. We were approaching this initially from a more global perspective, what is the benefit of dopaminergic system evaluations. And if we are moving to the more precise one, then we need to look -- we consider the database that we have, plus the new information to see how it all fits together and then determine what else might be needed. Yes, someone in the audience has a comment. AUDIENCE PARTICIPANT: Hi, I am Tom Shalley again from North Shore University Hospital. I would like to mention this point that we have done a lot of cases in the case of when there -- transplant was going on. I am not going to talk about the -- of the -- transplant, but I am talking about the assessment of the patient using Fluorodopa. We did that one before and after. And those studies where -- to our patients that the symptoms and the -- butamben -- it was very level. So those studies are very useful studies and Fluorodopa was very useful to us in those patients. Whether the operation was successful or not, that's not what I'm talking about. I am talking about the patient -- DR. RAMAN: Thank you for your comments. Yes, I am familiar with the work you are talking about. In fact, although I did not review these like I did for the others, I have it in my database. The problems that I encountered there were the numbers. There were several studies which included four or five patients, six patients, and -- [Simultaneous conversation.] AUDIENCE PARTICIPANT: [Off mic.] Twenty patients. DR. RAMAN: I would be very happy to look at those one more time. Additionally, also, these were cases of end-stage Parkinson's disease. These were patients who had failed all other treatments. And these were post-surgical patients. In fact, there were two articles that I thought were very strong, but the number was only two to three and two of those articles followed the patients and they even had autopsy information on that. AUDIENCE PARTICIPANT: [Off mic.] We have followed the patient -- Thomas Shalley from North Shore University Hospital. We have followed the patients up to two years now -- two to three years now. In some cases there was a tremendous increase in the progress in the patients, you know, there was no tremor, not much tremor, so there was tremendous improvements in certain cases. So for assisting these patients the Fluorodopa is needed, and we need that. DR. RAMAN: I agree with you. Perhaps it may be useful, but, again, there are also -- there are strong clinical outcomes that were used and that were compared with. And, again, I have not formally reviewed these and the initial impression, at least looking at these post-transplant patients were the numbers of these patients that were studied across the board. AUDIENCE PARTICIPANT: This is Thomas Shalley again from North Shore University Hospital. I would like to say the patients came there were at the end stage. At that point they are not worried about the efficiency, efficacy or anything like that. They just want to know, are they going to get some benefit out of it. And some of the patients came and they got some benefit. So if it is useful, I think you should give some kinds of risks for Fluorodopa. DR. LOVE: I think that what we have heard today has been very helpful as far as FDOPA is concerned and we are going to take a look at it. We have been sort of passing notes back and forth amongst ourselves thinking about what is going to be our next step. Obviously a lot of this information needs to go back before the advisory committee and there are number of questions for clinical utility for use in Parkinson's disease that would be very relevant and would be helpful to have a broader discussion. But I think this additional information that you have identified would be useful for us to consider. We will be looking at dates, do we want to -- go ahead, Jane, you wanted to say something? MS. AXELRAD: Well, I think that we both sort of heard each other out for the first time on FDOPA and for the second or third on O 15 water. I'm not sure where we are. And I think it's fair to say that we are not there yet in having sort of a meeting of the minds entirely of what the data show. Now, we have said that we are going to go back and look at the data that we have and see if we can -- and the indications that you've suggested and see if we can, you know, reach some conclusions that are closer. I do not see the likelihood that we will be able to have another public meeting like this before we go to the advisory committee. So, if we are talking about going to the Medical Imagery Advisory Committee sometime this fall, probably late October or November. We will just have to go with whatever it is we have. I think that, you know, we can have a status call probably to let you know, you know, where we are on it, that would not be in a public meeting, but we could just, you know, talk status about what we are doing so that you would know whether we will be presenting, you know, a recommendation that it be found to be safe and effective for a particular indication like we did at the last advisory committee, or whether we'll be saying, we're not there yet, we have questions about it, and raise issues to the advisory committee. In which case we would make a presentation much like we have done today supplemented by any additional data that we may have. And then presumably you would make a presentation at the advisory committee making whatever case you would like to make with regard to both compounds. And then we will get the advisory committee's advice and then decide what we are going to do from there. I mean, I sort of think that would probably be the best way to proceed if everybody is in agreement on that. DR. BARRIO: No, I'm not. [Laughter.] DR. BARRIO: I think the best way to proceed would be -- I don't think that the MIDAC forum would be the best to debate whether this is, you know, good or not. We welcome the notion of going to MIDAC, of course, and if this is the alternative, we will go. But I think I am very confident, you know, that we -- I'm biased, of course. I spent 20 years of my life working on that. If I asked Dr. Brookes the same question, he would probably tell me the same thing and other people working in the field. You know, we have to remember also this is not a condition, but something very important, there is nothing out there that could make a non-invasive assessment of the center function of that system, period. There are many agents, many things for this and that, we can replace one from the other, but there is nothing here for this. Therefore, again, granted, I would like to make sure everybody understands that we are not saying, hey, there is nothing out there, therefore you have to prove it. This is not the point. The point is there is a consideration, I guess; it's an important consideration. And I think that since it has been shown already and demonstrated in animals and humans that this agent is able to determine abnormalities in the central dopaminergic function. We are really talking about in what particular situation it would be helpful, and I understand what the issue is. But I think we are confident that with the existing clinical trials that are occurring, David is telling me that in Europe this is something that is happening. You know, it's ironic that we have clinical trials sponsored by the FDA for different drugs in which Fluorodopa is used as a gold standard. Then we have a situation like this, you know, ultrapain had been used and with Fluorodopa and has been matched, you know, the utility of ultrapain based on Fluorodopa status or results. Then I think it's -- and also there are pharmacological trials of drugs, procedures, you know, it's the -- Fluorodopa is used to assess the range of the pharmacological effect. Therefore, you know, we have a situation here that we are trying to decide whether this is good and at the same time we are using it. I think, you, Peter, wanted to have a comment on the water situation? DR. LOVE: Just a point of clarification, we haven't sponsored trials on FDOPA. We understand that there are products that might be under IND that may or may not be using FDOPA as one of the control arms in their studies. But -- DR. BARRIO: I probably misspoke. I didn't mean that you are sponsoring. I mean, that there are interventions with FDA participation, I mean intervention -- pharmacologic interventions or some surgeries or whatever in which Fluorodopa is used to assess the intervention and there is a clinical trial being worked or response or whatever the word is by the FDA. DR. LOVE: There may be INDs that have FDOPA as one of the control arms. I did want to ask a question. Your response to Jane was, no, you didn't think this was the way to go. I just was curious what was your recommendation or suggestion on the approach to take? DR. BARRIO: Well, of course, my recommendation is that we can get some decision right now or make -- but I think my recommendation -- I know my recommendation because it is something you have to determine, of course. But I think we are confident that the data we can provide to you in the meantime are coming from David's work and other work that existed from Dr. Eitelberg and whatever else you haven't seen yet could be sufficient to justify probably an indication of these agents in which we can say that, you know, it's like an FDG situation, it's used for the determination of abnormalities of this entire pharmacokinetic system and then to assist in the diagnosis of the movement disorders or something like this. Assist in diagnoses, not to make the diagnoses; assist the physician to decide whether this is as described by David. But, you know, this is something that the new data will help us design. If we cannot be convincing enough, therefore, then I think Jane's alternative will be the only one, I guess, get to MIDAC and resolve the issue there. DR. BROOKES: Just ask. One of the problems with the trials that was flagged up was they weren't blinded. But if one was able to go back and retrospectively do a blinded analysis of these data, would it help at all? DR. LOVE: There are times that we have accepted that, yes. Usually we just need to make sure that the retrospective read is prospectively designed. So it still needs a prospectively designed protocol for the retrospective analysis. DR. BROOKES: And what I am trying to do is get a feel for what trial you guys would -- not that I am likely to be doing it -- DR. LOVE: Right. DR. BROOKES: -- but what you had in mind. DR. LOVE: Right. DR. BROOKES: Because I think the trial looking at cases where we just don't know is never going to work in reality. DR. LOVE: Exactly. You know, obviously we are more comfortable with a prospectively designed study, but in some MDA situations where the data were still available, we have accepted retrospective reads. But, again, it has to be very carefully planned and we probably want to talk about that. MS. KEPPLER: What I see as perhaps being useful is after you all have gone back and had a meeting of the minds of the things that you've learned here, I think our folks have a better understanding of what you see as weaknesses in both the methodologic and the clinical literature. If we could have a telephone conference call, at that point, not a week before MIDAC, because I agree that the idea of arguing in front of MIDAC is not the best approach. But instead to, you know, work towards a solution and that might be a good next step. DR. LOVE: Thank you. I think we are agreeing at this point, yes, that we would go ahead and take the information that you are going to be sending to us and we will talk internally and probably call and have a conversation on where we go next before we definitely establish the next MIDAC or not. So we will, I think an interim discussion step is needed. MS. AXELRAD: I think what I would like to do, you know, we have been very careful about trying to have all of our meetings in a public forum. So if we were going to have a conference call in which we were going to be discussing substance as opposed to sort of the status of things, we would certainly want to at least do minutes of a meeting and put them in the docket, you know, on the web, so that people can see. I think it's just really expensive to have these public meetings. It costs thousands of dollars, you know, to get the room. You all have to fly here and everything, so I don't think that we'll have enough to talk about to have a full day public meeting. So if it's all right, I would suggest that we handle it that way. We will do minutes of the call and put them in the docket so everybody will be aware of where we all are. And then, of course, if we decide that we are going to go to the advisory committee, that will be a public discussion. MS. KEPPLER: That would be the best. I mean, I think we have done that before. DR. BARRIO: I would like to say something that would reflect on a conversation I had with Dr. Brookes over lunch. It is absolutely amazing what you have done with 472 papers. Coming from -- you know, trying to understand a very complex issue, really, after nowhere because that is not -- you know, what is your field, of course -- but you know, the imaging thing is not, and there are so many technical issues that even people working the field are still arguing really, you know, that K, the other K, that procedure, the other one, or the other one or whatever, then it's amazing that you guys have done this with so much competence and I would like to really indicate that and recognize what you have done. It is very good. DR. LOVE: On behalf of the review team, I will say, thank you very much. DR. CONTI: I would like to make a couple of comments. Again, I will reiterate what Jorge said, I think it's rather remarkable that just about anyone can learn what we do now. [Laughter.] DR. CONTI: The comments are related more to issues on the definition of the indications and the categorizations. It's difficult for us to understand what is meant by these various categories because they are new. And I think it's showing on your part as well, that you're trying to assimilate them into your system so it's really -- it's difficult for us to decide which way to go when we don't know necessarily which is the best option and there's no track record to guide us in a sense. And we are talking about functional versus diagnostic indications, et cetera. That sort of is the general comment. Having said that, the issues pertaining to the 15O-Water are a little bit more complex in a sense that this is never going to be a commercial product. So we are really up against the wall to some extent with regard to commercial sponsorship of clinical trials. It's rather limiting in terms of what we are going to do. The second fact is that this tracer has been used as the gold standard in many publications now for other tracers. I've got five of those papers in front of me right now. It makes it very difficult for even an academic center to go and do another study now to pick the tracer that was used to be studied comparing 0 15 and reverse the table. In other words, use that approved tracer now to prove 15O-Water. So that's a little bit hard to swallow in terms of getting a graduate student to do that type of research study. So I think part of the problem is that we haven't really formulated some realistic indications for these two drugs so that we can go back and say, well, which is the best? We have talked around the issues about the various aspects of a pharmacology, the toxicology, the clinical studies, but we really haven't put in front of us a couple of options as far as where we want to go with indications or what categories we really want to be focusing on and maybe eliminating some others that we don't want to spend time doing. So I guess we are relying on your folks to take the data and make a suggestion to us as to what you think is the best category. But keeping in mind that I would rather see it go to a minimalistic approach where there is less restriction and then we can build on that by off-label indications down the road as opposed to trying to shove in various types of clinical entities that there may be sparse data for. So at least for the 15O-Water. Now, maybe with Fluorodopa we might want to consider focusing on a more specific patient population because the indication is more solid. So, while we have some conceptions, we also need to be able to figure out how we are going to mesh them with yours. I don't see the mechanism for that to happen right now. DR. LOVE: I think from a mechanism standpoint I would offer the same general approach that we are just going to take with FDOPA in the sense that we will look at the other information that you are going to provide and then have another TCON to try to move forward in terms of steps. But I think from a perspective I would say, number one, in terms of what indications, I still see us looking at a couple of different things, and I think our overall goal is still to move towards a more general functional indication since it is used in a wide variety of settings, we need to just determine exactly which settings would be beneficial. On the other hand, I think the brain mapping is still one that is worth thinking about. For us that would be equivalent to a priority indication meeting an unfilled need. And the kinds of information that we might be looking for could certainly be different in that kind of a setting. We still need, obviously, solid information to help, but it seems to me in many ways the kind of trial design and so forth that might be considered, and that setting might be substantially different the numbers of patients and certainly would be very different from say an Alzheimer's study with long-term follow-up. And also in Alzheimer's there's a much more complex situation. So I think that it's worth continuing to consider the brain mapping type of indication whether it's worded "brain mapping identification for neurosurgery" or whether it's a more general cerebral blood flow of the eloquent cortex. Who knows, I mean, those are language phrases that we'll have to sort out over time. But I think that continuing to look for information in that area would be helpful. We did look specifically at some of the neurosurgical literature itself, not just the imaging literature, but there may be some other information particularly from many institutions where you are located where you know that you're actually doing a lot of neurosurgery guided by PET. So 15O-Water specifically. So I think that our first step, back to the last part of your question, what's the mechanism, I think it's for us to look over the other information that you brought with you and the other articles that were identified during the discussion and then we will look back at that and consider what's been said today and then talk on it during the next TCON. DR. CONTI: I would like you again to specifically look at the issue of cerebral vascular disease and the current state-of-the-art and use the literature that does compare with PET. I want that to be considered in this evaluation. We can look at brain mapping and we may be able to design a trial to supplement what is at and what is missing perhaps. But I don't want to lose the fact that there is an established literature base on cerebral vascular disease and profusion imaging. DR. LOVE: I think that I'm saying that we're going to do that. Part of what I understand that you've brought with you are some of the articles that affect that. So we certainly will be looking at it. I think our challenge is going to be the one that we've had before and that's if we're looking at cerebral blood flow alone, or are we looking at other measurements. And if the articles are based primarily on other measurements then we will have to figure out how are we going to deal with that issue if that's the primary endpoint. That's basically the problem that the advisory committee identified. But I think if there is something else there that we can use, then certainly we would, absolutely. Thank you. MS. AXELRAD: Okay. I think we have a plan for how we will proceed from here. Just a couple of little sort of other things. Somebody came up and asked me about applications for the submission of applications for FDG and ammonia and I said, by all means, we are encouraging anybody who can to submit one. It isn't required yet. And we will be working on finalizing the guidance document. You know, we have gotten a few comments on it and we will be analyzing those comments and finalizing it and we are also working on -- I think we're calling it a sample application these days, where we are actually trying to use the guidance document in filling out the chemistry section of it and completing an application that we could post publicly so people would have the benefit of that to know how to do that. I also believe that we are going to try and schedule around the time of the ICP annual meeting if we can make it, I'm not sure we are going to make it. I may have to be later than that. We are going to aim for that because you are going to be in Washington anyway, to schedule our next big meeting on current good manufacturing practices and some of the other issues that we last dealt with in February. So what we are hoping to do is have another discussion about that and have a lot more detail in the regulation and the guidance document on that subject to discuss at the public meeting. So does anybody have any comments on that otherwise? DR. BARRIO: No, I was wondering when you were saying that if then the work or the meeting with MIDAC probably wouldn't be in the fall, right, maybe later? Is that your plan, or together, similar time? DR. LOVE: No, I think we are saying we are deferring the final decision n the next MIDAC until we have a chance to talk. DR. BARRIO: I understood that, but Jane mentioned -- I don't know, sometime ago -- in this meeting that probably we could have the MIDAC meeting in the fall. Then, you know, since we are having decision in the fall too, right, October is still the fall, I guess. MS. AXELRAD: When is your meeting? When is the ICP annual meeting? DR. BARRIO: Late October. MS. KEPPLER: October 15th through the 18th. MS. AXELRAD: Yes, I don't think we will be ready for GNPs by then because we sort of feel that in order to have a meaningful meeting we would have to get out publicly available the regulation and the guidance several weeks in advance of that and we're just not going to be able to do that, people's schedules. We brought new people in to work on that when Tracy Roberts left. We have new staff working on it and they have, you know, things scheduled to working around their schedules. We are just not going to be ready by October 15th. So we are looking at November, I would say at the earliest, depending on when, you know, when things develop. And I think that what we should do is just sort of keep track of where things are in the clinical discussion if we could work the two together since everybody would have to come, you know, whether it's in late November or in January, you know, we could work around that. But try and do it. It's really different sets of people are involved on the clinical side than on the GNP side. So theoretically they could be back-to-back or something like that. If we had the work done on both topics and could schedule it that way to minimize travel time for everybody. So let's just see how it plays out and when we get the work done and are ready to sit down and talk about it. DR. LOVE: Okay. So thank you, thanks to everybody who came. I think, you know, we have made some progress. [Whereupon, at 3:06 p.m., the meeting was adjourned.] [--- Unable To Translate Box ---] [--- Unable To Translate Box ---] MILLER REPORTING COMPANY, INC. 735 8th Street, S.E. Washington, D.C. 20003-2802 (202) 546-6666 MILLER REPORTING COMPANY, INC. 735 8th Street, S.E. Washington, D.C. 20003-2802 (202) 546-6666 [--- Unable To Translate Box ---] MILLER REPORTING COMPANY, INC. 735 8th Street, S.E. Washington, D.C. 20003-2802 (202) 546-6666