UNITED STATES OF AMERICA

P-R-O-C-E-E-D-I-N-G-S

8:30 a.m.

CHAIRMAN SUZUKI: The Dental Products Panel of the CDRH Medical Devices Advisory Committee. My name is Jon Suzuki. I'm serving as the Chairman of the Dental Panel. And I would like to call this meeting to order.

The Executive Secretary, Michael Adjodha, will make some introductory remarks.

Mr. Adjodha?

EXECUTIVE SECRETARY ADJODHA: Thank you, Chairman Suzuki.

My name is Michael Adjodha, Executive Secretary of the Dental Products Panel.

Allow me to introduce the members of our panel. Please raise your hand as I call your name.

The Chairman of the panel is Dr. Jon B. Suzuki. Chairman Suzuki is a periodontist and immunologist, and is the Associate Dean of the School of Dental Medicine at Temple University in Philadelphia, Pennsylvania. Note that change from the agenda. This change is recent.

Joining him are the following panel members:

Dr. Salomon Amar is a periodontist and is Professor at the Department of Periodontology and Oral Biology of Boston University, Boston, Massachusetts.

Dr. David L. Cochran i s a periodontist and is Chair of the Department of Periodontics at the Health Science Center at the University of Texas, San Antonio, Texas.

Ms. Elizabeth Howe is a consumer representative and is the Outreach Coordinator for the National Foundation for Ectodermal Dysplasias in Auburn, Washington.

Ms. Allison F. Lawton is our drug industry representative and is Senior Vice President for Genzyme Corporation, Cambridge, Massachusetts.

Dr. William J. O'Brien is a materials engineer and is Professor at the School of Dentistry at the University of Michigan, Ann Arbor, Michigan.

Mr. Daniel R. Schechter is the Device Industry Representative and is General Counsel for Parkell, Inc., Farmingdale, New York.

Dr. Domenick T. Zero is a cariologist and is Chairman of the Department of Preventative and Community Dentistry at Indiana University, Indianapolis, Indiana.

And Dr. John R. Zuniga is an oral surgeon and is Professor at the School of Dentistry of the University of North Carolina at Chapel Hill, Chapel Hill, North Carolina. Dr. Zungia is recovering from an automobile accident and we're pleased he could be with us today.

Joining the Panel members if the following consultant: Dr. Inder J. Sharma is a biostatistics consultant and is an Associate Professor at the Department of Community Health and Preventative Medicine of Morehouse School of Medicine, Atlanta, Georgia.

Joining us at the table is Dr. Susan Runner, Deputy Director of FDA's Division of Anesthesiology, Infection Control, General Hospital, and Dental Devices.

I will now read into the record a memorandum from the Center Director regarding voting status of our Panel Consultant.

Pursuant to the authority granted under the Medical Devices Advisory Committee charter, dated October 27, 1990 and as amended on April 20, 1995, I appoint the following consultant as a voting members of the Dental Products Panel for the meeting to be held on Tuesday, July 13, 2004. Inder J. Sharma, PhD

For the record, this individual is a special government employee and is a consultant to this Panel under the Medical Advisory Committee. He has undergone customary conflict of interest review and he has reviewed the material to be considered for the meeting. Signed Daniel G. Schultz, MD, Acting Director Center for Devises and Radiological Health, July 8, 2004.

Next I'll read into the record a conflict of interest statement for the this meeting.

The following announcement addresses conflict of interest issues associated with this meeting and is to be a part of the record to preclude even the appearance of impropriety.

To determine if any conflict existed, the agency reviewed the submitted agenda for this meeting and all financial interests reported by the Committee participants. The conflict of interest statutes prohibits special government employees from participating in matters that could affect their or their employer's financial interests. The Agency has determined, however, that participation of certain members and consultants, the need for whose services that waives the potential of conflict of interest involved is in the best interest of the government.

Therefore, waivers have been granted for Drs. Cochran, O'Brien and Sharma for their interests in firms that could potentially effect the panel's recommendations.

Dr. Cochran's waiver involves a grant to his institution for the sponsor study for which he had no knowledge of the funding and had no involvement in the data generation or analysis.

Dr. Cochran's waiver is limited in that it allows him to participate in the panel discussion but excludes him from voting.

Dr. O'Brien's waiver involves a grant to his institution for the sponsor;s study for which he had no knowledge of the funding and no involvement in the data generation or analysis. Dr. O'Brien's waiver is limited in that it allows him to participate in the panel discussion but excludes him from voting.

Dr. Inder Sharma's waiver involves a philanthropic contribution from the firm at issue at his institution for which he has no involvement and is uncompensated.

Dr. Sharma's waiver allows him to participate fully in today's deliberation. Copies of these waivers may be obtained from the Agency's Freedom of Information Office, Room 12A-15 of the Parklawn Building.

We would like to note for the record, the Agency took into consideration on other matters regarding Dr. Domenick Zero. This panelist reported past and current interest involving firms at issue, but are matters that are not related to today's agenda. The Agency has determined, therefore, that this panelist may participate fully in all discussions.

In event that the discussions involve any other products or firms not already on the agenda for which a FDA participant has had financial interests, the participant should excuse him or herself from such involvement and exclusion should be noted for the record.

With respect to all participants we ask in the interest of fairness that all persons making statements or presentations disclose any current or previous financial involvement of any firm whose products they may wish to comment on.

I'd like to request that everyone in attendance at this meeting take the time to sign the attendance sheet available at the front door.

Now transmitting you back to Chairman Suzuki.

CHAIRMAN SUZUKI: Okay. Thank you. I note for the record that voting members resent constitute a quorum as required by 21 CFR Part 14.

We will now proceed the first of two open public hearing sessions for this meeting. The second open public session will follow the panel discussion this afternoon. At these times public attendees are given an opportunity to address the panel to present data or views relevant to the panel's activities. No individual has given advance notice of wishing to address this panel. If there's anyone now wishing to address the panel, because identify yourselves at this time. Okay. Thank you.

I'd like to remind public observers at this meeting that while a portion of this meeting is open to the public observation, public attendees may not participate except at the specific request of the Chair. You will be given no more than 10 minutes for your presentation.

I would like to ask at this time that persons addressing the panel come forward to the microphone and speak clearing, as the transcriptist is dependent on this as a means for providing an articulate transcription of the proceedings of this meeting.

If you have a hard copy of your talk available, please provide it to the Executive Secretary for use by the transcriptist to help provide an accurate recording of these proceedings.

We're also requesting that all persons making statements during the open public hearings disclose if they have financial interests with the sponsor of the products under consideration.

Before making your presentation to the panel, in addition to stating your name and affiliation, please state the nature of your financial interest in the product under consideration, including who is paying for your attendance at this meeting.

Okay. At this time we'll follow the agenda and we will present with the sponsor presentation on the product GEM 21S. Mr. Mark Citron.

MR. CITRON: Good morning. My name is Mark Citron. I'm Vice President of Regulatory Affairs at BioMimetic Pharmaceuticals.

On behalf of BioMinetics we would like to thank the panel and the FDA for the time and attention that the FDA and the panel have spent in reviewing our PMA and meeting today to provide your recommendation regarding approval of our device.

We have the privilege today to present to you the results of decades of what began as scientific research, progressed to product development and clinical trials leading to today's presentation of the GEM 21S control comparison randomized study results. For the next 60 minutes we will present these preclinical and clinical results and respond to any questions you may have.

I will begin by introducing today's speakers and our agenda.

First, Dr. Samuel Lynch, President and CEO of BioMinetic will provide the brief overview of the GEM 21S device and the development of the device. Dr. Lynch is a periodontist and has conducted extensive scientific research on PDGF as well as other growth factors involved in tissue repair covering many years.

Next, Dr. William Giannoble of the University of Michigan will speak on the mode of action of GEM 21S with particular emphasis on the growth factor component recombinant human platelet-derived growth factor.

Dr. Ron Nevins, a clinical professor at the Harvard School of Dental Medicine, who is also in private practice, will present the animal and human histology data.

Dr. Bob Genco, who is currently the Distinguished Professor of Oral Biology and Microbiology at the State University of New York at Buffalo and recently appointed the Vice President of Research at the State University of New York at Buffalo will present the results of the randomized control clinical trial.

Finally, Dr. Lynch will provide concluding remarks to the formal presentations.

We welcome the panel's questions, and we have available today several of the key scientific researchers who have been involved in the GEM 21S program, and they are prepared to respond to your questions. These include the study statistician Dr. Phil Lavin. He's an Associate Professor of Biostatistics at Harvard Medical School and President of Averion, which is a biostatistics consulting firm.

We have Dr. Charles Hart, our Vice President and Chief Scientific Officer.

Dr. Jeffrey Hollinger, the Director of the Carnegie Mellon University's Bone and Tissue Engineering Center.

Dr. Michael Reddy, a clinical professor at the University of Alabama, Birmingham.

And finally Dr. Mark Reynolds from the University of Maryland Dental School.

Dr. Lynch will now begin.

DR. LYNCH: Thank you, Mark. And good morning to the panelists, members of the audience and the FDA.

I would also like to thank the panel for your time and consideration today as well as the FDA for their support and recommendations during the development of GEM 21S.

I believe it is important to note that our meeting today is the culmination of over 15 years of scientific research by multiple investigational groups working independently and sometimes collaborative. We are fortunate to have many of these research groups represented here today.

Two persons who are not here this morning but who deserve substantial credit for the development of the GEM 21 product, and who I would like to take this opportunity to acknowledge and thank, are Dr. Ray Williams, Chairman of Periodontology at the University of North Carolina and formally Chairman of Periodontics at Harvard. My mentor, counselor and friend.

And posthumously, Professor Harry Antaniales, whose lab conducted much of the early research on PDGF, who inspired much more of the scientific work in this field and in whose lab I trained.

Finally, I would wish to acknowledge my appreciation to the Biomedics Clinical and Regulatory team for their hours of preparing the PMA submission before you today as well as the entire GEM 21 group of clinical investigators who rigorously conducted the pivotal clinical study from both academic research centers and private clinical practices thereby providing us robust data from both important clinical environments.

We are fortunate today to have three individuals who are involved in the pivotal clinical trial and who are widely recognized for their expertise in clinical and basic scientific research to speak in favor of the approval of GEM 21.

Our first speaker today is Dr. William Giannoble of the University of Michigan and Director of the Michigan Center for Oral Health Research. Dr. Giannoble was a clinical investigator in the GEM 21 pivotal clinical trial, and is a recognized expert on the biology of growth factors including platelet-derived growth factor or PDGF.

As Mark mentioned, Dr. Giannoble will discuss the mode of action of GEM 21S with particular emphasis on the protein growth factor component.

Next Dr. Ron Nevins, a former President of the American Academy of Periodontology and currently the editor and chief of the International Journal of Periodontics and Restorative Dentistry who also finds time for a busy private practice will present will present the animal and human histology data demonstrating the effectiveness of GEM 21 in promoting periodontal regeneration including new cementum and periodontal ligament coronal to the original apical extent of calculus.

Dr. Nevins is uniquely qualified for this presentation, having participated in a GEM 21S pivotal trial also as well as having been the lead investigator for many studies evaluating the human histological response to a number of different drafting materials including PDGF and periodontal bone defects.

And finally, Dr. Bob Genco, past President of the International Association of Dental Research and editor and chief of the Journal of Periodontology will present the results of our randomized control double blinded prospective multi-center pivotal clinical trial.

Dr. Genco was the independent medical director for the overall GEM 21S clinical program and has many years of experience in designing, conducting and evaluating the scientific integrity of clinical trials related to periodontology including having served as the formal of this august FDA Advisory Panel.

Let me now set the stage for these speakers by briefly describing the GEM 21 product, it's development history and the unmet clinical need that it is designed to satisfy.

Next.

GEM 21S, as we have alluded to, principally consists of two main components. One component is a particulate beta-tricalcium phosphate or Beta-TCP, which is filled into a cup and terminally sterilized.

The other principal component is a physiologic solution containing recombinant platelet-derived growth factor, which is aseptically-filled into a syringe just to facilitate handling of the material.

At the time of the surgical procedure, the surgeon or surgical assistant simply peels back the lid of the cup, adds the growth factor solution to fully wet the graft particles. After a few minutes sitting on the surgical tray, specifically we're recommending approximately 10 minutes, the material then forms a cohesive mass of particles which are then packed into the alveolar bone defect.

Next, please.

One of the main and principle attributes that we would like to stress today is the extensive scientific research known about both principal components of this product, both the PDGF and the Beta-TCP.

There are well over 200 publications on PDGF that deal specifically with its beneficial effect on wound healing. These studies have been conducted in a variety of models and systems including in vitro self-culture systems using primary cultures of osteoblast or well qualified osteoblast like cellnoids, primary cultures of periodontal ligament cells and gingival fibroblast cells and many, many other cell types.

All of these studies in vitro have clearly demonstrated the receptor binding of the BDGT to the receptors, as you will hear from Dr. Giannoble.

In addition, there are multiple publications showing the beneficial effect of PDGF on a wound healing in vivo in mice, rats, rabbits, canines, swine, nonhuman primates and human clinical trials. As you can see, it's a very well studied molecule.

In addition, PDGF was the first recombinant human growth factor to be FDA approved as a wound healing agent. It is currently marketed under the trade name Regranex by Johnson & Johnson. Has been on the market for over 5 years and is absolutely well documented safety record with no elicitation of antibodies or any adverse responses in commercial use.

In addition, the beta-tricalcium phosphate has is an FDA cleared bone augmentation device. It is the Beta-TCP that we incorporate into GEM 21S. Is on the market in a larger particle form under the trade name Vitoss by Orthovita for orthopedic bone regeneration procedures.

Next please.

Thus, as you will hear this morning the benefits of GEM 21S are it is a fully synthetic bone regeneration system supported by hears of research that have elucidated its mechanism of action and demonstration a strong safety profile. And again, rigorously conducted clinical trials and commercial use.

The PDGT component has specifically been shown to enhance periodontal regeneration in both animals and humans. Our pivotal clinical trial has demonstrated that the product accelerates the attachment level gain and enhances or improves significantly radiographic evidence for bone regeneration.

Finally, we hope to show today that this product demonstrates minimal risk and has the potential for strong benefits in clinical practice.

Thank you very much. And I would now like to turn the presentation over to Dr. William Giannobble to discuss the biology mechanism of action and highlights of some preclinical data on the product. Thank you.

DR. GIANNOBLE: Thank you, Dr. Lynch.

And I'd also like to thank the FDA and the FDA panel members for the opportunity for me to present to you this morning some of the basic biology in the extent of preclinical data that have demonstrated some of the safety and effectiveness of platelet-derived growth factor the GEM 21S system for the promotion of periodontal regeneration.

So as we look at periodontal disease, which typically it's a disease that results from a microbial infection that leads to the resorption of alveolar bone through to its cementum and periodontal ligament. There are a variety of different factors that appear to be critically important to the reconstruction of periodontal wounds; those being the appropriate cells within the lesion, they can repopulate the wounds such as osteoblast, cemental blasts, periodontal ligament fibroblasts within the presence of the appropriate scaffold that will then allow cell ingrowth and vascular invasion into the lesion.

And then the usage of signaling molecules or growth factors that can direct the migration of cells into the wounds, promote proliferation of the cell types within the defect and stimulate matrix biosyntheses.

In addition, given that the structure is a vascular, it is critical to provide an angiogenic environment to promote new blood vessel formation to reconstruct these periodontal wounds. And so in my presentation this morning I will focus on platelet-derived growth factor and the scaffold, the osteoconductive scaffold beta-tricalcium phosphate for use in promoting periodontal regeneration.

So as we look at the two key components of the GEM 21S system, the first being the Beta-TCP, as this is an osteoconductive scaffold that promotes cell attachment ingrowth, it also has been demonstrated to prevent soft tissue collapse into the soft tissue defects, and also facilitates blood clot stabilization during the initial wound repair process.

Recombinant human platelet-derived growth factor BB there is a very extensive profile in terms of its demonstrated ability to promote chemotaxis of the key cell types involved in tissue repair. It is also mitogenetic or promotes proliferation of these various cell types such as periodontal ligament fiberblast and osteoblast.

And PDGF has also been demonstrated to be to be an angiogenic molecule by recruiting smooth muscle cells that are important in the formation of new blood vessels.

Next slide.

So to go into a bit more depth on beta-tricalcium phosphate, this is a synthetic purified calcium phosphate ceramic that has a very extensive history in the FDA as well as a device used in dentistry and in orthopedic applications as a bone void filler. And in this long history of usage there have been no demonstrated adverse events utilizing beta-tricalcium phosphate as a bone void filler in these varieties of applications. And recently the FDA Advisory Panel recommended a reclassification of Beta-TCP from a high risk device to a lower risk device for use in dental applications.

This slide demonstrates scanning electron microscopic views of beta-tricalcium phosphate at lower power magnification here and a higher power magnification. The low power view demonstrates the beta-tricalcium phosphate granules which in the formulation for the GEM 21S system range in particle size from 250 to 1,000 microns in diameter. This higher magnification view demonstrates the very open pore structure of the Beta-TCP used in the GEM 21S. It has a 90 percent open pore structure which then this porosity, this ranging from 1 to 1,0000 microns in diameter thus allows cellular ingrowth and vascular invasion. This lower panel demonstrates at a different microscopic view the growth of osteoblast like cells on top of the beta-tricalcium phosphate demonstrating that it does promote cell attachment and proliferation on the device.

Recombinant human platelet-derived growth factor has been an extensively studied molecule in the area of wound healing. So it's a natural wound healing hormone released from platelets during normal wound repair.

The scientific established mode of action is that PDGF has been demonstrated to promote connective tissue formation, also osteogeneses and angiogeneses by the induction of vascular endothelia growth factor in the recruitment of smooth muscle cells.

This diagram depicts the binding of platelet-derived growth factor, which is a dynaric protein which binds to cell surface associated tyrosine kinase receptors. These receptors dimerize and then elicit autophosphorylation of the receptor. This autophosphorylation event then leads to a variety of different signal transduction pathways which will then led to the elicitation of the variety of different biological effects such a mitogenesis or cellular proliferation, directed cell mitigation or chemotaxis, and also the blocking of program cell death or promoting cell survival.

So PDGF more specifically as we examine its ability to promote periodontal regeneration within the periodontia, platelet-derived growth factor and its associated receptors are naturally induced during normal tissue repair, both soft tissue repair and during the fracture healing procedure.

PDGF has been demonstrated to be chemotactic for a variety of cells derived from the periodontia as well as promoting cellular proliferation and matrix biosynthesis. And there is a large body of work supporting the variety of effects as shown here.

PDGF also promotes cell survival since a PDGF alpha receptor encodes for a growth arrest specific gene. So PDGF will promote or prevent apoptosis or programmed cell death.

PDGF also enhances angiogenesis specifically by promoting the proliferation of smooth muscle cells or parasites around the newly formed blood vessels and it compliments the actions of VEGF or vascular endothelia growth factor that's critically important for blood vessel formation and maturation.

This slide published by the San Antonio group demonstrates the effects of recombinant human PDGF in an artificial wound model on promoting cell repopulation. And so what we can see in this slide looking at percent wound fill or cell repopulation of periodontal ligament fiberblast versus a low serum control, this graphic demonstrates that over a period of ten days the significant increase in cellular repopulation into artificial wound defects by the application of recombinant human platelet-derived growth factor.

Next slide.

This slide demonstrates the ability of platelet-derived growth factor applied onto the beta-tricalcium phosphate osteoconductive device for its release and then subsequent biological activity of the release PDGF. And so this slide shows treated Thymidine incorporation as a method to determine DNA synthesis over time when PDGF has been applied to the beta-tricalcium phosphate device. And so what one can note is that there is a rapid release over the first 24 hours and the PDGF that is released is indeed biologically active as measured of the promotion of DNA synthesis.

This slide demonstrates results from an in vivo animal study done in Beagle dogs where fenestration bony defects were created on two surfaces and then autoradiography was performed to look at cells that were demonstrating active proliferation within the periodontal wound compartment.

So the variety of different cell types examined were those important in periodontal repair such as fibroblasts, cementoblasts, osteoblasts, perivascular and endothelial cells. And what was noted that it was compared to control or surgery alone defects, PDGF promoted at least a three to five full increase in cellular DNA synthesis as noted by the autoradiography. And you can see this in a multitude of different cell types that were found within the lesions, thus demonstrating that the PDGF has pleiotropic effects on promoting a variety of parameters associated with periodontal regeneration.

This slide published by Bob Genco's group several years ago in a canine model of surgically created critical size defects in dogs. These are class 3 furcation defects that do not typically heal.

The defects were treated with guided tissue regeneration, a standard treatment modality for periodontal regeneration versus recombinant human platelet-derived growth factor applied to the tooth root surface combined with GTR. And looking at histomorphen metric analysis to determine the amount of regeneration that occurred within the defects, what was noted was that PDGT strongly augmented the degree of newly formed bone and periodontal ligament, while at the same time blocking really the production of the granulation tissue or scar formation that resulted after this healing period.

This slide demonstrates the potent effects of platelet-derived growth factor on promoting osteogenesis. This is a study published several years ago that examined an osteoporosis model where female rats were ovariectomized which induced a rapid bone loss. And the slides on the left demonstrate the metathesis of the tibia in these animals either in an osteoporosis saline control or animals that were delivered a three times per week infusion of 2 milligrams of recombinant human platelet-derived growth factor.

What one can note from these long bones was that there was a significant increase in the boning trabecular in both the primary and secondary spongiosa in these bones that were treated with -- these animals that were treated with recombinant human PDGF.

Using histomorphic metric analysis of the vertebral body and then tibial metathesis once could also note a statistically significant improvement in bone density measures nearly two-fold in both of these different bony sites.

This slide now demonstrates the platelet-derived growth factor's ability to promote periodontal regeneration. This is a natural disease model in the Beagle dog that will result in loss of connective tissue and alveolar bone. So this slide demonstrates a through and through class 3 furcation defect that typically will not heal on its own. These animals were delivered a single application of recombinant human platelet-derived growth factor in a beta-tricalcium phosphate carrier, and this slide shows six weeks after this single application of PDGF plus the Beta-TCP. The promotion of new alveolar bone, a periodontal ligament and tissue consistent with cementum.

These various preclinical animal studies performed in dogs were also followed by in nonhuman primates in the monkey model Macaca Fascicularis. And what this side is demonstrating is the consistency of effects in the animal model Macaca Fascicularis versus humans when platelet-derived growth factor was combined with insulin like growth factor one. So this study looked at animals that were treated with a single application. If you look at the parameter of ostis defect fill, there is a striking similarity between the monkey model and this is -- the human data here is derived from a multi-center trial done, it was a phase 1 phase 2 trial done at the Harvard School of Dental Medicine and at the University of North Carolina. And essentially the bottom line of this study was demonstrating that similarity between the animal model and the human.

Next slide.

The next few slides will now demonstrate the extensive track record for the various components used for the GEM 21S product. And with the Regranex product that has been FDA approved for, it's been over five years now, it has a very extensive safety record. And so the results shown here are actually a compilation of six randomized controlled trials where the Regranex product demonstrated extensive safety. There was no neutralizing antibodies that were developed. And these patients received the treatment of the Regranex every other day for up to 140 days of a concentration of 100 milligrams per mil of the PDGF. And so to date there have been at least 17 million doses applied of Regranex, demonstrating its safety.

Also you have provided to you very extensive confirmatory biocompatibility tests. As you can see on the list here, in terms of cytotoxicity, sensitization, acute systemic toxicity, genotoxicity and muscle implantation for the GEM 21S product. And so all of these tests have demonstrated that GEM 21S is both biocompatible and safe.

So what I would like to summarize for you this morning is that we have demonstrated that the mechanism of action of platelet-derived growth factor is well established as shown in vitro studies as well as in vivo applications demonstrating its potent ability to promote periodontal regeneration, i.e., tooth group cementum, periodontal ligament and alveolar bone.

This is also a very long history of safety for both of the components, the beta-tricalcium phosphate in both dental and orthopedic applications and the platelet-derived growth factor component, i.e, in the Regranex product for the treatment of neuropathic diabetic ulcers.

The results have also been demonstrated to be quite consistent amongst the large body of research done with a variety of different clinical investigator reclinically that bridge and demonstrate consistency to some of the human clinical studies that have been performed.

I would like to thank you for your attention and I look forward to the discussion this afternoon.

I will now introduce Dr. Myron Nevins who will present the proof of principle data on the ability of platelet-derived growth factor to promote periodontal regeneration in humans.

DR. NEVINS: Good morning.

I'd like to take this opportunity to thank the FDA and the panel by allowing us to demonstrate the evidence of regeneration, periodontal regeneration that we've been able to achieve with GEM 21S.

The definition of periodontal regeneration is histologic. It has evolved from proceedings of two world workshops in clinical periodontics and it is inclusive of information of new bone, new cementum connected by a functional periodontal ligament on a root surface that has previously been pathologically exposed.

The hierarchy of evidence in periodontal regeneration has taken years to evolve, but because of the histologic definition, it's clear that the most compelling evidence are human studies that have histologic evaluation. In lieu of the obvious difficulties in obtaining this information, other means have surfaced to measure success, including RCTs with measuring clinical invaded ethic parameters. Perhaps the more contemporary benchmark has been the use of, be it surgical reopening, which would more closely mimic the radiographs in terms of interpretation.

Proven principle assessment is established to demonstrate the safety and the effectiveness of a product. Safety would determine histologic tissue reactions, healing response and provide a clinical assessment for safety. Effectiveness provides human histologic evidence of regeneration, in this case for vertical intrabony defects and also for Class II furcation invasion problems.

This design include 11 intra-osseous defects around teeth scheduled for extraction, six intrabony and five Class I furcation defects were treated. They were treated with a combination of recombinant BDGF plus a carrier. At nine month post-operative follow-up recordings of the CAL, the pocket depth gingival recession and linear bone -- were recorded. At that time the teeth were abstracted with a small amount of surrounding tissues and submitted to blind histologic analysis to assess regeneration.

I should mention at this moment that informed consent was obtained from the patients. The patients were rehabilitated from the site with bone crafting, dental implants and a prothesis to reestablish or in all senses to provide them with a dental solution that they would not have been able to have otherwise.

The intrabony defect results demonstrate a pocket depth, a mean pocket depth of 9.7 millimeters and at nine months at time of the harvesting block, 3.3 millimeters. This is a change from baseline of 6.42.

The importance of this is related to length of roots in a human model. The smallest, the shortest roots are the incisors, the central incisors with 11 millimeters and of course the longest would be the cuspids which approximate 18 millimeters.

If we accomplish a 6 millimeter correction, this definitely changes the prognoses of the tooth.

The CAL gain started -- level started at 11.1 baseline and was 4.9 at nine months. Once again, for a change from baseline of 6.7, which is consistent with the pocket depth reduction.

Bone height change shows radiographically a 2.14 improvement.

This will become in a few minutes when we look at the histologic measurements, because there will be a correlation between what was here radiographically and histologically.

The furcation defects from a pocket depth began at 6.2, in nine months were 2.8 for a change from baseline of 3.4. And the clinical attachment level changed with a change from a baseline of 4.

Since these were horizontal as well as vertical probing depths, these are very significant in reversing the invasion of the furcation by inflammatory periodontal disease.

When the teeth are described as being candidates for extraction, it's most effective when we look at a clinical photograph. And here we see a maxillary cuspid with bone defects both in vertical dimension and bone morphology that would be serious candidates for extraction.

We determine the level of the root that has been exposed to disease by the presence of calculus. So at the base of the calculus a notch is made with a small burr to designate that we actually have evidence that disease occurred at that point.

The calculus, of course, is removed before we continue on to place the crafting material.

Next.

After nine months when the block was removed or harvested, we now have an opportunity to witness the histology in evidence.

Here we see a lower power and we're going to observe three different situations. One, the area of the notch where we can see new cementum, new bone and a new mature well vascularized periodontal ligament. If you look closely, you can see sparky fiber attachments on both the bone and the cementum side. So the area of the notch which was placed at the base of the calculus has responded appropriately.

Now, in the next observation we'll look at mid-root and then we'll look at the mouth of the defect.

As we move occlusally we again witness new bone, new periodontal ligament and a functional vascular periodontal ligament with clear evidence of sparky fiber attachments on both sides indicating its function.

Coming to the mouth of the defect, the new cementum has come all the way to the beginning of the bone defect and we have new bone and, once again, the functional periodontal ligament with supercrestal fibers that show very little evidence of any inflammatory infiltrate.

This completes the picture of that cuspid that we witnessed.

Looking at a second vertical defect just to demonstrate quickly that this occurred more than one time, we again observe a notch. We see in the notch new cementum, new bone connected by a functional vascular periodontal ligament, and in fact we have new cementum and new bone all the way to the top of the defect.

We have studied several different materials, but this astounding to see complete regeneration of the defect.

Next.

However, the most exciting observation that we encountered was the response in Class II furcations where there has been evidence to suggest that we fulfil the definition of periodontal regeneration with any of the materials that are presently available.

The notch designated the extent of the calculus and if we take the excerpt from the box, we see new cementum and new bone connected by a new functional periodontal ligament, again with evidence of sparky fiber attachments and no evidence of epithelium.

The outstanding observation in my estimation is that even at the fungus of the furcation there is no evidence of epithelium and we've completely resolved the definition of periodontal regeneration with new cementum, new bone and new periodontal ligament. This offers us the opportunity clinically to provide resolution for a problem that escaped clinicians indefinitely.

Next.

The results and conclusions of this human histologic evidence demonstrate safety; there's normal bone and ligament remodeling. The clinical measurements were demonstrated to be significantly improved. Radiographs were consistent with bone fill. We have no evidence of root resorption or ankylosis.

Actually, the histo micromophy that was performed very closely related in size or dimension to the radiographic analysis that was performed.

There is no evidence of root resorption or ankylosis whatsoever, so there is nothing to discuss along those lines. And the histologic evaluation we just saw revealed regeneration in both the intrabony and Class II furcation defects.

It became time to design a pivotal study. And in doing so, the transition was made to GEM 21S. There were two issues: One, select a carrier and the other to give some consideration to dosing.

Allograph was used for the histologic study. Since it's not formally approved by the FDA, a lot of questions -- a lot of producability remained and consideration was given to trying alternatives. Since beta-tricalcium phosphate and allograft were shown to provide comparable delivery properties of recombinant PDGF. The kinetics are similar and the BDGF release from both matrices simulated bone cell proliferation.

The study objectives were to compare the in vivo performance of PDGF with the two carriers, beta-tricalcium phosphate and allograft.

It was also to access the dose response of the recombinant PDGF.

The study designed is a randomized control blind trial and in canine this critical size periodontal defects. Six defects were made in each group and there is an eight week follow-up.

Looking at the results we can see that the beta-tricalcium phosphate by itself demonstrated new bone formation, but particles of the carrier remained and obviously it has left a significant portion of the furcation without periodontal regeneration. However, when the product GEM 21S is used the combination of the recombinant PDGF with the beta-tricalcium phosphate received a notch and received complete regeneration with cementum and periodontal ligament indicating a much more favorable response in the type of clinical end point goal that we would hope to achieve for our patients.

Next.

Evaluating the results of the canine study we see results with both TCP and allograft. And it's clear that the dosage of .3 mg/ml with the TCP and PDGF outperformed the other possibilities.

Next.

This led to the overall conclusions that GEM 21S, a truly synthetic system is safe and biocompatible with no risk of disease transmission. The BDGF when used with beta-tricalcium phosphate or allograft significantly improved the periodontal condition. This is measured in the formation of new bone, new cementum connected by a functional periodontal ligament.

There was sufficient evidence to now initiate a pivotal clinical trial and the decision was made to use the .3 mg/ml of BDGF because of the greater effectiveness that was shown.

And now I have the pleasure of introducing Dr. Robert Genco, the Director of the Periodontal Disease Clinical Research Center at the State University of New York in Buffalo. Bob has carried out five phase three and pivotal trials of periodontal products that have previously been accepted the FDA, so he's an old hand at it.

DR. GENCO: Thank you, Dr. Nivens.

And I, too, would like to thank the panel for your special efforts in reading that mass of material that was submitted to you. I have the file; files and files of those submissions and I know what a tremendous effort it is.

I would like to also thank the FDA, as Sam did, for their help during the design and analyses of the pivotal trial.

Now, I've worked in this area for about 15 years. One of the groups that Sam mentioned was the Buffalo group that looked at BDGF and other growth factors, and I have a tremendous interest in seeing this come to the benefit of society. And I'm very pleased to present this material today.

I have an official role with the company. I'm the Chairman of their Scientific Advisory Board. And longstanding interaction with Dr. Lynch.

I'd like to talk about the pivotal trial and share some of the results with you, the highlights the results. The next slide shows the nature of the trial. It was a double -- prospective randomized control trial with 180 patients randomized to three treatment groups.

Group one was Beta-TCP plus 0.3 milligrams per mil of recombinant PDGF beta subunit.

Group two TCP plus 1 milligram per mil of recombinant PDGF.

And then group three is interesting in that it's an active control. It's TCP with buffer, no recombinant PDGF. And it's an active control in that it's a product that's already on the market for bone regeneration used in orthopedics extensively. And we used a super fine fraction of that Vitoss. And it's a newly formulated form with increase porosity and increased surface area. And really it hadn't been systematically tested in periodontal disease. So, the design is really -- puts a high hurtle to show an adjunctive or additional affect of recombinant BDGF, and you'll see some of the results that bare that out.

It's a six month follow-up study. And we looked both at clinical and radiographic pinpoints.

The study was carried out in 11 centers, four university centers and 7 private clinical offices.

Next slide, please.

The investigators, patients, sponsor and monitors and radiographic assessment was all masked. The patients were randomized to one of the three groups by a variable block design, and all of the investigators; that is the examiners who were separate from the operators, separate from the surgeons, the examiners were calibrated, both at baseline and at six months to ensure inter and intra examiner standardization.

Next slide, please.

The study was independently monitored for quality and safety performed by Target Health, and it was independently analyzed by both Target Health and Averion, Dr. Phil Lavin's company, and he's here.

Next slide please.

The key inclusion criteria were: Age, 25 to 75 years; the pocket depth of the treatment site had to be at least 7 millimeters deep and had to have an intrabony defect at the time of surgery of at least 4 millimeters.

Any configuration of pocket was allowed. It could be 1, 2, 3 combination, combination with circumferential and combination with Class I or II furcation. So these are real live complicated complex intrabony lesions.

We allowed smokers who smoked up to one pack per day. The rational was that many, many patients who suffer periodontal disease are smokers. So we wanted to make sure that this was a treatment that would work in smokers who are known to heal less well than nonsmokers.

Next slide.

The key exclusion criteria included pregnant women or women intending to become pregnant during the study. This was not excluding women of childbearing age. Only those that were pregnant, lactating or intending to become pregnant.

History of oral cancer or HIV. Signs of acute infection or abscess at the site, the test site, Class III furcations, surgery under study too from the last year; all of these were exclusion criteria.

Next slide, please.

The outcome measures are very important to comment to. Clinical attachment level was assessed both at three months and at six months. Linear bone growth and percent of bone fill are quantitative measurements of radiographs. These were assessed as companion outcomes. And as we have heard from the previous presentations, in this complex disease, periodontal disease, the pathology involves both soft tissue and hard tissue, so it's reasonable from the clinical pathologic standpoint to assess both tissues, hard and soft for clinical outcome.

Then we also used a composite outcome where we blended or we merged, melded both the clinical and the radiographic technique. And the rationale for that is to get at this question of clinical significance. To try to address the issue of what percent of the target population benefitted from therapy. It wasn't meant to look at statistical significance to prove the efficacy. It was to get at this very difficult question. I know I was on the panel for a number of years. We always wrestled with the question of is significant. Did it benefit a significant portion of the target population? And that's why we used this composite outcome.

We also looked the pocket depth reduction, gingival recession and wound healing. And at the direction or suggestion of the FDA we compared it to currently approved FDA therapies that were sort of comparable.

Next slide.

This is the study time table. I draw you attention to day zero to the day the surgery was carried out. At least or less than two weeks prior to that baseline examination, examiner calibration and radiographs had to be made. At least two months prior to that the patients had to be screened, informed consent obtained and an initial preparation carried out.

After surgery the patients were followed at three months and at six months, radiographs were taken and all of the clinical measurements made both at three and six months.

The next is a videoclip of the actual preparation of the material. And this shows the dry material in a dappen dish and the PDGF solution added to it from the sterile syringe. And then the material is next. And this is done approximately ten minutes, at least ten minutes before the material is applied. You can see how the particles adhere to each other, and it's actually a very easily managed material to place in the mouth when you're having to place the material in upper lesions or mandibular lesions. It's actually quite easy to work with.

The next is a videoclip of the actual surgical procedure. And this is from one of the clinical sites. You can see the initial probing was carried out. And the -- we'll get some -- there we go. The videoclip shows the depth of the pocket. You see the tissue is quite firm after the initial preparation.

Then the root is thoroughly debrided. The issue is removed. All the granulation tissue is removed from the lesion. The lesion, you can see the dimensions here. It's a 3 wall intrabony defect. The root is cleaned absolutely clean. And then the root is treated with tetracycline to condition it. And then material is placed in the lesion to fill the lesion to the brim. And it's packed gently into the region.

You can see how easy it is to handle.

The operators, the surgeons were all standardized. They were standardized to a standard way of making an incision, incision design, to a standard debridement of the root, to a standard use of tetracycline concentration, duration.

And as you can see here, see the incision is a scalloped incision and we standardized the suture technique so that the buckle flaps could be opposed to get primary tension healing. Very, very important in these regenerative techniques to make sure that we get full coverage inasmuch as possible of the lesion with the soft tissue.

Now the examiners, as I mentioned, they were different than the surgeons. Different set of people. They were calibrated. They were calibrated to look for reproduceability of their own measurements, and that's inter-examiner calibration. And they were calibrated against a gold standard. Someone on the research team who had an intrinsic low error, all of the other examiners were calibrated against that person to ensure consistency across sites so we had more confidence to prove the data.

The next slide shows the actual results. The Kappa for the intra-examiner reproducability was 0.94 and for the inter-examiner consistency was .89. Both very, very high levels of reproducability and consistency exhibited by those Kappas.

Now the radiographic analysis. Care was taken in that also. For example, the films were taken under a uniform height quality field conditions using the renperil system, and every investigator's team was standardized to take these x-rays at a high quality uniform way.

Then the films were sent to a central site, University of Alabama, and Dr. Reddy and his team used standardized techniques and validated measurements which they and Dr. Genco for a decade had developed over the years to measure both linea bone growth as well as percent bone fill. And this is really percent linear bone fill. It's not a volume. It's a two dimensional measure.

And I'll show you those measures on the next slide. This is a graphic diagram of the radiograph and the landmarks that were measured at a synitho enamel junction, or if there is a filling, it would be the apical portion of the filling, restoration, root apex, crust of bone, base of defect at baseline, and similar measurements at six months. And then the next slide shows how these calculations were made.

First, linear bone growth is very simply the measurement from the CEJ to the base of the pocket at baseline, and subtracted from that is the measurement from CEJ at the base of the pocket six months later. Now in this instance, it turned out to be the original value of 6 millimeters, and at six months 3 millimeters, so we have three millimeters of linear bone growth. It's that simple, but very precisely measured.

Now percent bone fill-in is simply the linear bone growth divided by the initial depth of the lesion. In this instance, it would be 50 percent, so the original depth of the lesion is 6 millimeters and the linear bone growth was 3, so you had a 50 percent bone fill. It's a linear bone fill.

Now as a matter, because these are field x-rays, they're not taken with stents or any other precaution except good technique. There was a control on elongation or foreshortening, and that was the measurement of the CEJ to the apex, and that was measured on all x-rays pre and post. If they varied by 15 percent either way, then the x-rays were adjusted. They were normalized, a very standard technique used in radiographic analysis. We've used it for years and it works quite well. Now in fact that happened in less than 5 percent of the cases which test to the quality of the x-rays site by site. Next slide, please.

Now once the x-rays are sent to the site, then there's a whole other set of calibrations and measurement variability assessed; that is, the actual measurement of the x-rays at the site. The technician made repeated measurements on randomly selected cases, and there was a less than 3 percent variability between measurements, which is very good.

Following assurance, all radiographs were then looked at, reviewed by an independent periodontist. And, of course, all the radiographs are blinded anyway so nobody in Alabama knew which group they were from, but the independent periodontist who was not connected to the study looked at the x-rays and looked at the measurements to see if they made sense; were there any really odd-ball measurements. And they were occasionally some measurements that didn't -- so those were remeasured, so there was another level of control placed on the measurements. Next slide.

Now the results. I'll first summarize all the results in the next slide, and then go into them individually. First of all, there were no device-related serious adverse effects, an expected result, but it had to be proven. You're using two FDA approved products, put them together, both are safe, together they're safe, but it had to be proven.

GEM-21S significantly improved, that is statistically significantly improved CAL at three months. It significantly improved CAL gained between zero and six months. And the area under the curve assessment showed that the three month gain was maintained, it was a really accelerated healing which was maintained at six months. The LPG, that's the linear bone growth, was significantly improved at six months, as was the percent bone fill at six months, and these were highly significant in the .001 range. And the GEM-21S exceeded the benchmarks of effectiveness as compared to Emdogain, an FDA approved product, PepGen P-15, an FDA approved product, Allograft which is FDA allowed, not necessarily approved but it's allowed, and open flat debridement. Next slide, please.

DR. SHARMA: Excuse me. I want to just clarify one thing there. These results you're talking about, they are baseline to certain time point.

DR. GENCO: That's right.

DR. SHARMA: Not to compare it with different groups. Right?

DR. GENCO: I'll get into which group, yes. It's the .3 milligram group. That dose group showed these differences, and not the one. Right. But I'll get into that in some detail. I just wanted to give an overview, the result of my judgment and education; that Aristotle technique of tell them what you're going to say, say it, and tell them what you said. So I just told you what I'm going to say. Now I'm going to say it.

DR. SHARMA: All right.

DR. GENCO: The number of subjects were 180, 178 finished with a 1 percent drop-off rate, which is amazing for such a study. Forty-three smokers, mean age 51, gender slightly more males than females, approximately 60 percent Caucasian, the rest distributed among Asian, African American and Hispanic. Next slide.

Now baseline defect characteristics, the message here - there were no significant differences among treatment groups, and you can see this in the data, this inspect pocket clinical attachment level, bone defect, percent one wall, percent two wall, percent three wall, circumferential. They're all approximately the same, which you'd expect that random variation you get by randomizing. No statistically significant differences. And this is extremely important as we'll see later, because the deeper the pocket, the more healing you're going to get, so you really must have all the pocket depths at the baseline comparable. Next slide.

Now let's look at total adverse events. No significant differences among the treatment groups with respect to any adverse events, serious, potentially related, unrelated. For example, subjects with at least one adverse event ranged around 70 percent. Well, they all had surgery, and what was that adverse event; pain after surgery, which is to be expected. Not different between the surgical control and the other treatment groups, so there's no effect here of increasing adverse events by adding PDGF to the TCP. Let's look at the serious adverse events. They were present. They were not different among the groups, but they were present. Let's look at them. Next slide.

There were four serious adverse events, none related to the study device; including bronchitis, basal cell carcinoma, spinal fusion surgery, and diabetic complications. These are things as we all know in a six month study with 180 patients, you're going to get these adverse events - not related to the device.

Now let's look at some of the measurement data. One assessment is clinical attachment level gained over three months and over six months comparing the .3 milligram, the 1 milligram and the TCP, and you can see that the .3 milligram was statistically significantly different than the control at three months. That 3.8 millimeter gain was more or less maintained at six months. However, what happened, I think, is that the TCP control gained - and we saw this with the dog study too. The control actually catches up to the treatment over time, so now the difference between .3 milligram and TCP is not statistically significant. And we'll look at this another way looking at the area under the curve analysis. Next slide, please.

Now if we compare the gain of GEM-21 with the CAL gain of Emdogain and PepGen using the three studies for Emdogain that were submitted to the FDA and the two studies for PepGen, using those studies as a baseline, you can see that 3.7 CAL gain versus 2.7 versus 1.7, at least it's comparable, highly unlikely that GEM-21 under-performs, but at least they're comparable. Next slide.

Now one of the problems with such analysis, you have to really be careful as you all know, is that the studies were not done head-to-head. They are separate studies. We're talking about three, two, in our study five different, six different studies compared. And the possibilities for making misinterpretations are great.

For example, if you look at the baseline pocket depth, they're pretty comparable between our study and the Emdogain studies, but look at the PepGen study. They started out with shallower pockets, so the comparison with PepGen is fraught with difficulties, because they started with lower pocket depth so they're going to get less healing. And, in fact, that's what we saw. So we really have to be very careful about the interpretations compared to products on the market, and we are.

So what we say is they're comparable, very unlikely that GEM-21 under-performs relative to the others, so I think that's a conservative way of stating those comparisons. Next slide, please.

Now the area under the curve is commonly used in wound healing studies, and its purpose is to detect differences in CAL gain among the treatment groups between baseline and six months. We're using data from zero, three, and six months. Next slide, please.

And here are the curves. The green line is the 0.3 milligrams, and you can see it out-performs the other two groups at 3 and at 6. Our interpretation is that this is an early gain over the other two groups, statistically significantly different at .3 milligrams and the control, and then it's maintained. And if you look at the area under the curve, there is a difference between 0.3 milligrams and the other two, which is statistically significant at the 0.54 level. There is one subject here who suffered an abscess during the healing phase who lost four millimeters of attachment. If you remove that attachment that subject just becomes 0.033. However, we didn't remove that subject because it's an intent to treat analysis. Next slide.

Now let's look at radiographic linear bone growth as the companion. Here the .03 milligram per mil out-performs the 1 milligram per mil, and both out-perform the TCP alone in terms of radiographic bone analysis. And these P-values are very, very strong, even taking into consideration multiple variable comparisons using Yates Correction or other techniques. These P-values are extremely powerful. Next slide, please.

Now if we compare against current therapies again with the caveats I mentioned before, clearly GEM-21 is comparable to Emdogain - 2.5 to 1.1 and probably better than Allograft, certainly better than surgery alone. Next slide, please.

Now let's look at this radiographic percent bone fill. That's this derived ratio of linear bone growth as related to the original pocket depth. And here the mean percent bone fill is on the X-axis, and the three groups are depicted by the bars. Again, the green bar in the 0.3 milligram per mil out-performs both the 1 milligram per mil, and both out-perform the TCP alone, and the P-values again are quite low, showing high levels of statistical significance. Next slide.

And now comparing radiographic bone fill with the predicate products, you can see that GEM-21 is comparable to and probably performs better than the other products with respect to percent bone fill. But certainly it's comparable too. Next slide, please.

Now if we look now, drill down into the data and look at the various types of lesions we're treating, you know the one and two walls are very difficult to treat as to the more contained three-wall and circumferential. And you can see that in the data. You look at all of these bars for the one and two-wall are lower than all of these bars for the three-wall and circumferential, so in general, these lesions heal better than these; yet, the 0.3 milligram per mil GEM-21 gave 50 percent bone fill in over half the subjects in those very difficult to treat one and two-wall lesions. Again, out-performing the 1 milligram per mil and out-performing the TCP. And if you look at the three-wall and circumferential defects, 65 percent of the lesions were filled, or the lesions were filled 65 percent of the bone with the 0.3 milligram, as compared to 34 and 21 for the other one control. Again, the .3 milligram out-performs the TCP and even these defects that heal on their own.

DR. SHARMA: Is this all this radiographic data at three months or six?

DR. GENCO: Well, it's at six months. All the radiographic data is at six months. The CAL data is at three and six. Now the reason the radiographic data at three months is not used is because the material is in the lesion at three months. You can see it on the radiograph. And from the histology, which we've done extensive histology both in man and animals, it's usually gone by three months histologically, you can't see it any more. So we felt safe in looking at the six month x-rays. Next slide, please.

Now this is a distribution of the cumulative proportion of bone fill and the curve to the right, right is better, left is worse. You can see the curve to the left is the control, and if you look at the proportion, let's say 50 percent of the subjects with control, 20 percent bone fill. In other words, in 50 percent of the subjects given the control, you got 20 percent bone fill, not very good. And 33 percent and 50 percent of the subjects given the 1 milligram per mil, they got 33 percent bone fill, but in 50 percent of the subjects with .3 milligram, we got 50 percent fill, so half the subjects, half the pockets were filled with .3 milligram, 20 percent of the subjects or half the subjects, 20 percent fill with the control. Again, statistically significantly different. This starts to address the issue of clinical significance. What proportion of the population actually benefits from this? It's not meant to be the statistical proof for efficacy, but some indication of how many people in the population are actually benefitting from this treatment which addresses clinical significance.

Now the rationale for the composite outcomes is just as I said, to get some indication of clinical significance. So we use the two primary end-points, CAL and bone, and then made a composite. And this is done in rheumatology and other areas of wound healing. As a matter of fact, we're carrying out a study of cardiovascular disease and we're using a composite of six different cardiovascular variables, so composite variables I think are gaining in attention in the clinical trial methodology area, and are extremely useful when done properly. Next slide.

So how do we define a successful outcome? Well, what we did is we took the PMAs for Emdogain and PepGen and took their best results, and we said all right, that's the attachment gain achieved by either Emdogain or PepGen, and the best bone fill or linear bone growth will accept as the cut-off point. Okay. So we put that together. The best attachment gain for Emdogain or PepGen was 2.7 millimeters, and the best bone fill for either was 14.1, so we put those together. That's our composite. If you reach that, we define that as "success". Same for CAL and LPG, 2.7 millimeters and 1.1. Let's see what the results are.

Now look at this overall. Success was achieved with .3 milligram dose using one of those composites in 70 percent of the subjects, and with the other composite, it's 60 percent of the subjects. Now did the percent success out-perform the control? Yes, but that's not the intent here. The intent here is not the statistical significance, although it was highly significant, but what percent of the population achieved this definition of success? We're quite pleased that 60 to 70 percent of the population benefitted from this product. Next slide, please.

Now let's look at one of these lesions. This is a lesion that after breaking the code, we know got the .03 milligram per mil, a 44-year old female with a lesion above the two-wall defect which was circled around the lingual with a class two furcation, and you could see the defect on the distal. You could see it here. And if you look now at the six month x-ray, this is new bone. This not TCP, this is new bone. The TCP is little particles that you could see on the x-ray, very characteristically different than the new bone. So when you make the measurements from the CEJ to the base of the defect, in this instance it's something like 6 millimeters, and in this instance it's something like 3, so we had about a 50 percent fill, just rounded off with a linear bone gain of about 3 millimeters, so this is a typical result of 50 percent fill and 3 millimeter bone gain.

Now as Dr. Nevins mentions, this markedly changes the prognosis of that tooth. The lingual furcation is filled and the distal lesion is pretty much 100 percent healed, and the mesial lesion is about 50 percent healed. This is a very good result clinically. Chances are we've gone from a 6 millimeter, 7 millimeter pocket to a 3 millimeter, 4 millimeter site. That's maintainable. Next slide, please.

So in summary then we had 180 patients who were fully masked in this perspective multi-center trial. Quality was assured by multiple mechanisms, including CRO, blinded investigators, blinded design, arm's length statistical analysis by CRO and another firm. There were no device-related serious adverse events. The GEM-21 has statistically improved CAL at three months, and the CAL under the curve between zero and six months, and the interpretation is it's rapid healing which persisted at six months that was induced by the .3 milligram per mil growth factor.

The linear bone growth was improved at six months as was the percent bone fill at six months in a highly statistically significant manner. We feel that the GEM-21 exceeded benchmarks of effectiveness but for caution we'll say it was comparable to the benchmarks, very unlikely to be less effective than already existing products on the market. Next slide, please.

Now let's look again at the comparability to the approved products, the GEM-21S, CAL gain 3.7, radiographic fill 2.5 are in the ballpark if not better than the other approved products. Next slide.

So overall then, we feel that GEM-21S, a fully synthetic and safe product. Take it off the shelf, don't have to worry about contamination with Bovine contaminants, or with Allograft problems, although that's not a major problem, but it is an issue in patient's minds and some clinicians, so this is fully synthetic and safe as a known mechanism of action demonstrated by over a decade and a half of very intense high quality research a mechanism of action of PDGF. And on the characteristics of the Vitoss, and in fact the Vitoss is a new product developed in 1999.

The recommitant PDGF-BB component enhances periodontic regeneration in animals and humans, and this is very reproducible result seen in many species. I personally have been involved in three dog studies, and they all show the same thing; complete fill of Class 3 furcations. Accelerates attachment level gain and radiographic evidence of bone regeneration, quite well documented in human study I've just mentioned, and demonstrates a favorable risk to benefit relations, so I would say in general, to sum up, my view is that this product gives a favorable clinical result in 60 to 70 percent of patients when used as indicated with very few side effects that we're not expecting. Thank you very much for your attention.

DR. LYNCH: Thank you very much, Dr. Genco, Dr. Nevins and Dr. Giannoble for those presentations. We had asked the FDA, Mr. Adjodja, Dr. Runner for just a few minutes at the conclusion of the presentation just to be available to answer any burning questions on the methodology. We don't want to pre-empt the discussion this afternoon. We understand there may be some more global questions. I think those might be more relevant for this afternoon, and I think that's where that discussion is planned, but we didn't want to leave any -- if there were any lingering burning questions on methodology or specific aspects of the presentation to let those sort of fester in your mind. So we'd be happy to again entertain any specific burning questions relative to methodology that you might have now, or proceed forward and we can address further questions this afternoon.

CHAIRMAN SUZUKI: Before we begin with the questions, I'd like to remind the audience that I'd like you to reserve any questions regarding the particular hearings until after the presentations. And then secondly, the FDA panel members will have the prerogative of asking questions first, including procedural questions. And I wanted to thank the presenters and the sponsor for presenting such a precise presentation and keeping us on time.

As Chair, I'd like to take the prerogative of asking perhaps the first two procedural questions, and that is with respect to the radiographic benefits - and I know Dr. Genco mentioned looking at the composite outcomes in total, but focusing in on the radiographic interpretations, I notice that there is a mean improvement of about 2.1 millimeters. I'd like an explanation as to why you think that this is clinically significant.

DR. LYNCH: Okay, certainly. And I'm going to moderate the session and probably defer that to many of our panel members since your question specifically refers to the clinical relevance of the radiographic bone gain. I think it's appropriate that the clinicians and the panel answer that, and I might refer to Dr. Genco and to Dr. Nevins.

DR. GENCO: The radiograph especially at six months under-estimates the healing. You saw the attachment gain was more like 3.8 with a bone gain for the radiograph of 2.1. I think that reflects the under-estimate. 3.8 millimeters in a 7 millimeter pocket, it's a 4 millimeters gain, 7 millimeters to begin with, you're down to 3 millimeters. And I think we and others have done studies showing that if a pocket is 5 millimeters or greater, it has a 6 to 7 fold greater chance of losing attachment in the next two years, so if you can get it below 5 or 6, that bodes well for the future. And I think this is what this study has shown, that the pockets are reduced from 7 to approximately 3, 3 and a half. The bone doesn't quite reflect that because it under-estimates the healing, but the pocket reduction and the attachment gain I think really are telling from the clinical standpoint.

CHAIRMAN SUZUKI: Okay. Thank you. My second question is a procedural one regarding the surgery itself. In the video clip that we saw, there were just a couple of procedure questions that I had. The first is that in your presentation of materials, you indicated that you use Tetracycline of the preparation of the root surface, but the video clip did not show that. Is there a reason why it was omitted or was that standardized?

DR. LYNCH: Yes, and Dr. Giannoble, who was an investigator, could comment, but I could certainly comment on that one, as well. It was omitted from the video clip simply to make that a very concise video clip for no other reasons. It was standardized as to the amount of Tetracycline that was used and the duration of the root conditioning so that was all pre-specified and the examiners or the surgeons were trained on that.

CHAIRMAN SUZUKI: Okay. The last question I had about the procedure is that frequently surgeons fenestrate the osseous lesion, and I noticed in the video clip that you did not. Was there a reason for selectively not using that particular step?

DR. LYNCH: Again, Dr. Giannoble, if you want to come up and comment, you could certainly feel free to. We did, I believe, allow fenestration at the discretion of the investigators, a very hard cerotic bone, the bony walls there. It wasn't necessary in the particular case that you saw, but if in the judgment of the investigator that the bone is very cerotic and sort of avascular, it was permitted for them to do perforations.

CHAIRMAN SUZUKI: So that portion of the surgical procedure was not standardized in terms of the fenestration.

DR. LYNCH: Yes, we felt that clinically certainly not all cases would require it, again as the case that you saw, but we felt like certainly some would. So again, in the investigator meeting prior to study initiation, we discussed certainly this very point, and the investigators that we chose, of course, are all very, very highly regarded, very seasoned clinicians, and they felt like you couldn't predetermine that all the lesions should have perforation because it was necessary in all the lesions, so we discussed the parameters or characteristics of the lesion that would require perforation and left that then to the surgeon.

CHAIRMAN SUZUKI: Okay. AT this time I'd like to ask the panel if they have any points of clarification that they would like answered. And before you do so, I'd like to ask that you identify yourself in the microphone for the transcriptionist, as well as presenters identifying themselves into the microphone, as well, before you respond. Okay. Dr. Cochran.

DR. COCHRAN: David Cochran. Dr. Lynch, I'd like to ask a couple of questions. First of all, in the documentation you provided for us, you used a couple of papers to reference for linear bone growth and percent bone fill. Dr. Genco talked about using the PMAs that were submitted prior for these other products. How did you go about choosing those, particularly there was a study from Greece, and then there was Rutger Persant was another one that you used.

DR. LYNCH: I'll turn this also over to the rest of the panel, but wherever there was data available from the PMA submissions, specifically the summary of safety and effectiveness for previously approved products, we utilized those, so to translate that means Emdogain and PepGen P-15. As has been mentioned, Allograft has never been formally "approved" or cleared for dental uses at any rate by the agency and so, of course, there are no formal FDA submissions that were available, so we did go back through the literature and did a very extensive literature search on specifically looking again at radiographic assessment of bone fill following Allograft treatment. And we used what data was available in the literature. As Dr. Genco mentioned, we're certainly not by any means claiming superiority to any of those materials that were used. We were just trying to, and at the agency's request, get some comparison of the effectiveness of the results seen in this trial benchmarked against other materials that the clinicians are using.

We also utilized or carefully reviewed the paper out of the San Antonio group that clearly shows that the radiograph assessments do under-estimate bone fill as compared to re-entry assessments.

DR. COCHRAN: That was a good choice. The second question is what are your thoughts on comparing some of your results to the enamel matrix proteins being that that's a protein-only therapy, and you're using protein plus graft material. Would you comment on that?

DR. LYNCH: Again just to stress that the comparisons that we did were simply to compare the results to other benchmarks of effectiveness that were available. The TCP that's used as a carrier, as has been mentioned, is fully resorbed within about three to four months and, therefore, we did not think that that would affect, for example, the radiographic assessment at six months, as was mentioned. That's one reason we did not do the radiographic assessment at three months.

Given that the PDGF is clearly gone by six months, given that the matrix is, as far as we can determine, certainly mostly, if not totally, resorbed at six months, as well; that would make that site then somewhat comparable, and I don't want to overstate this to the Emdogain where there was no matrix observed.

We know in clinical practice that many people do mix Emdogain, as again has been investigated in your university, with other bone substitute materials to try to contain it in the bone defect. And I think that that was one of the rationales for using a matrix in our product, was to provide the clinician a standard matrix that they could use to mix with the recombinant protein, as compared to Emdogain where the clinicians are often just kind of taking whatever they have on the shelf, so to speak, or whatever grafting material they like and mixing it with the Emdogain, so we feel this provides a more standardized product.

DR. COCHRAN: The last question would be in the documentation there's a product mentioned called Vitoss Plus, which is a similar product or an approved product. What is that?

DR. LYNCH: It's just a different name that -- sorry for the confusion there. Actually, the names of this product that we're reviewing today have sort of transitioned from Beta TCP Plus, at some point I think in the documentation it was called Vitoss Plus, now it's called GEM-21S. It's the same product.

CHAIRMAN SUZUKI: Dr. Amar.

DR. AMAR: Salomo Amar. I'm going to ask a more general question. Dr. Genco mentioned that at a certain point the control catches up with the rest of the experimental. And my general question, if at six months the control or the experimental -- the control catches up with the experimental, what would be the added benefit of using this molecule as compared to TCP? Is it just for the early reading improvement parameters or are we talking about long-term maintenance?

DR. LYNCH: Well, Dr. Genco, why don't you come up here and I'll provide my interpretation, but I would like -- I think Dr. Amar would like to hear your's, as well. I think what Dr. Genco was referring there was for the clinical attachment level of the soft tissue. Certainly, the bone fill as measured radiographically never catches up in the control versus the treatment group, so there's always a strongly significant improvement or benefit in the radiographic bone fill.

DR. GENCO: That's one point. We didn't see that phenomena in the bone. Of course, we only looked at 1.2. With respect to early healing it's, of course, benefit to get that healing pretty much underway in the first three months, and you can get on with the rest of the therapy. We think that if you're involved in a complex case that requires implants and other treatment, that to have this early result at three months is of great benefit, so it's an accelerated treatment that fits in with the treatment of advanced case, and it's a definite benefit.

DR. AMAR: If the clinical attachment level is the primary outcome and it catches up, aside from the bone failing, it looks pretty similar.

DR. GENCO: Well the point is if you -- let's say if you have to put a crown on the tooth, you could start putting the crown on in maybe two to three month rather than waiting six.

DR. AMAR: The other question that I had is regarding the resorption. I saw sections by Dr. Nevins, at nine months I believe on the furcations showing probably some deposit of Beta Tricalcium Phosphate.

DR. GENCO: Myron, do you want to --

DR. NEVINS: If you're referring to the --

CHAIRMAN SUZUKI: Can you identify yourself, please.

DR. NEVINS: I'm sorry. Myron Nevins. If you're referring to the human histology, we didn't use Beta Tricalcium Phosphate. That was an Allograft study. The Allograft was the carrier for that. The only thing I showed with Beta Tricalcium Phosphate was the one slide at the end on a K-9 study, and that was at eight weeks. And on the GEM-21 there was no evidence of Tricalcium Phosphate at all in the control which was the Tricalcium Phosphate by itself. There were pieces of Tricalcium Phosphate.

DR. COCHRAN: To follow-up on that, I think in that result didn't you get 70 percent regeneration with the TCP alone in that K-9 study?

DR. NEVINS: I would have to --

DR. COCHRAN: It's 37 percent, not 70.

DR. NEVINS: It's the blue column over TCP.

CHAIRMAN SUZUKI: Okay. Any other questions, Dr. Cochran? Okay. Any other questions from the panel? Dr. Sharma.

DR. SHARMA: According to the protocol, this is Inder Sharma. According to the protocol, the primary comparison was to be between the high dose and the control. Only if it was significant, then 0.3 which is low dose, was to compare with control. But I see the focus of presentation have been mostly on the low dose, so I'm wondering what happened that we are now focusing on what we said in the protocol that this will be about primary comparison, because primary comparison is not significant whether you look at three months or six months.

DR. LYNCH: I think you're looking at an older version of the protocol. There were formal amendments that were submitted to the agency and approved by the agency throughout the study for various things that we had under discussion with them, so the primary comparitor was the .3 mg/ml dose level versus the TCP control.

DR. SHARMA: The second question I have is about the composite end-point. Was this a pre-planned comparison or was it decided to do that after the fact?

DR. LYNCH: I'm sorry. I'm not sure I understand your question.

DR. SHARMA: The composite end-point.

DR. LYNCH: Oh, the composite. Okay.

DR. SHARMA: Was it a pre-planned comparison using composite end-point, or was it later decided to be --

DR. LYNCH: I think Dr. Phil Laven will address that. He's our biostatistician.

DR. LAVEN: Hi, Philip Laven, biostatistical consultant to Biomonetics. My company is Avarion. When we came up with the idea for the composite end-point, it was done before the database lock, and it was planned at the same time that we planned looking at the AUC. That end-point was reflective of the fact that we knew that the disease was more extensive than just looking at the delta CAL measurement, and that the composite end-point had to address both the radiographic end-points, as well as the clinical end-points, so this was all done prospectively before the database lock, but was not in the original protocol.

CHAIRMAN SUZUKI: Dr. Zero.

DR. ZERO: Domenick Zero. I have a question about how the statistical analysis was done, although that's not my main expertise. On looking at the distribution of females, smokers, African Americans, and the CAL values, there are although not statistically significant differences, there are some numerical differences that are noticeable just in looking at the different groups. The data reported, was that an adjusted statistics, or was that just the raw statistic?

DR. LYNCH: Dr. Laven.

DR. LAVEN: Yes, Philip Laven. Those gain that you are seeing there in the report are unadjusted statistics. At the request of the FDA over the last month, we did prepare additional analyses where we did look at controlling for those factors and looking for treatment interactions with factors like smoking, the location of the tooth, whether it was a molar or not, and we did assess those analyses. And those analyses, just to give a sense for where they turned out, there was no treatment interaction with any baseline co-variates, so the treatment advantages that you're materially seeing there, although they're uncorrected, do represent the state-of-the-art for what happened in those groups.

DR. ZERO: Thank you.

CHAIRMAN SUZUKI: Dr. Zuniga.

DR. ZUNIGA: John Zuniga. A couple of very simple, hopefully, questions on the study protocol, just more for clarification. I notice in your management of your post-operative patients, you included the use of NSAIDs for any analgesia. Was that a -- why did you do that, and do you have concern about NSAIDs in this product?

DR. LYNCH: No. Sam Lynch. There are no specific concerns about NSAIDs related to this product. We just knew that NSAIDs as a general class of drugs had ability to affect wound healing, and we didn't want some patients to be on NSAIDs by some investigators, and other patients not to be, because we thought that that might affect the -- especially the immediate post-op healing, and we did have an end-point that looked at wound healing over that first three week period, so in order to standardize that regimen we just elected not to use NSAIDs.

DR. ZUNIGA: And has that been explored using NSAIDs?

DR. LYNCH: It, again, was not in the pivotal clinical trial. We would have to look at the patients in the human histologic study to see if they were given NSAIDs or not.

DR. ZUNIGA: And then the second question, again relative to the study protocol, is the antibiotic use and some of the post-operative instructions for the patients. Were there any patients that did not comply with the antibiotics? And if so, were there any difference in effects? And then finally, you have pretty strict protocol for soft foods and diet eating on the other side. How compliant was that regarding effects on the patient outcomes?

DR. LYNCH: Right. I think as is relatively customary, post-op instructions were given to these patients. I don't believe that there were anything unusual about our post-op instructions compared to many that we give our periodontal patients. In terms of any specific, like protocol violations that were reported where the patient reported chewing on the site of the surgery or that kind of thing, and I don't know - Mark, do you care to comment on that? I don't think there was any -- there wasn't certainly any significant violations along that line. There may have been isolated cases.

MR. CITRON: No protocol violations.

DR. LYNCH: Okay. So there were no protocol violations.

CHAIRMAN SUZUKI: I have a couple of final questions. With respect to the dose, .3 versus 1.0 milligrams per mil, were dose response curves completed prior to your selection of these doses, or were these doses taken from the literature? And secondly, why isn't more better?

DR. LYNCH: I think the second half of your question, Dr. Suzuki, we might want to table to this afternoon. It's certainly a very excellent question, and I don't mean to put it off. We could address it here, but for broader questions in terms of that, we might defer those to this afternoon at your discretion.

In terms of how we selected the .3 and the 1 mg/ml, as Dr. Giannoble showed one slide from the study by how co-workers at Harvard and UNC several years ago, that study utilized .05 mg/ml, and a .15 mg/ml of PDGF. And then that study used also accommodation with the insulin like growth factor, and it showed that there was no effect, no beneficial effect at the .05 mg/ml. There was a significant beneficial effect on bone similar to what we frankly saw in this study at the .15 mg/ml, so we used that as information. We also then conducted a canine study that looked at the .3 mg/ml, and you say well how did you get from .15 mg/ml in that initial clinical study a few years ago to .3. And the rationale, right or wrong there was that because that initial study utilized a combination of two growth factors, we felt like we might need to utilize the total growth factor dose, if you will, and so that would be .3 mg/ml. So that was the justification for the low dose in our pivotal trial. And the justification for the high dose was taken just as a XXX multiple from that. And as was reported, I believe by Dr. Nevins, we did see absolutely consistent results in the canine study that the .3 mg/ml provided the most beneficial response, so that was again the reason for determining that that was our primary comparitor, was based upon the canine study.

CHAIRMAN SUZUKI: Thank you. John Suzuki again. My last question is with respect to the T-inclusion criteria, and the age range of your periodontitis patients were from age 25 to 75, I believe. In the submitted materials, you indicated that the aggress of periodontitis patients were excluded from this patient group; yet in the oral presentation that was not in the particular slide. Is there a reason for that?

DR. LYNCH: Sam Lynch. That was just omitted off the slide just for sake of brevity, and we couldn't include all of the criteria on the slide. But certainly, patients that were considered to have aggressive periodontitis, what we used to call juvenile periodontitis, were excluded from the study.

CHAIRMAN SUZUKI: Dr. Amar.

DR. AMAR: I have just one more question. You're going to do x-ray analysis and Dr. Genco mentioned that there was no stent, am I correct? There as no stent. And what was the percentage of elongation accepted, and you mentioned 15 percent. Am I correct?

DR. GENCO: I can start the answer. Bob Genco. Maybe Dr. Reddy can continue. But what I presented was that the elongation -- or for shorthand, 15 percent, it was seen in less than 5 percent of the x-rays. Is that the question?

DR. AMAR: I guess my question is if it's really 15 percent, the cut-off value on a tooth that's -- a root that is say 10 millimeters, 15 percent is 1.5 millimeters, that the effect size - it's about the effect size that we would see on bone fill. And I would have some kind of concerns about that.

DR. GENCO: Well, that was used to then adjust the x-rays to normalize.

DR. AMAR: So there was no more than 5 percent elongation.

DR. GENCO: Well, let's let -- well, 5 percent of the cases exceeded the criteria of 15 percent elongation or foreshortening. Therefore, required normalization, so the extent to which there were over 15 percent - I think that Mike could answer that. Dr. Reddy.

DR. REDDY: I think I understand what Salomo is asking. Hi, Michael Reddy. I'm an investigator from the University of Alabama. I did radiograph analysis. You want to know what percent -- what was the highest range of elongation and foreshortening. And I have to look at the database to tell you exactly, but some of them were up to about 25 percent, but there were very few x-rays. Remember, this was an intent to treat analysis, so we simply couldn't say that that didn't make our radiographic criteria, so we included them, and then retrospectively corrected it mathematically. You have to remember that a 15 percent increase in the overall root length, which may be 15 millimeters in length, may just vary the measurement of the bone growth by about 10 percent even if you didn't correct for it, even though we did correct for it. So if you have 2 millimeters of bone growth, you're really only going to change it to about 2.2. But the case that did happen, we did mathematically apply a formula and run an algorithm to correct those. Those are very few. We had a great fear, the same fear you had at the start of the study. That's why we incorporated that into the analysis, this 15 percent cut-off because we were afraid that these are field x-rays, and we may have 50 percent of them with elongation and foreshortening, and it turned out that they were actually very clinical radiographs. Of course, again, we had to get a good x-ray of one tooth --

DR. AMAR: So if I understand you correctly, there was an area of elongation about 15 percent of max that was corrected.

DR. REDDY: No, only if the area was over 15 percent was it corrected. There were some sites that were 25 percent foreshortened or elongated.

DR. AMAR: That could translate into if the root is 10 millimeters into 1.5 millimeters change?

DR. REDDY: It could if it wasn't corrected for, and that's the reason why we put the correction in, exactly.

DR. AMAR: Can you elaborate on the correction?

DR. REDDY: Yes, the correction simply to go ahead and correct everything back to the baseline. We consider whatever length we measured at the baseline from CEJ to apex as the gold standard. If it differed by more than 15 percent, we used a ratio of the original CEJ to apex measurement to new CEJ to apex measurement to mathematically correct all measurements that were subsequently taken, so we wouldn't lose the data.

DR. AMAR: Thank you.

DR. SHARMA: I have one.

CHAIRMAN SUZUKI: Okay. Dr. Sharma.

DR. SHARMA: Inder Sharma. I have one final question. Interim analysis were planned for the study and I was wondering where they conducted? And if they were conducted, was there a DSMB or who had access to those results?

DR. LYNCH: Would you mind repeating the question, please?

DR. SHARMA: The interim analysis for the study --

DR. LYNCH: Interim analysis --

DR. SHARMA: They were planned, and my question is if those interim analysis were conducted, and who had access to the results, was it a DSME independent diversity of monitoring the worker?

DR. LYNCH: Yes, I understand. Sam Lynch. There was an interim analysis conducted per the protocol. The analysis was conducted on the first 90 patients to complete three-month follow-up. This was, again, an analysis that was agreed upon with the agency. It was done in a fully blinded fashion by the independent clinical research organization, the CRO that was responsible for monitoring the study, so there certainly was no breaking of the blind or anything.

The only data that we got back was that and the reason the FDA had asked us to do that interim analysis was to one of sample size, should we adjust the sample size at that point? Do we increase the number of patients, because we had agreed not to decrease the number of patients because we did not want to take a statistical penalty for the interim analysis, so we had all along said we're going to do 180 patients, even if that result was just fantastic. But the question was would we need to add patients at that point? So that the answer that we got back and which we provided the agency was based on the statistical analysis performed by the CRO, there was no need to adjust the sample size.

DR. SHARMA: But for the final analysis, when you do interim analysis, there is adjustment for Alpha to accommodate the analysis done earlier. And I don't see that in the final analysis.

DR. LYNCH: That was an issue that we discussed with the agency very carefully. I'm not sure if we should address that now or if we should wait and let the agency have their time and present that.

DR. RUNNER: Well, I think when we have our statistical presentation, that may be --

CHAIRMAN SUZUKI: This is Susan Runner speaking.

DR. RUNNER: I'm sorry, Susan Runner. When we have our statistical presentation, that may be addressed.

DR. SHARMA: Okay. Thank you.

CHAIRMAN SUZUKI: Okay. If there's no other questions we'll take a 15 minute recess.

(Whereupon, the proceedings in the above-entitled matter went off the record at 10:31 a.m. and went back on the record at 10:47 a.m.)

CHAIRMAN SUZUKI: I'd like to try to stay on time and begin the next part of our session and to begin with, a presentation by the FDA on GEM 21S and our first presenter is Dr. M. Susan Runner, Chief, Dental Division Devices Branch and Deputy Director.

Dr. Runner?

DR. RUNNER: Thank you. This morning I'd like to have FDA give you some input or give you some input on our feeling about GEM 21S and this morning we'll start with the presentation by Ms. Angela Blackwell, who is a biomedical engineer in the Dental Devices Branch, who will review some general information about information submitted in the PMA and go over device characterization and some submitted pre-clinical studies. Then I will go over some of the issues with the study protocol, clinical results and submitted device labeling and then Ms. Judy Chen from our office will also go over a brief statistical analysis.

And just one additional comment, FDA has received some additional information from the company which was not included in your panel pack. However, we will not be discussing that information today. It will not be discussed by the company or FDA because it has not had time to be reviewed by the Agency, but that's just for your information.

MS. BLACKWELL: GEM 21S is a combination product which consists of a device bonding material beta tricalcium phosphate and a drug, recombinant human platelet growth factor. This product is regulated as a Class 3 PMA product. GEM 21S is intended for the management of interosseus periodontal defects. Beta TCP is classified for dental indications in 21 CFR 872.39.30. For dental indications beta TCP was viewed as a drug by the FDA for the device amendments. After the adoption of the device amendments, it was transferred to CDRH as a transitional device. This made it automatically Class 3 and subject to PMAs.

Beta TCP for dental indications still requires a PMA but a proposed rule to reclassify beta TCP and other bone filling materials to Class 2 was published on June 30th, 2004. Under the Orthopedic and Rehabilitation Devices Panel beta TCP is regulated as a Class 2 device.

rhPDGF is regulated as a therapeutic biological product in our Center for Drugs and requires a biological licensing agreement. Prior to submission of the preliminary protocol, it was determined that this combination of product would be regulated under a PMA and that CDRH would be the lead review center. CDER would act as the consultant center.

The beta TCP component in this submission is regulated for orthopedic indications as Vitoss. Other beta TCP products are regulated for dental indications under PMA regulations. In addition, when beta TCP is combined with other dental bone filling materials, it's regulated under 510K.

Dr. Runner will now review the PDGF pre-clinical data and the clinical data.

DR. RUNNER: I'd like to go over some of the safety data that was submitted to FDA that indicated that GEM 21 should be safe for the stated intended use prior to the initiation of the clinical study. The previously submitted data submitted by the sponsor relating to recombinant PDGF consisted of in vitro data, animal data and data form human feasibility studies as well as data from the diabetic fool ulcer study for the Regranex drug product. The in vitro data, as you've heard, consisted of biocompatibility studies, studies of the effect of PDGF on cultured cells and studies of PDGF released from grafting materials. The animal studies, as you also heard, consisted of the evaluation of the effects of recombinant PDGF on bone healing in several animal models.

The clinical studies as was also reviewed by the company consisted of human feasibility studies as well as the pivotal studies related to the Regranex product. In addition to review of all of these information prior to the approval of the clinical study, FDA did a review of our adverse event data bases. This was for both the drug product, Regranex and as well as for the device component of beta TCP. The review of these data bases revealed that the relevance of the data in the data bases was of little significance or questionable significance relative to this product. As you heard from the sponsor, previous effectiveness data from preclinical and feasibility studies have indicated that this product may effective for the stated intended use. Based on these safety data and effectiveness data, FDA approved the IE study for GEM 21S.

Just to go back a little bit, you've heard a little bit about Regranex. The product was cleared for diabetic foot ulcer treatment and the data came from three large Phase 3 clinical trials using the product of rhPDGF in a vehicle. The product, as you heard, is known as Regranex. The primary end point for these diabetic foot ulcer studies was the percent wound closure after 20 weeks of daily application of approximately .1 mg/ml rhPDGF. Regranex was found to be safe to use and was approved but the wound healing results were not statistically significantly better than good foot care alone.

As you heard, the sponsor submitted a pivotal clinical study to try to assess the safety and effectiveness of GEM 21S in the management of interosseus periodontal defects. The study hypothesis, as you also heard, was that GEM 21S promotes greater soft tissue and bone regeneration as measured by clinical attachment level and radiographic bone measurements than beta TCP alone. As you also know, there were three treatment groups, a low dose, high dose and a control group.

The measurements, as you also heard, included pocket depth probing, clinical attachment level, gingeral recession and radiographic measurements of linear bone gain and percent bone fill. The primary end point, as stated in the protocol, was the change in clinical attachment level at six months. The sponsor retrospectively added a change in clinical attachment level at three months. In your discussion this afternoon, was will want your input as to the validity of retrospectively adding an efficacy end point of this type to the data analysis. The sponsor also had secondary end points including linear bone gain, percent bone fill, probing pocket depth and such that has been previously described. These secondary end points, as you know, are used in many bond graft, in studies reported in the periodontal literature and are important parts of this statistical analysis of the clinical study.

The sponsor also had secondary end points comparing current data to historical data and those are also being more frequently used in bone grafting studies reported in the periodontal literature. The numbers that are used for both Emdogain and PepGen P15 are a means derived from previously approved PMAs. Many of you may know, we've heard Emdogain and PepGen P15 thrown around a lot today. Emdogain is not a bone grafting material but a gel derived from porcine tooth buds that is applied to root surfaces while PepGen P15 is a bone grafting material containing a synthetic biological response modifier.

The linear bone gain and percent bone fill numbers here are also averages derived from the literature. FDA, again, would like your input as to the clinical relevance of the results reported from these end points. In addition to the three-month retrospective analysis end point, the sponsor also added two composite study outcomes and an area under the curve analysis. Both of these analyses were a combination of a clinical and a radiographic end point and tend to be a composite of clinical and radiographic results. Again, FDA would like your input as to the clinical relevance of the results of the area under the curve analysis and the composite analysis.

These next tables show the results of the statistical analyses and I'd like you to please focus your attention on the last column to the right where it lists statistical significance for the analyses performed. The first table here compares that low dose to the control group and the next line will compare the high dose to the control group. Please note that the results for the primary end point were not statistically significant. Please note also which end points were perspective and which end points were retrospective as well as which are statistically significant and which are not.

We have a prospective at the top and retrospective at the bottom as you can see. This slide similarly compares the high dose versus the control. For the high dose the only significant results were for linear bone gain and percent bone fill. Again, please note which end points are prospective, which end points are retrospective as well as which are statistically significant and which are not. This table sort of gives a summary of the statistical significance for the study and please note that three statistically significant prospective results were secondary end points.

FDA would like you to discuss the clinical relevance of the results reported for these statistical end points as well as the possible importance of the non-significant end point data. The specifics of the statistics will be discussed by Ms. Chen in a little bit. In terms of the safety end points, as you heard, the primary safety end point was the number of adverse events and as you also heard, most of the adverse events reported in the GEM 21S study were associated with the surgical procedure itself not attributable to GEM 21S and this is consistent with other periodontal studies.

In the device labeling, the package insert claims these items and the primary device labeling claim was that the GEM 21S was shown by both clinical and radiographic measures to be effective in treating moderate to severe interosseus periodontal defects within six months of implantation. I think it's important to note that in this PMA the sponsor has expanded its indications for use to include deficient alveoli ridges, cystectomy, apicoectomy and treatment of extraction sockets.

To summarize the safety data from this PMA, the protocol as you heard, was followed without any protocol violations. There were no safety concerns related to the GEM 21S or its components and the safety data collected were consistent with previous studies. The efficacy summary as summarized, showed that the study failed to meet its primary efficacy end point. A high percentage of patients completed the study which the sponsor should be commended for. Results from many secondary points, end points, demonstrated statistically significant results and a retrospective analysis was positive for the low dose group at three months.

In conclusion, one of the things that we're going to ask you to discuss later today is how we should look at reliance on secondary end points form approval, possible approval of this PMA, reliance on retrospective statistical analyses in a possible approval of this PMA and also the clinical benefit for the addition of the recombinant PDGF to beta TCP. And finally, we would like your input on the expansion of the indications from you -- for you from the simple periodontal indication to the other indications that I mentioned previously.

And now, I'd like Ms. Judy Chen to continue with our statistical review.

MS. CHEN: My name is Judy Chen and I'm a statistician from FDA. Right now, I will present to you from my statistical perspective of the clinical studies. I'll just go through these quickly because we have gone through this two times already. This is a multi-center, randomized, three parallel groups trial of 180 patients and Group 1 included beta TCP and plus a low dosage of the subject material PDGF.

And the second group is the beta TCP plus the high dose of PDGF and there's the third which is used as the control and that is only beta TCP and the study is blinded and the study size, the subjects and the monitor were all blended to the treatment of assignment, which, of course, is a very good point.

The primary effective end point defined in the protocol is the change in clinical attachment level between baseline and the six months post-surgery and the comparisons between the low dose versus the no dose. And the study in the protocol also specified a group of secondary end points which is improvement in linear bone growth and percent bond fill at six months and improvement in clinical attachment level at six months but this is for the high dose versus the no dose. The reduction at six months, change in gingivae recession at six months and the bone healing.

Okay, the study hypothesis is to promote greater soft tissue and the bone regeneration as measured by clinical attachment levels and a bunch of secondary end points. Then an osteo-conductive scaffold alone and also then the historical controls. Statistically, the hypothesis -- the study hypothesis is stated as such; that the objective of the study, we like -- we want to show that the alternative hypothesis of the -- is the 2H1s that the clinical attachment level improvement is greater than 1.5 millimeter and also that the clinical attachment level in the low dose group is significantly better than that in the low dose group. I order to do that, we need to reject the known hypothesis which is the improvement in the CAL less than 1.5 millimeter or the improvement in CAL in the low dose group actually is less than or equal to the clinical attachment level in the no dose group.

I know that in order to show the effectiveness we need to reject the known hypothesis. But in doing so, we always will come into two different kind of error rate. One is that false positive error, that is actually where the device is not effective yet it become proven that it is effective. Conversely, the false negative error is that when the device is effective but we concluded that the device is not effective. So both of these error need to carefully controlled. If we don't control this error, both the error rates, we really cannot trust our conclusion.

Okay, that was the statistics will help you but the statistics allow us to based on the clinically significant treatment of fact, the expected standard deviation and at acceptable false positive rate it is conventionally use the five percent. We allow five percent positive error, false positive rate. The statisticians can estimate the simple size which needed to test the intended hypothesis, hypothesis with adequate statistical power so we can be able to control the false negative rate.

So we will find that the difference in statistical significant reject known hypothesis and conclude that the alternative hypothesis of device effectiveness is correct. And also subsequently, when we have the data, the statistician can based on the observed data rejected known hypothesis and conclude that adding the PDGF to beta TCP will improve the six months count when the false positive rate is under control.

Okay, in the protocol, the symbol size is estimated such that for expected treatment difference of one millimeter with standard deviation of two millimeter, and actually the sponsor used the one-sided false positive rate of .05, which I don't completely agree but that is leave it there for now, and a false negative rate of .2 or a power of 80 percent. According to that, 50 patients per treatment group were needed and adjusting for potential missing data, 60 patients were enrolled in the study.

So that the study actually has adequate power to detect a treatment difference of one millimeter or larger. That is the false negative rate is under control in this study. Now, let us look at the result we have after we have the data that mean care improvement at six months are 3.7 millimeter in the low dose group and the 3.5 millimeter in the no dose group, so that the treatment difference is only .2 millimeter with standard deviation, 2.2 millimeter which is statistically not significant even at the one-sided P value, one-sided P value is .2, that's 20 percent.

And also the 95 confidence interval for the treatment difference is minus .35 to .55 millimeter. So based on the data, if we reject the end -- the known hypothesis and the claim there is added benefit, the false positive rate, which is one-sided false positive rate will be 20 percent.

So this is really very much larger than two-sided five percent value. So now we cannot conclude that device is effective so what shall we do? Can we construct new end points? Can we add more patients, be more helpful? Let's first look at the situation of adding more patients. Well, here we can all see that the observed treatment of difference is only .2 millimeter which is much smaller than the precision of instrument which is one millimeter and also the study is designed for one millimeter so fail to reject the known hypothesis due to small treatment difference not due to large standard deviation because if it is due to large standard deviation, statistic procedure under limited circumstance will allow several estimation but here the treatment difference is really way too small.

And also the blinding is broken so if we do that bias will be a problem. And here just supposed if we would do that, we would add more patient to detect a .2 millimeter difference with standard deviation 2.2 millimeter, the study for what we had, the same size we had in the present study the study's power is only 17 percent. To increase power to 80 percent, we need actually instead of 50 patients per treatment group, it needs 750 patients per treatment group.

Of course, ultimately the question is, does the difference of .2 millimeter make any real clinical difference? So we're -- no more patients will not be helpful, so let's look at additional end points. Now, we have set some secondary end points which is specified in the original protocol and also there are new end points which is added later. Well, however, in both situations, you know, the false positive to conclude that a device is effective cannot be controlled like we can for the primary effectiveness end point.

Here let us look at the secondary end point that among these six comparisons, at least six comparisons, I only counted the low dose versus no dose, high significant treatment difference is detected in both linear bone growth and the percent bone fill marginally significant difference is seen in total gain measured as area in the curve and the other variables are not even -- didn't -- it's not -- didn't even reach the five percent level. However, since we made multiple comparisons, we have to know that. Statistician has long noted that if the critical -- usual critical values are used, or the usual P live I show in the previous slide, when there are multiple comparisons the false positive rate is greatly increased over the nominal level. So actually what you have just seen are the false positive rate is too small so that the AUC variable, as you recall, the P value, just reached five percent, can now become considered significant but the other two variable, linear bone growth and percent bone fill will still be significant but the are secondary end points.

Okay, there are more new -- there are other new end points which is the care at three months and gingivae recession at three months and also there are two additional composite end points. We have gone through this before, when the linear bone growth and the other is bone fill. Here are the raw P values of these new end points. Let's focus on the low doses, that's the subject therapy that you can see that both gain at three months and gingivae recession are statistically significant just look at the P value. However, here we made four comparisons, notice the previous slide that for -- if we did multiple comparisons the error rate actually are increased so both these can now become considered significant so that left with the composite end point.

However, they are result based on end points constructed after the blinding. This is my understanding of submission but it certainly is the new end point broken and results on the pre-specified unknown are not reliable. False positive rate are inflated and cannot be statistically adjusted. Okay, there are also results on comparison to baseline values and also here I just compared to standard value of 1.5 millimeter that the improvement for all three treatment groups are all statistically highly significant. However, know that this is comparing the low dose treatment -- this is a comparison between the low does treatment to no treatment, not to the additional benefit by the growth factor and also compared to baseline values. There are other problems such as placebo effect, change due to disease and natural history or regression.

So my conclusion for this submission is that data in the pivotal study demonstrated that adding PDGF to beta tricalcium phosphate does not statistically significantly increase clinically attachment level improvement which is the primary end point. Statistically significant benefit are detected in the secondary end point, percent bone growth, percent bond fill and linear bone growth.

The statistically significant treatment I found in the two composite end points CAL linear bond growth and CAL person bone fill are not reliable since false positive rate are inflated and cannot be statistically adjusted. Thank you for your attention.

CHAIRMAN SUZUKI: Okay, I'd like to now ask the panel if there are any points of clarification for the presentation at this time. Ms. Lawton?

MS. LAWTON: Yeah, I have one question and I don't know whether we should save it for later, but I'm hearing two different things. I thought I heard earlier when the question was asked of the company about when they -- when they came up with these additional end points that it was done prior to data base lock and unblinding of the data, but what I'm hearing from the FDA is that they were constructed after the blinding was broken and I'm trying to understand which one actually happened.

CHAIRMAN SUZUKI: Okay, maybe Ms. Chen can provide that clarification. Dr. Runner?

DR. RUNNER: This is Susan Runner. It is our understanding that the three-month look of CAL for the low dose group was a retrospective analysis. If I am incorrect, please correct me. That's correct, that's incorrect? Would you please clarify?

MR. BEASLEY: Bill Beasley with BioMedic. The three-month CAL was actually part of the original -- it was an original composite end point that was obtained and it was -- that analysis was done before the data base -- or sorry after the data base was locked. All clinical measurements were analyzed after the data base lock.

CHAIRMAN SUZUKI: Ms. Lawton?

MS. LAWTON: And can I just follow up, the two composite end points, were they done before or after data base lock?

MS. BLACKWELL: This is Angela Blackwell. None of the composite endpoints appeared in the submission until the PMA arrived, in other words in the IDE study. So they could have been -- so it's not very clear. You know, they were not added to the IDE study before the study started and that was the case with the three-month CAL. The original protocol called for them to gather the data of CAL at three months but it wasn't considered a primary end point. The six-month was a primary end point. It was only after the data was gathered that they came back and said, "Well, we want to depend more on this one because there was a difference in the data".

DR. SHARMA: Excuse me.

CHAIRMAN SUZUKI: Dr. Sharma?

DR. SHARMA: Inder Sharma. My question is at what point the company came over to ask for the analyze at three months? Was it before the data lock or after the data lock?

MS. BLACKWELL: The data was gathered as part of the original protocol. They changed the way they were analyzing things after the fact. So the data was gathered, it was part of the protocol but they came in later and asked for -- there were some changes in the course of IDE. If you remember you were asking about the change in the dose? That's something that happened during the course, you know, before the data was unlocked but after the original approval.

CHAIRMAN SUZUKI: Thank you. Dr. Amar?

DR. AMAR: Salomon Amar. And this is a question to the FDA. At a certain point, I believe that the FDA had asked the sponsor to perform a meta-analysis with published data from Reynolds and Genobaly. I want to know what was the rationale behind asking this meta-analysis and was that -- if I understand correctly, was that meta-analysis asked after the blindness was broken, am I correct?

DR. RUNNER: I believe it was after the blinding was broken and it was the result of the fact that the primary end point was shown to be not significant.

DR. AMAR: The other question that I had, did the sponsor or the FDA and from a statistical perspective, perform an analysis between Group 1 and Group 2?

MR. CITRON: My name is Mark Citron and I spoke earlier and I'm with BioMedics. There appears to be a certain amount of confusion on the radiographic and the statistical analysis as well as the CAL and if I could take just a second to try to explain the dynamics of the process, I was reviewing some of the documentation and it brought to my mind that submission that we made last summer after the data had been collected and we'd always collected CAL and radiographic information. The data base was still blinded to all of us, to the investigators, to the company, to the patients so we never broke the blind for these analyses and what happened was -- and this kind of helps put it in the proper context, when we made the submission and trying to summarize the situation, I indicated to the agency that the revision to the statistical plan provides a change from the original plan in order to add the radiographic assessment in addition to -- and to include a radiographic assessment was in response to the FDA's original request at the study onset, that the company assess the validity of the radiographic evaluation as an efficacy end point. The collection of the radiographs was pre-specified in the pivotal clinician study protocol. The radiographic analysis conducted appointed fashion and conforms with applicable good clinical practices standards. We're governed by a separate IRB for the protocol, performed by an independent reviewer in a single investigational center that was not involved in the clinical portion of the pivotal study and finally performed after -- after an initial radiographic qualification study demonstrated that the quality of the radiographs supplied by the clinical sites was sufficient to accurately perform the radiographic assessment.

In other words, what happened was the radiographs were collected at the onset of the study. They were necessary to do the safety portion of the assessment as part of the interim analysis and then we were required before we used them for any analytical purpose to insure the integrity of the data, the validity of the data, which we did through a very comprehensive program, separate protocol, filed with the IRB and the results of that study allowed us to go forward with the use of the radiographs, but we didn't do that until we had gone through the process and let's face it, the radiographs were not useful for doing efficacy in three months because they light up and, you know, the data is not valuable.

So we needed to wait for the six-month time to get the valuable radiographs to do the study. We continued to blind the study during the entire course of that assessment and only until after the integrity of the data was assessed were we allowing ourselves to go back and unblind it to run the numbers. Is that -- okay. I'm just trying to clarify.

DR. AMAR: I was just asking whether a comparison was done between low and high dosage.

MR. CITRON: And I'm -- specific to your question, yes, there was an assessment done between the lower concentration of PDGF and the higher concentration of PDGF. In the data tables that were provided in the panel packet kind of delineates all the differences. The statistical analysis, I don't believe found any statistical differences between the low concentration and the high concentration. The statistical differences were between the .3 mg per mil at the low concentration and the TCP alone.

DR. AMAR: At least in regard to the secondary outcome say for example, the one that played the most on Group 1 which is linear bone measurements and percentage bone fill, there was some kind of difference between Group 1 and Group 2. One would say 56 percent, the other one was 33 percent.

MR. CITRON: Oh, you mean for radiographic and --

DR. AMAR: Yes.

MR. CITRON: Yeah, I don't remember the -- yeah, we'll -- we can address that this afternoon. Thank you.

CHAIRMAN SUZUKI: Okay, yes, Dr. O'Brien?

DR. O'BRIEN: Bill O'Brien, I have a question for Ms. Chen. Early in your presentation you made an objection to BioMimetics and their study using the one sided test for significance as opposed to a two-sided test for .05, which is often done since the hypothesis clearly stated a test if something was greater. So I can see why they would use a one-sided test. Did you redo that in terms of a two-sided test for your conclusions or did you stay with the one sided test that they proposed?

MS. CHEN: Yes, in my opinion *** 11:29. And actually in most case the *** already past. Here we need to show that the combination is better than the tricalcium phosphate alone. *** because we are concerned about whether the experimental group actually turns out to be worse than the control group but *** but in this study the significance never seemed to make any difference in the combination because even looking at taking the one sided the other way for determining the end point is *** between the low dose and no dose is the PR is .2 which ***

CHAIRMAN SUZUKI: Dr. Cochran?

DR. COCHRAN: I'd like a clarification from Ms. Chen as well. You said, if I understand this correctly, that the composite secondary variables was not statistically correct based upon the high negative error rate. How about the other secondary variables that were significant, the bone fill and the linear bone growth, are those valid assessments, the secondary end point verbiage?

MS. CHEN: Yeah, since both of them are pre-specified in the protocol, they are statistically -- they are highly significant number so even in the presence of multiple composites I would conclude they are significant but as secondary end point. Why I say it cannot be, you know, adjusted is because I do not know how you put the secondary end point, put it as if they were primary end point because we did do that, taking that end point after the primary.

CHAIRMAN SUZUKI: Any other questions from the panel? Dr. Runner?

DR. RUNNER: I think that the question you asked and Ms. Chen's response is part of what we'd like you to weigh in on in terms of the clinical significance of what the results of this study are.

DR. AMAR: Who selected primary and secondary outcome?

DR. RUNNER: Well, initially the company selected the end points. I believe we requested, if my memory is correct, that we would not like more than one primary end point, is that correct?

MS. BLACKWELL: Yes, Angela Blackwell. Initially there were not as many secondary end points as you see here. Initially the company was going to take the radiographs but they did not have something specified that they were going to evaluate them. They were included in the data they were gathering and then later on they said, "Well, we have this data, so we want to analyze it now".

We had actually suggested since they were gathering it anyway that they perform some type of analysis. So later in response to us, they came back with the two primary end points and then after they had done the study, then they came back with the composite end points that you saw that were the radiographic and the clinical combined.

CHAIRMAN SUZUKI: Dr. Lynch?

DR. LYNCH: Just to comment very briefly on that, I think everybody's statements are correct. The original protocol the CAL was assessed as the primary end point following a pre-IDE meeting with the agency, and the reason for that is that's the primary end point that had been used for both pre-PMA approved products, both Endogain and PepGen. Neither of those products met the statistical end point of CAL for their primary either, very similar to us, but it was felt that since that was the primary end point for the two previous PMA approved products, that that would be the appropriate end point for the study.

And then as Ms. Blackwell indicated, based upon actually the FDA's recommendation, we did then add secondary end points, the radiographic analysis of bone fill.

DR. AMAR: I guess my question now is and I'm asking that to the statistician, can we retrospectively consider linear bone gain as a primary outcome?

MS. CHEN: Judy Chen. They are, they are pre-specified secondary end point and they are -- the secondary are pre-specified but they are secondary specified as secondary end point. The primary -- it's not a primary end point, that's what I want that to be clear. But they are pre-specified and they are statistically highly significant.

DR. AMAR: The reason I'm raising this issue is that in periodontal disease and periodontal treatment, it depends on how we look at the success of treatment. If it's viewed as closing of the pocket, it's one of the end point. And if it is viewed as bone gain and regeneration, that's taken from another perspective and then the choice is unequivocal here but if it's taken as closing of the pocket, then that becomes another regenerative issue. That's the reason I'm considering whether statistically we can now retrospectively take linear bone gain as a primary outcome and that's -- forgive my ignorance.

DR. RUNNER: I think that, you know, just the comment, that's why we --

CHAIRMAN SUZUKI: Dr. Runner?

DR. RUNNER: Yeah, this is Susan Runner. I think that's what we want you to weigh in on, given your clinical experience what the clinical significance of the results are.

DR. SHARMA: Inder Sharma, may I speak to this? No, you cannot. Yeah, certainly for new study you can plan other end points which are more promising to you as -- or even considering, even I would say based on what you learned from the previous computation that they could not achieve the statistical significance that you're hypothesis could have been non-inferiority or non-superiority hypothesis and there are a lot of other considerations to take into account. But you cannot have the secondary at this point retrospectively extended.

DR. LAVIN: Yeah, Philip Lavin. I think that I would agree with Dr. Chen's (sic) perspective that the secondary end points are valid statistically. Where it gets murky and where it gets complicated is the interpretation of the composite end points, but I would stand on and agree with Dr. Chen that the two secondary end points stand alone in terms of their statistical significance.

CHAIRMAN SUZUKI: Dr. Zero?

DR. ZERO: Domenick Zero. Since the radiographic analysis is becoming more important in our discussion, I want to go back to the protocol a little bit and my understanding was that the initial radiographs were taken mainly for the purposes of safety to make sure that the subjects that were enrolled met the inclusion criteria. What I'm -- we're now taking that information and now moving that baseline forward and then using the six-month radiographs as a way of interpreting the study.

The fact that in the protocol, there isn't a very good description of how the radiographs were taken, and the fact that they weren't going to be used as a outcome measure for the study, was the standardization, the training, the condition that the x-rays were captured under, do they meet, you know, what we call scientific criteria for using radiographs as an end point?

CHAIRMAN SUZUKI: Dr. Lynch?

DR. LYNCH: Sam Lynch with the company. Again, as part of the pre-IDE meeting we had with the agency, we had this discussion about the radiographic -- the use of the radiographs and you are correct, in the original protocol that was approved, and what we had agreed to do was to collect the radiographs, certainly analyze them for safety purposes as well as eligibility criteria but had agreed as a course that that -- of those discussions that we would look at their validity actually to your very point, as using those radiographs as a basis for measuring bone formation, linear bone growth and percent bone fill.

We had a discussion at the pre-IDE meeting as to whether or not -- in fact the agency, again was encouraging us to use those and we actually had gotten some frankly mixed messages as to the validity of using radiographs and what we came back and suggested was that we would collect the radiographs. We then in a separate pre-specified protocol that was fully IRV approved and so forth through Dr. Reddy's site at Alabama, did a qualification study on the first 25 sets of radiographs that were collected, full sets of radiographs, baseline three and six months, to answer the very question, are these valid, you know, radiographs. And are they useful and could Dr. Reddy do the analysis.

It was at the conclusion of that qualification study and based on Dr. Reddy's report which was submitted to the agency, that then we said, "Yes, we do believe that these radiographs are of sufficient quality and that the analysis is sufficiently robust that we could rely on the data that was again, submitted as a formal IDE supplement to the agency clearly specifying exactly what those radiographic end points would be for linear bone growth and percent of bone fill.

DR. RUNNER: This is Susan Runner again. I think from our experience with previous periodontal studies, the radiographs tend to be important when it comes down to the final look at the data and I think that's one of the reasons why we suggested that to the company.

DR. ZERO: Domenick Zero, just to follow up, there are limitations to what you can do with radiographs after capture based on image analysis or -- I mean, you can stretch it but you can only go so far and you do approach a quality issue of whether it's acceptable and scientific. So I recognize you can do some post capture modification of a radiograph but there are limitations.

DR. LYNCH: Absolutely, Sam Lynch again. And as part of the investigator meeting, prior to initiation of the study is we -- Dick mentioned during the course of our presentation. We did clearly describe with and discuss with the investigators the importance of the radiographs, the standardization technique using the Rand *** instruments and the fact that -- and I think this comes out in the data, that they should analyze those radiographs, you know, clinically, visually, prior to releasing the patients to make sure that in their clinical judgment, the radiographs were of sufficient quality to be you know, analyzed. And so we clearly, you know, made a big point of this with our investigators and I think that the fact that 174 cases out of the 178 total cases in the study were able to be analyzed radiographically, speaks to the fact that again, we and more importantly our investigators were very careful in the quality of those radiographs and if there was a radiograph that had a poor emulation or so forth, they retook it before the patient left that visit.

CHAIRMAN SUZUKI: Dr. Cochran?

DR. COCHRAN: Yeah, David Cochran. I have another statistical question for all our support statisticians. This product is used to stimulate healing responses and I understand Dr. Lynch's use of six month as their clinical attachment level end point due to the prior submissions that had occurred, which makes a lot of sense, but this product actually as opposed to those previous products, does speed up the healing.

So looking at a three-month clinical attachment level certainly makes sense from the mechanism of action. But my question is, by going back and re-analyzing it, does that change anything that we should interpret from a statistical point of view looking at the significance of CAL at three months.

CHAIRMAN SUZUKI: Ms. Chen?

MS. CHEN: Judy Chen. Yeah, sure, of course, there has to be modifications. I think in my slide, I don't know which one, that at the three months, the three months CALs comparing to low dose and the no dose, it -- the -- let me try to find it. It's like -- it's quite marginal, that's my point, yeah, three months is .04, but here we just look at, we make four new end points, we didn't even consider the other ones, even just consider four, with multiple comparisons, the P of .04 will not be statistically significant if we consider multiple comparison, which we should.

Now we have -- we may and if I can have a clarification actually that for the previous question by Dr. Amar, that about whether secondary or primary end point, how do we consider the P value, I have to clarify that I didn't say a secondary end point can be taken as primary end point. What I actually said that statistically that there's no method that we can adjust a secondary end point to a level of primary end point that I know. So I honestly that statistically we cannot, I don't know and I think that the question has to be clinically determined.

CHAIRMAN SUZUKI: Thank you.

DR. COCHRAN: Let me follow up on my question, please, that I asked. David Cochran. The data was blinded and had -- was collected, the three-month data, prior to analysis but going back and looking at the three-month time point, although you said that when you make the one comparison, it's marginally significant, but if you did a multiple end point analysis, it wouldn't be significant. But it is significant at three months in the two comparison; is that right?

MS. CHEN: Because we have many end points that -- I have a slide that shows -- backwards. Yeah, multiple comparison and the significance level. Because we make more than one comparisons, the false positive rate is inflated so that at nominal -- what we see all these P values are nominal value and they are too optimistic that the actual false positive rate is much higher. The most, you know, single adjustment is that if you make four comparisons, in order to achieve an overall five percent rate of false positive, you need to divide that .05 by four which means, .125, then the P value we have really is much, much -- didn't reach that. It's too high.

DR. LAVIN: Philip Lavin, sponsor's statistician weighing in. In response to Dr. Cochran's you know, point about looking at the three-month end point and the six-month end point separately, that was part of my rationale for coming up with the area under the curve and we did send that into the FDA and this was done, you know, before we broke the blind, so the area under the curve analysis was to integrate together the chain from three months with the change from six months, so that we could actually avoid this multiple comparisons question with the AUC.

CHAIRMAN SUZUKI: Okay, any other panel questions? If not, we will adjourn for lunch and we will convene at 1:00 o'clock.

(Whereupon at 11:48 a.m. a luncheon recess was taken until 1:02 p.m.)

CHAIRMAN SUZUKI: On the record. I would now like to call this meeting back to order. We will now continue on the agenda with the panel discussion. Drs. Amar and Sharma have presentations and will lead the discussion. Dr. Amar will present first.

DR. AMAR: Good afternoon. My name is Salomon Amar from Boston. What I'm going to be trying to do with your permission is to summarize. Although you have seen the data being presented this morning by the sponsor, by the FDA, I will summarize the important data and take them into a somewhat a clinical perspective as to what is expected from a product like that if it ever comes into the market.

One of the mandates of these Federal agencies is to look for safety and I think that I can address that very specifically that in terms of safety my personal appreciation or evaluation of this component has been such that the product whether taken alone or in combination does not jeopardize or provide important concerns as far as me as a reviewer for the public in being safe to be presented to the public. So the question that comes to the panel is the efficacy or effectiveness in regards to predicates that are available in the market.

One of the questions that comes to the panel is basically considering the statistical result that is there a clinically significant benefit from the addition of the recombinant PDGF ββ-tricalcium phosphate. I got asked this question this morning to Dr. Genco as well as the rest of the Panel and the way I would rest or at least momentarily I would say is that there is some kind of a clinical significance associated with adding the PDGF whether in terms of early wound healing that has been seen or definitely if one is interested into taking linear bone measures and bone fill as an outcome in terms of bone regeneration.

We all know the periodontists that are in the room that for maintenance of a tooth in terms of longevity the presence of an increased bone is a factor. I think that this trial, and I would like to congratulate the sponsor because this is, within itself, was well conducted and we cannot allow yourself into comparing, as Dr. Genco said this morning, two other trials, but within itself, the trial was well conducted.

If the outcome from a periodontal perspective or periodontal surgeon is regeneration per se, then the choice is unequivocal. There is clinical significance as compared to either other product or as compared to itself, the control. If, now, closing of the pocket is an outcome, we can consider that there is a significance after three months. Although that significance doesn't hold on because β-tricalcium phosphate closes also at six months, but there is no bone support underneath.

So it all depends. It does catch up to the control, but the control doesn't have bone filling as was to the β-tricalcium phosphate with PDGF of having bone fill and that's a little bit of the perspective that I wanted to bring to this panel.

Furthermore, I didn't have a chance this morning to address the sponsor, but I wanted to know, and that you could address later on, why gingival index, bleeding index and plaque scores there are across the board that take any multiple clinical trials as being standardized, the index not being used in these studies as opposed to a Lowborn (PH) 1986 scale that in all likelihood was just by a few trials. So the comparison with other measurement, in addition to the fact that the way I looked at this scale, is a little bit more subjective as opposed to a bleeding and probing being there or not there or bleeding the Eastman index of putting the stroking into the pocket, determining to us presence of inflammation and not presence of inflammation. That would have been a better marker of wound healing in that area. In addition to the fact that pre- and post-computer digitized image were not taken and that could have an issue and it still is a remaining issue, although this was somewhat addressed but not to complete satisfaction.

In terms of the result, I summarize the result. There is clinical significance in terms of clinical attachment levels at three months, but there is no differences with the control and that could well be the case as opposed to if an individual is interested into maintaining and looking at regenerative procedures, bone is a component and a major factor of over that and clinical significance was achieved over there.

The other concern that I had, and I tried to vocalize that this morning, is that I haven't seen a comparison between Group I and Group II which is low dose v. high dose. I'm going to tell you why. The way I look at it at this point is that to my understanding Group II did not perform or performed worse in terms of bone healing as well as linear measurements, 56 percent as compared to 33 percent. The linear measurements were 2.4 millimeter for the low dose as opposed to 1.5. That's a little bit of a, I would say, "detrimental effect" in terms of the bone.

My question to the sponsor is that shouldn't we be concerned by a practitioner using several dosage for several grafting area and therefore increasing the local concentration of PDGF reaching probably 0.75 milligram or 1 milligram and reaching detrimental conclusion or detrimental bone support on that. So given that Group II had poorer results than Group I, I would be concerned about grafting multiple sites at this point and that would be one of the recommendations.

As I said earlier, adverse events are not an issue. Here it was concluded recently from the FDA that the mode and the ERS data that are an adverse event, the database reviewed, do not have too much relevance to this product probably because any adverse event that was reported separately from each of the components was due to patient underlying conditions. So a risk benefit analysis was performed according to the IS. So I indicated that the benefit of using the device outweighed the risk associated with itself. Body compatibility appears to be safe and intended to use. Now like I said, I will reiterate some of the concerns that I had, concerns all raised at this point for lack of information to the practitioner regarding safety in large and multiple defects where PDGF local concentrations could raise to a substantial level that could lead to a bolus concentration and therefore, bring to poor result.

The clinical study data component of the device really did not include two areas for which the device did not perform well as opposed to 3-wall defect and 1-wall defect. So my recommendation, if I may say so? Although there are several criticisms in regard to the IDE study, one of them was mentioned this morning where the sample size was too small as well as the analysis of 1-sided t test could have been, 2-sided t test could have been, better for the analysis.

But even taking into consideration the criticism, I tried to bring to this panel the perspective that there is clinical significance as compared to predicate in the market. Therefore, I would recommend approval of the component with several considerations. Thank you very much.

CHAIRMAN SUZUKI: Thank you, Dr. Amar. The next presenter will be Dr. Sharma.

DR. SHARMA: This is Inder Sharma. Looking at the data I reviewed presented by the company as well as data was provided to me by the FDA, it appears to me the device is safe. As to effectiveness, the data is a little bit mind-boggling to understand why it's working for low dose and not working for high dose.

There could be several reasons there, maybe looking at a more understanding of the population and are there any factors which are confounding the issue. So there would be an important thing to be considering for maybe a future study to be conducted to redesign the study with appropriate end points and appropriate stratification is it is necessary.

Based on efficacy data, I would suggest that these are some of the things which should be clearly looked into that whether we want to go with also superiority or non-inferiority. But it's possible there may not as much room. It's always harder when both are active devices or drugs. It's harder to prove unless you have a very large sample size and it depends on also a choice of endpoint. If there is an endpoint which can show a significant difference, sizable, clinically-meaningful difference, then you can even have a smaller sample size and still be able to prove your point whether you take superiority or non-inferiority. So these are some of the considerations to talk through, looking at and revisiting the design of the study to get a superiority kind of claim.

In terms of just by itself, I see effectiveness is there in all the three groups. All the three groups are proving that to a certain threshold they are all working. So I see the effectiveness there, but in terms of superiority, I see that that's where the challenge is. To understand the design, to understand the population and maybe reading through that future study should be conducted to get that kind of claim that you want, superiority. Thank you.

CHAIRMAN SUZUKI: Thank you, Dr. Sharma. To guide our discussion, the FDA has a few questions for our consideration. Ms. Blackwell.

MS. BLACKWELL: Yes, I'm going to read the questions. They are going to appear on the screen and I think, Mark, you wanted to say something.

MR. CITRON: We'll wait until after.

MS. BLACKWELL: Okay, and then I think the Company had some things that they wanted to say after we've gone through the questions.

(1)"Clinical Benefit of the PDGF. Preclinical and feasibility study data appears to indicate that the PDGF is safe for use in humans. The primary efficacy endpoint results do not demonstrate a clinical benefit for the addition of the PDGF to B-TCP at six months. Please discuss the clinical ramifications of the clinical attachment level results at three months versus six months."

(2) "Endpoints and retrospective analyses. Please discuss the validity and clinical significance of relying exclusively --

CHAIRMAN SUZUKI: Excuse me, Ms. Blackwell.

MS. BLACKWELL: Yes.

CHAIRMAN SUZUKI: Could we go back to the first question?

MS. BLACKWELL: Well, I'm reading through them right now. You'll be able to see them again.

CHAIRMAN SUZUKI: We were going to go through each one at a time.

MS. BLACKWELL: Okay. Well, hold on.

CHAIRMAN SUZUKI: Panel members, is there any discussion on question number one? Dr. Cochran.

DR. COCHRAN: This is David Cochran. I think it's important for us to keep in mind several things that are going on here. One is that we're looking at lesions that are very complicated biologically that a lot of different factors are coming and going during the healing process and it's not a simply system we're looking at where we're looking at one type of tissue that has to regenerate.

We're looking at multiple types of tissues. So when you're looking at product that's going to stimulate regeneration and you're delivering it at one point in time, I think that we have to consider that there are going to be multiple issues that are going on which are going to cloud the potential significance of the data when you pick a point in time.

Secondly, the sponsor has picked an endpoint that was based on prior PMAs that had come before the panel so they were led to think that six months was a good time point. The other thing I think that's very important is that when you have a product that is going to stimulate healing, the chance of seeing early healing is certainly where you might see the difference based on this type of product. So I think given those types of issues they are very important for us to consider when we look to consider whether this product should be approved or not. Thank you.

CHAIRMAN SUZUKI: Thank you. Any other issues or questions from the panel members? Dr. Zero.

DR. ZERO: Domenick Zero. Since not being an expert in this area, I would like from the panel or from the sponsors of the application to have a better understanding of the connection between the relatively short-term outcomes and long-term health of the patient in terms of tooth retention, mobility, functionality of the tooth, susceptibility to subsequent disease later on, those issues because again to make a distinction between these three-month to six-month outcomes without that understanding I have difficulty in making a good decision.

CHAIRMAN SUZUKI: Would anyone like to comment on that? Dr. Lynch.

DR. LYNCH: Sam Lynch, and again I just want to serve as a moderator here and I think our clinicians should speak to that. So Dr. Genco and perhaps Dr. Nevins could talk to the clinician benefit that was observed in this trial, if that's really the point to the question.

DR. ZERO: No, the point is that we have a very short-term clinical benefit. That three and six months is a very short-term clinical benefit. How predictive and how valid is this outcome measure for predicting future health of the patient. In other words, if you have this gain but it's lost in a year, that's a very little benefit to the patient considering the pain, the cost and the various factors the patient has to endure.

CHAIRMAN SUZUKI: Dr. Runner.

DR. RUNNER: Could I make one comment just in terms of how we looked at clearing the IDE? In some of our past periodontal studies, we have looked at even a year data and through working with clinicians and representatives of the periodontal clinic community, we've been led to believe that in periodontal research after a six-month period, in particular, you start to get maybe recurrent disease process taking over again and clouding the results. Therefore, that's one of the reasons why we pull back to the six-month timeframe as one endpoint area that sometimes, you can correct me if I'm incorrect, after six months you sometimes get a clouding of the picture in terms of the disease process as a whole.

DR. GENCO: I think Susan had stated that.

CHAIRMAN SUZUKI: Dr. Genco.

DR. GENCO: Excuse me. Bob Genco. Many conferences have been held with the FDA workshops on clinical trial design and this is a very important issue that was originally or let's say maybe five years, nine months. Six-month result holding up for another three, that was the standard. I think what we all saw was that you could get second episodes of disease. I think that's your question.

What about recurrence? The studies from the point of view of looking at risk for periodontal disease, we've done a lot of that work. Previous periodontal disease puts you greater risk for future disease. I mention this morning that one there's a cutoff point about five or six millimeters of pocket depth.

If you have less than that, then you're at much less risk of having future disease. So in answer to the question, Solomon's question this morning, so we went from seven millimeter pockets essentially to three or four. So that's good, predictive of long- term health.

The second thing, and Ernie Houseman did a lot of this work with us, is looked at the level of bone loss as a predictor of future bone loss. So if we can get more bone to grow, then it bodes well for the future. So I think in terms of those two measures what this product does is it gives us both. It gives us closure of the pocket and that may be due to the surgical procedure and the care of a plaque control and good suturing, good technique, as much as the product.

We understand that, but certainly the bone formation, I think, bodes well for the future. If we can get more bone around the teeth, this predicts health in the future, the best we can.

DR. ZERO: Okay, but using the basis that if you have the bone you're better off, you don't have the answer to the question that if you form the bone are you better off.

DR. GENCO: Yes, I think the assumption is that if you'd lost it and then you'd regained it, you're better off than if you'd lost it and didn't regain it.

DR. ZERO: But it's an assumption at this stage.

DR. GENCO: Yes.

DR. ZERO: That's what I want to understand.

DR. GENCO: I think it's pretty well shown from cross-sectional, longitude and epidemiological studies. Thirdly, Sig Sacraski (PH) has just shown that six millimeter pockets harbor pathogenic flora and the shallower ones don't. So there's this cut-off point of five or six that seems to be real in terms of predicting future loss.

DR. ZERO: But that speaks to just the CAL outcome.

DR. GENCO: Right, but again this is a complex disease and you have to look at all of these issues, the flora, the pocket depth which is reflective of CAL and the bone.

DR. ZERO: Okay. Thank you.

CHAIRMAN SUZUKI: Dr. Nevins representing the sponsor.

DR. NEVINS: Myron Nevins. I would like to address your question, too, with a background of 39 years of experience in clinical periodontics added on to my advocation of education. There is a distinct difference in what we might anticipate with how we approach a patient clinically. But I think it's incumbent upon us or obligatory upon us as clinicians to determine what is in the best interest of the patient. And I think that's the primary interest to you as a panel.

There's a difference between CAL and periodontal regeneration. Given the same grouping of teeth for the same amount of bone loss, if we make the mouth clean and we open the tissue and we clean the roots and more and we put the tissue back without any significant change in the supporting structure of the tooth, we'll see an improvement in the CAL just by eliminating the inflammation.

It would be nice to reflect on that as a success and to move on with our career. However, the lack of compliance on the part of our patient population is too evident. There is no report in our formal recordings, our literature so to speak, of patient compliance after five years of being more than 33 percent.

Therefore, it becomes incumbent upon us to try to reach the best possible level we can for a patient. Therefore pocket elimination, CAL are all interesting, but the endpoint goal from productivity of every periodontist should be to try to restore the periodontum that's been lost. Obviously, using the example of a 10 millimeter root, whoever did that this morning, I don't remember who used that number, if we had a 10 millimeter root and we had seven millimeters of bone, we'd be much better than if we had four millimeters of bone.

In this instance where we're talking about how we're going to measure regeneration, the radiograph really has to take precedence because although it's not a substitute for histology as I presented this morning, you can not make the mistake of looking at a radiograph and determining that you've achieved regeneration. That's already been demonstrated to be a fallacy with histology, but it's a much better indicator than CAL as to whether you're going in the right direction.

Now we've presented a series of matters of information. Granted, it's very difficult to use an active control and fulfill the dream of a comparison between a new product with and without that control. The bottomline is that all of the information that we've provided has been going in the right direction. I don't think there's any question on anybody's part that our bone measurements as secondary evidence shows that the patient has benefitted.

We can't solve the CAL at six months because as David Cochran said, there's so many things going on. We have new cementum being formed, new bones being formed, remodeling of the bone, the periodontal ligament, the attachment apparatus that we may not really have the ability to make a judgment on periodontal regeneration. This may be why all the products stumble when they try to get to this because some products, we mentioned Emdogain, might be 14 or 16 months until we really see what the endpoint goal of that is.

Your question. The most germane question that may have been asked today other than, of course, the primary what's the patient -- The primary question should be what does the patient benefit, not what is a CAL. The question that you're asking is what happens when you regenerate and you get to six months or you get to six years what's the benefit.

Well, from clinical observation when things go well, we seem to go smoothly indefinitely. This isn't one person's observation, but it's very difficult to do in RCT long term on a group of patients because their compliance is terrible. You can't tell who's going to die, who's going to move away, who's going to become unhappy with the services of your practice or your university or however it goes.

We ran into the same thing with the NIH with trying to plan long-term RCT answers with dental implants. There was implant failures and trying to find matched defects to do bone growth for dental implants. There are some things that are hard to do. There are some questions that are difficult to answer with the investigative methodology that we have even in a case like this where I think we've done a pretty good job with an RCT. And 99 percent patient retention, it's almost unheard of.

But you still cannot bridge the gap of the unknown. When you hang everything on what the CAL's going to be in six months, you tend to overlook what is the benefit to the patient. If you can show regeneration of Class II furcation with human histology as we were able to do today, that's an eyeopener. That's something we have not been able to do in 100 years of being periodontists.

I mean the American Academy of Periodontology is about 85 years old now, 90 years old, and we're just finally getting to the point where we have a product. We can do this. To get hung up on statistics and overlook the benefits, even they are secondary because you only allowed one primary, may overlook the question you're asking. What is the benefit to the patient and the benefit to the patient is significant. That's our job as clinicians, our job as educators, your job as a panel.

CHAIRMAN SUZUKI: Dr. Nevins, this is John Suzuki. As a follow-up question, I guess, Dr. Nevins, would you consider your cohort population of 180 patients at six months to be the same at risk for future periodontal disease as other periodontal patients?

DR. NEVINS: I think that these people probably now have been committed to a careful period of observation and unfairly, they've probably been tutored to a point where they may have a better chance of taking care of their mouth because maybe they'll pay more attention. When people participate in a study like this, sometimes they pay more attention than those individuals that go through the daily rigors of the practice because you're taking special records and you're taking photographs and you're doing things that they catch on to the thought that well, maybe they're going to do better. But I'm terribly optimistic with the results that we've seen compared to what I've observed with so many other studies that I've participated in.

CHAIRMAN SUZUKI: Dr. Zero.

DR. ZERO: Domenick Zero. That just triggers the another question I had. Is this a special population? In other words, was this population preselected because they demonstrated they were able to maintain a 15 percent plaque index?

DR. NEVINS: The protocol demonstrates that these people did not go through a long preparatory period, an observation period, before they participated in the study. These people, other than some exclusionary data that we looked at this morning, basically showed that they had a four millimeter, curving depth intraboning.

DR. ZERO: Right, but they weren't selected because of the potential for being good compliers.

DR. NEVINS: You know what? I wouldn't be able to do that if I tried.

DR. ZERO: Okay.

CHAIRMAN SUZUKI: Dr. Zuniga.

DR. ZUNIGA: Jon Zuniga. The question has been broached earlier and I guess this is another time to broach it once again. That is can someone in sponsor or on the panel explain the lack of a dose effect on either the CAL, the two, three and six month and translate that to clinical significance?

CHAIRMAN SUZUKI: Dr. Lynch.

DR. LYNCH: Sam Lynch. We actually -- I don't know how possible it is to put our slides up because we had a slide prepared specifically to address that question. Is that possible or not?

DR. COCHRAN: Jon, this is David Cochran. Let me make a comment on that. It's not unusual for an optimal dose to be lower and give you a better effect. This happens with growth factors a lot. TGF-B, for example, very sensitive to dose response meaning if you go too much, you're going to get less response.

Even in the Emdogain data, we've done a lot of work on the enamel matrix proteins. You can get too much of that material as well and the results start getting more and more negative. So I don't think it's a real surprise in any sort of way that there would be an optimal dose that's lower than expected.

The fact that you're using a carrier that's going to keep the material around may also contribute to that. There are so many factors involved that are going to influence the amount of the protein that actually gets released there that because you don't continue to get a higher effect with higher doses is really not surprising at all.

CHAIRMAN SUZUKI: Dr. Amar has a comment also.

DR. AMAR: Yeah, Dr. Amar. In fact, it's a curve, it's a bell, that you have before and after and you have an optimum concentration in many of the growth factors. I have a question for Dr. Nevins. Dr. Nevins, when you did this 11 patients and you did some histology, were you able to calculate the amount of cementum regenerated above the notch?

DR. NEVINS: The histologic analysis was done by Robert Shenk in Bern, Switzerland and he did a histomorphologic analysis. I don't have it at my fingertips, but we did have the radiographs and they were very similar. When the radiographs were measured and we had the histologic measurements, it appeared that they were a very accurate measurement of bone development.

DR. AMAR: How about cementum?

DR. NEVINS: And cementum. They were both the same level.

DR. AMAR: Do you have any idea how --

DR. NEVINS: The cementum was actually coronal to the bone.

DR. AMAR: Do you have any idea how much on average -- I guess I'm trying to compare the data and this is an eyeopener as you just said. There's no question about that in terms of doing some kind of trials and having the possibility of doing histological approach is very remarkable. I must say that. All I want to determine is the amount of cementum and try to compare it with what we know on human clinical trials.

DR. NEVINS: I realize. I just want to turn my head to Sam and see if he had -- Do you remember the analysis of the cementum from Shenk?

DR. LYNCH: Just as you alluded to, it was to the place slightly coronal.

CHAIRMAN SUZUKI: Can you come to the podium please, Dr. Lynch?

DR. LYNCH: It is always coronal to where the bone level is and there's a periodontal ligament and you can always see the epithelial edge and stuff and, in short, the cementum. I didn't come here prepared with the numbers in that study unfortunately.

DR. AMAR: Are we talking about two, three, millimeters?

DR. LYNCH: I think you're talking about four millimeters.

DR. AMAR: Four millimeters.

DR. NEVINS: If you look at that canine data, it was 70 percent bone fill but as we pointed out this morning, it was about 36 percent regeneration in the canine, but the bone fill was pretty dramatic.

DR. AMAR: Above the cementum.

DR. NEVINS: Yes, but the cementum always runs above the PDL and the bone. Ron's exactly right.

DR. AMAR: No, I'm trying to bring this into the perspective, the Bower's study, if I remember correctly.

DR. NEVINS: Bower studied the average cementum if I'm not wrong and it was 1.8.

DR. AMAR: 1.2.

DR. NEVINS: Okay.

DR. AMAR: I'm just trying to --

DR. NEVINS: It was 1. something. In this instance, I think it was more 4. something.

CHAIRMAN SUZUKI: Okay. Dr. Lynch was going to comment on the inverse relationship of the dose or was that answered?

DR. LYNCH: Unless there were further questions, I think that was answered.

CHAIRMAN SUZUKI: Are there further questions on the dose?

DR. NEVINS: Dr. Amar, did I answer your question?

DR. AMAR: Absolutely.

CHAIRMAN SUZUKI: Dr. Zuniga.

DR. ZUNIGA: Dr. Zuniga. In relation to a practical as a surgeon who's approaching the site and may be size dependent on the site, you have a prepackaged unit. You might be tempted to put two of those units in or three of those units in or two number 18 is too close to 719, in the relationship. There are questions that you might be two no. 18, 17 or 16, whatever, and you have a large accumulation in an area. Is that a negative? Are you now going over a certain threshold of effectiveness? That has not been defined to that.

DR. COCHRAN: This is David Cochran. That was Dr. Amar's concern.

DR. ZUNIGA: Right.

CHAIRMAN SUZUKI: Dr. Lynch.

DR. LYNCH: I think we have several people that would like to respond. Dr. Giannoble, would you care to respond?

DR. GIANNOBLE: So the issue regarding the dose applied, it was the dosing that was utilized in this study included a concentration. So the grafting material received the application of the growth factor, the PDGF, at either the 0.3 milligrams per mil dosage or 1 milligram per mil.

The way it's put together at chairside is that the granules are saturated with the material and so then you're only able to adhere, absorb, as much PDGF that's within that specific solution. Actually the practitioners utilized in the surgical design of this study, they applied that entire amount of the PDGF to the granules. It is concentration specific.

Some of the defects were much larger than others as we looked three dimensionally in terms of the volumetric measurements. Some of the lesions were wrapped around lesions that encompassed the mesial and distal and part of a furcation and so the entire amount was utilized in those studies for comparison. Some of the other dosing and release studies were done by Dr. Hollinger who I think can speak to that point a little bit more if he'd like.

CHAIRMAN SUZUKI: While Dr. Hollinger is coming to the podium, Dr. Amar, do you have a comment?

DR. AMAR: Did you ever or any of the clinicians come into situations were you had to open a second or third package to be able to feel several sites and how did you manage all this?

DR. GIANNOBLE: So the study protocol was to treat the target lesion. On average, I think we have the numbers on that, but it averaged about half of the granules which were actually applied with the platelet derived growth factor. It was never more than one package applied to a lesion.

DR. AMAR: Okay. So the question is clinical situation and in clinical setting, and you know that better than me probably, you have a patient coming in your office, upper maxillary region that we open five or six teeth together and they are different lesions. We run out of the material. What do we do? Are we or aren't we allowed to open a second packet and can we graft each one separately when we run out, use another package?

DR. GIANNOBLE: I think that what we would do is with that particular formulation of the platelet derived growth factor you would still use it in the same concentration dependent manner applying it to the granules. Then you would treat multiple defects because basically in essence we would need to saturate the particles and then apply them to as many defects that this material could be fit into.

So essentially given if there were multiple large defects, then one would open up another package and apply it in the same controlled manner as what was done in the clinical trial, i.e. there would be the cup. The PDGF would be placed into that, absorbed onto that B-tricalcium phosphate and the delivered into the defect. That would be a way to control for the amount of PDGF on the device itself. Does that address your question?

DR. AMAR: I'm still a little bit, I have to be honest, unsettled.

DR. LYNCH: I think again we'll have a couple more comments on this.

DR. NEVINS: What's important is the concentration. So if you had a cup of the TCP and you had the vial of the liquid, the recombinant PDGF, and you put that in there, now you have a concentration. Let's say what you're describing three lesions in a quadrant. You apply that concentration. That's 0.3. Now you run out of material. You open another package. You have your TCP. You put your liquid in. It's still the same concentration. What you're seeing in the chart is different than what the concentration is.

DR. AMAR: I have no problem with that as long as we label it that it's concentration dependent and we should not mix different preparation together in an attempt to fill larger defects.

DR. NEVINS: And we agree with you.

DR. RUNNER: This is Dr. Runner. And that's a labeling issue.

DR. NEVINS: That's a labeling question. We don't have any disagreement with that.

CHAIRMAN SUZUKI: Okay. Any other comment on question no. 1? Okay, Ms. Blackwell.

DR. RUNNER: This is Susan Runner. Could we just go ahead and read the question without that way we won't have to keep going back and forth?

MS. BLACKWELL: Yes, I was wanting to read them all.

DR. RUNNER: Well, let's just do one at a time. Let's go to the second one. Just read it.

MS. BLACKWELL: "Endpoints in retrospect of analyses. Please discuss the validity and clinical significance of reliability exclusively on the secondary endpoints and retrospect of analyses identified by the sponsor for approval of this PMA."

CHAIRMAN SUZUKI: Okay. I would like to open this question for panel discussion. Comments? Dr. Cochran.

DR. COCHRAN: David Cochran. I think we you know have to consider the light of this trial. The sponsor actually picked the toughest control they could have picked because they picked an active control, if you will. There's a lot of literature available that suggests that pretty most of the bone graft materials that we put in there you're going to get a pretty good response. Some of the literature indicates for Lars Hale, 1997, with the enamel matrix protein 66 percent defect file.

If you look in the paper, that was from Greece that was used by the sponsor, in their discussion section, they said that with DFDBA they get 58 to 78 percent fill. So the range for any kind of bone graft material alone is going to be fairly significant.

When you use an active control that's going to give you a very positive response, it means the chance to distinguish a difference between adding some type of protein, the chance of showing that difference is smaller. So given the fact that we have secondary outcome variables that are very much clinically relevant with a bone graft with the x-ray data, I think it's not unreasonable to accept the secondary outcome variables as a proof of concept.

CHAIRMAN SUZUKI: Okay. Any other comments from the panel or the sponsor? Question No. 3, Ms. Blackwell.

MS. BLACKWELL: "Invented use. The sponsor studied GEM-21S and interosseous periodontal defects. In the PMA, the sponsor is requesting approval for the following intended uses: periodontal disease, cystectomy, apioectomy, deficient alveolar ridges, and tooth extraction sites. Are these claims support by the data and the information submitted?"

CHAIRMAN SUZUKI: Okay. I would like to open this question for panel discussion. Ms. Howe.

MS. HOWE: Elizabeth Howe, Consumer Representative. One of my key interest area has to do with benefit to patients who might have problems with healing, albeit the elderly. There was an indication that somebody in the population had diabetes. One piece of information I couldn't get my hands on was an Appendix 16-2 that actually wrote down individual subject data.

I'm wondering if there is any anecdotal or data information that is available to address this particular population, maybe the elderly in the group. I just saw an summary of age ranges but not necessarily breaking out the people who were 60, 65 and over and if, in fact, this could be a great benefit to this particular population.

CHAIRMAN SUZUKI: Would the sponsor like to respond?

DR. LYNCH: Let me comment from a preclinical data perspective if I could and then I may ask Dr. Lavin to enlighten us as to any stratification that has been done to-date on the different age ranges. But I would like to draw your attention to just the one piece of data, the one site, that we used to illustrate the effects of PDGF in particular in an osteoporotic model, so it would somewhat address the question.

As you may recall from that data, that was in a well-acceptable osteoporosis model where they had removed the ovaries from adult female animals. They become estrogen-deficient and they get osteoporosis very similar to estrogen-deficient, post-menopausal women so it's a very clinical relevant model.

In that study, we did demonstrate a strong statistical benefit by addition of PDGF basically throughout the skeleton. The study was done at Mass General Hospital and the investigators looked at both histomorphometry in terms of tribecular bone density as well as quantitative CTs as well as DEXA bone density scans as well as biomechanical strength testing and all of those endpoints and parameters showed a benefit with the addition of the PDGF to again increase bone growth and bone density in an osteoporosis animal model. Dr. Lavin, would you like to speak to any stratification that has been done to date with the different populations?

DR. LAVIN: Yes. Philip Lavin. We did look at age as a stratification variable in multi-variate analyses. Now these were just recently submitted to the FDA so I'm looking at Dr. Runner. She may not want me to go into the specifics of those analyses.

DR. RUNNER: The question has been asked. You can respond to it.

DR. LAVIN: Okay. Well, in those analyses, we did do cuts of subjects over and under 50 and the treatment advantages were very consistent in those over 50 as well as those under 50 for the low dose versus the control.

DR. LYNCH: And maybe one final point I could make is again it's not directly the GEM-21S product but certainly Regranex which is recombinant platelet derived growth factor in a topical cream or ointment for treatment of chronic diabetic foot ulcers. It have been FDA approved and therefore, it obviously has been shown to be safe and effective in a chronically diabetic, very severe efficacy diabetic patient and I don't recall the exact data, Dr. Schaumberg is here , but I would suspect that certainly many of those patients are also elderly.

CHAIRMAN SUZUKI: The question I have is with respect to inclusion of tooth extraction sites and perhaps the sponsor can respond. There are many diverse environmental conditions when a tooth is extracted including that of infection which alters the environment in pH sometimes dramatically. Do you have evidence of how your product behaves in such different pH and environmental conditions to be able to justifying including tooth extraction?

DR. LYNCH: Maybe I'll -- Maybe I won't sit down here. Sam Lynch.

DR. LYNCH: PDGF, in general, if I could use this approach to address your question, Dr. Suzuki, is known as a very stable protein and certainly those panel members and investigators that have worked with it can testify to that. It is very stable, for example, to your point in acidic solutions.

In fact, it is the most stable in slightly acidic or even highly acidic environments. For example, many people will store PDGF in one mil or more acetic acid for a long term of storage conditions. It also appears in the literature to be stable more at least than many other molecules too, for example, proteases, and that may explain some of the benefits of PDGF relative to other tissue growth factors in chronically inflamed environments. Thank you.

CHAIRMAN SUZUKI: Thank you. Dr. Cochran.

DR. COCHRAN: I think there's a major difference in some of these indications particularly in the last two. My concern is not so much the stability, but the question comes up, "Do you want to stimulate the cells in a tooth extraction site and in deficient alveolar ridge because that's a bone regeneration site and not a PDL?"

All the data we've looked at today has been directed towards periodontal ligament regeneration. Periodontal ligament regeneration involves cementum, PDL and bone and the last two are strictly bone sites where PDL is not going to be regenerated. Now I don't know if there's any data that would support this material in a pure alveolar defect site.

And then to comment on the top or the two above of that, cystectomy and apoicoectomy, I'm not sure how the labeling reads for the other devices if those indications were included or not, Susan. I don't know if you remember that or not.

CHAIRMAN SUZUKI: Dr. Runner and then Dr. Nevins.

DR. RUNNER: I would have to check on the Emdogain and Peptin P-15. However in other 510-K bone filling devices, HAs and bioactive glasses, etc. it's commonly given the range of indications, however, again they are different in that they are specifically bone void fillers.

CHAIRMAN SUZUKI: Okay. Dr. Lynch.

DR. LYNCH: Sam Lynch, and then I'll turn the podium over to Dr. Nevins. This list of indications was drafted very closely resembling the indications that have already been allowed for β-TCP which is, of course, the matrix used in our product. We don't see any reason why the addition of PDGF shouldn't limit the labeling beyond what has already been cleared by the Agency for the matrix itself.

As we've discussed this morning, PDGF has very beneficial effects on bone cells. It submits a proliferation and came out with great improvement of those bone cells and the revascularization of the site. It's hard to envision from our perspective in a situation where the addition of PDGF would actually be more limiting than the use of labels that have already been approved for β-TCP.

DR. COCHRAN: I would make a comment to that. David Cochran. That certainly I'm sure is true. My concern is just that as a panel member without any data in any of the bone-only site that we just don't know what that effect is. So it just makes us a little more uncomfortable when there's no data to support that.

CHAIRMAN SUZUKI: Dr. Nevins.

DR. NEVINS: Myron Nevins. David, I could only see that the same way. However, if we take the thought process a little bit further, both you and I recognize that the necessity of producing a periodontion is a more differentiated regeneration than that of producing bone by itself. Where the carrier has the ability to be or has been recognized to be a treatment method, it would seem hard to imagine that this wouldn't do the same.

Now what I don't know and I'm unprepared to discuss is what evidence there is that the β-TCP results with the treatment of an extraction mode, but it's very well known in oral surgery we use autogenous bone in extraction sites with relatively little proof. We use allograft. We use xenograft. It's hard to imagine that using the β-TCP together with the PDGF would be any less success than any of them. I understand that there's no study.

DR. COCHRAN: It may not be any better either though. That's the thing. It may be you're adding something that's not giving you any benefit at all.

DR. NEVINS: But some of the others, I'm not so sure it would give you a benefit.

DR. COCHRAN: Right. It think there are people looking at that now.

DR. LYNCH: One maybe further point.

CHAIRMAN SUZUKI: This is Dr. Lynch.

DR. LYNCH: I'm sorry. Dr. Lynch. Again I'll remind the panel of the data in the osteoporosis animals, clearly a pure bone deficiency, not showing the positive benefit effect of PDGF on simulating bone formation in a pure bone site, if you will. Although not exactly the same, we recognize that there are publications as you are familiar with, I'm sure, showing the use of PDGF in combination with, in that case, insulin-like growth factor for treatment of peri-implant bone defects, again, pure bone defects adjacent to dental implants that were canine studies which also suggested a benefit to the addition of PDGF.

CHAIRMAN SUZUKI: Dr. Zuniga.

DR. ZUNIGA: Jon Zuniga. When you add the addition of ***1:57:21 in tooth extraction, you are asking to expand the use of this material into sites that I don't think were tested. For instance, when you add those indicators, you could potentially be exposing your material, PDGF, to nervous tissue, peripheral sinus tissue.

If your indication is for cystectomy or you're referring to tooth bearing areas or you're referring to non-tooth bearing areas, including the cranial facial structures, now you could be potentially exposing brain tissue or joint tissue or synovial tissue to the chemical. If you have cystectomy, it's not usual to remove a five centimeter, ten centimeter, cyst and it may not necessarily be odontogenic. It could be neoplastic.

So you're exposing a growth factor which differentially accelerates osteoplastic and fibroplastic activity to potentially pathological environments. I think there is a little bit more that's involved especially, I think, when some of the testing, it's not mutagenic or teratogenic, but I believe it was found to be a mild irritant and certainly that material placed against the peripheral or even dura could potentially cause inflammatory processes that we don't know.

CHAIRMAN SUZUKI: Do you have a comment? Dr. Lynch.

DR. LYNCH: Just a very brief comment in terms in last point. I believe just a point of clarification that with GEM-21S product, there was no more irritation than in just the TCP alone or any problem with particular graft material. Just another small point is that the Regranex which is used, again recombinant PDGF, for treatment of severe chronic skin wounds involving subcutaneous tissue would potentially also expose the patient to many of the sort of considerations that you were alluding to, for example, the nerve endings and so forth in that area and they have not seen any adverse effects there.

DR. ZUNIGA: Jon Zuniga again. It's a significant difference between the end terminal versus the peripheral trunk.

DR. LYNCH: Okay. Great.

CHAIRMAN SUZUKI: We'll have an open discussion at 3:00 p.m. Comments? Okay, the next question.

MS. BLACKWELL: Number four: "Assurance of safety. Does the information submitted by the sponsor provide a reasonable assurance that the device is safe under the conditions that are used, prescribed, recommended or suggested in the purposed labeling. If the data and the information submitted does not provide reasonable assurances of safety, what information is needed to establish safety for the claimed intended use?"

CHAIRMAN SUZUKI: Okay. This question is now open for panel discussion. Any other comments? Okay.

MS. BLACKWELL: "Assurance of efficacy. Does the information submitted by the sponsor provide a reasonable assurance that the device is effective under the conditions of used, prescribed, recommended or suggested in the purposed labeling? If the data and information submitted do not provide reasonable assurances of device effectiveness, what information is needed to establish that the device is effective for its intended use?"

CHAIRMAN SUZUKI: This questions is now open for panel discussion. Dr. Sharma.

DR. SHARMA: Given that the data we have looked at, the effectiveness is established within each one of the groups. But if we are talking with the labeling and looking at the comparative of this particular device with the control, that is not established. To establish that, one has to be able to design that study and able to prove that in the -- hypothesis that we are able to prove our efficacy point.

CHAIRMAN SUZUKI: Thank you. Other comments? Dr. Zero.

DR. ZERO: Domenick Zero. This is sort of an open question to the sponsors. If you were going to redesign this study, would you choose a different primary endpoint?

CHAIRMAN SUZUKI: Dr. Genco.

DR. GENCO: I wonder if I could be allowed to give a little bit of a history of periodontal endpoints to answer your question. I remember when pocket depth was the endpoint for periodontal studies and there were some real problems with that because it was so highly dependent upon inflammation.

Then we got sophisticated and started using attachment level because we could measure from a fixed point CEJ or the bottom of the restoration to the apical portion of the pocket appropriate with depth. Then I've seen over the last eight or ten years a move to look at the radiograph assessment. You've heard all the reasons. It's probably a better indicator, not definitive, but a better indicator of regeneration. When you're looking in particular at a regenerative product, it makes sense to look at that.

So I see you're plotting sort of the transition of the thinking of the field in terms of endpoints and I mentioned this morning the cardiovascular research that are routinely using composites. So I guess if I wanted to cut down and I came to you next month, I would probably suggest a composite.

I don't think that's justification for going back and doing the study over. I think, in your mind, we have a done a composite for you. I think that you'll get the same result if you do the study over except that instead of saying six-month CAL will be the endpoint, it will be a composite or some other, but it will be the same result. I think it would wasteful. I think you'll be keeping a product from the patients that show clear benefit to require that a new study be done with a new endpoint.

DR. ZERO: Well to answer the question, probably a new endpoint would be -- I'm not asking you to do another study at this stage. I'm just asking the question in a hypothetical basis to understand your thinking and then the logic that you were to choose another endpoint how rigorously would that be and how would that hold up in court.

DR. GENCO: I think what you would have to do would be the study that you're already suggested. I think Salomon asked the question. What is the correlation between the histologic and the radiographic? I think the data, as a matter of fact we talked about that, is probably there to do that analysis on the 10 or 11 patients that Dr. Nevins talked about to show that there is this correlation. Then you could use the radiographic as a surrogate for regeneration with a little more confidence.

DR. ZERO: Domenick Zero. From the literature you've shared in the application, there are some articles on this subject and again I was learning so I read them. There was a number of statements there that the linear bone height and these other bone fill parameters do not correlate well when you go back and do a reflapping and look at the site.

DR. GENCO: Right, I think they are underestimates. They are consistently underestimating, but that's really a different question. If the systematic error is underestimating, it still could be a good surrogate. It's always underestimating by 40 percent, but it's still a good surrogate. If you look from point to point, millimeter to millimeter correlation, you're not going to find it between the flapback measurement of bone, the radiograph and the attachment. You won't find that. But overall, they tend to go in the same direction with the same constant error or difference and I think that's what you'd be looking at.

DR. ZERO: If they did that, they would be well correlated.

DR. GENCO: Yes, they are correlated, but they are not one for one correlation. In other words, two millimeters of radiographicing does not relate to two millimeters of bone on flapback does not relate to two millimeters of attachment gain. One is less -- The level of sensitivity is the bone would show less. The flapback would show more and the attachment would show even more. But I'm sure there would be a constant relationship among that three and that's what you need for a good overall correlation. These are theoretical. I appreciate your allowing me to answer that question.

CHAIRMAN SUZUKI: Thank you, Dr. Genco.

DR. SHARMA: I have one other comment.

CHAIRMAN SUZUKI: Dr. Sharma.

DR. SHARMA: My comment is that you said that you can get the same results if you had a composite endpoint for the next study, but I seriously doubt that it can be exactly the same results because the patient population might be different. There could be several differences.

One of the things I noticed is there were, in the high dose group, more smokers compared to other treatment groups and that may be one of the reasons that efficacy was low compared to other groups. So it's hard to get the same results unless you have exactly the same duplicate thing. It was just a comment.

CHAIRMAN SUZUKI: Okay. Dr. Amar.

DR. AMAR: I'm just going to make a quick comment. I guess this panel is a little bit hung up on a very

DR. STROMBERG: -- was the Chairman of the committee which evaluated PDGF in chronic diabetic ulcers. And we gelt that it was important to exclude from incorporation in this labeling patients who had Grade 4 ulcers, which extends down into bone and ligament. The pre-clinical studies again gave us some concern about correct exposure of PDGF to bone, and I will read from the Regrantix package insert which reads, "The effects of Peckopermin which is PDGF on exposed joints, tendons, ligaments, and bone have not been established in humans. In pre-clinical studies, rats injected at the metatarsal with 3 or 10 micrograms per site approximately 50 to 100 micrograms per kilogram of Veckopermin every day for 13 days displayed microscopic morphological changes of accelerated bone remodeling consisting of periosteal hyperplasia, and sub-periosteal bone resorption and exostosis." I thought this would be useful for you all to hear.

CHAIRMAN SUZUKI: Okay. Thank you. Other comments? Dr. Lynch.

DR. LYNCH: Sam Lynch. Yes, we considered that part of IDE, the FDA raised that very question and that's part of their review of the IDE. And I think we all felt comfortable that all those issues had been addressed from the five years when the Regrainex was approved in December of' 97 to now. And obviously, we've presented a whole wealth of data this morning addressing effects of PDGF on bone.

CHAIRMAN SUZUKI: Okay. Thank you, Dr. Lynch. Dr. Runner.

DR. RUNNER: I was just going to say that that was how we got them to approve the IDE for use.

CHAIRMAN SUZUKI: Okay. Other comments? Dr. Stromberg.

DR. STROMBERG: Well, you must consider that for Regrainex, the safety issue excluded the issue of exposure because the indication excludes Grade 4 diabetic ulcers, so we don't have that exposure-base available to us unless it was prescribed off-label. And in that sense, it's not done within a clinical trial so we don't have any feedback on adverse effects.

CHAIRMAN SUZUKI: Okay. Thank you. If there's no other comments at this time, we'll have a 15 minute recess.

(Whereupon, the proceedings in the above-entitled matter went off the record at 2:13:12 a.m. and went back on the record at 2:19:51 a.m.)

(Missed Audio 2:08 - 2:10)

DR. SHARMA: You had a composite magnifying for the next study, but I seriously doubt that it can be exactly the same results, because if patient population mechanism, there could be several differences. One of the things I noticed, there were more smokers compared to other treatment groups, and that may be one of the reasons that the efficacy was low compared to other groups. So it's hard to get the same results unless you have exactly, exactly the same duplicate thing. That was just a comment.

CHAIRMAN SUZUKI: Okay. Dr. Amar.

DR. AMAR: I'm just going to make a quick comment. I guess this panel is a little bit hung up on very philosophical as well as clinical issue that we all as periodontists have struggled for many years, and I'm still young, but I will struggle probably in the future with it, is that when we look at success, clinical success, and we look at tissue as clinical attachment levels as opposed to how tissue level, is a tooth that is regenerated with say just bone attached, or better bone attached, better in the long term for maintenance, or just clinical attachment levels even at the expense of the long junction apically be more a -- prepared for future recurrence of the disease. And I think that we go back and forth to this issue.

I stand at this point, myself, in the camp of saying that I do believe that a tooth that has regenerated some structure lost as a result of the disease is better prepared for future recurrence.

CHAIRMAN SUZUKI: Thank you. Any other comments on Question 5. Ms. Blackwell, are there any additional questions?

MS. BLACKWELL: That's all.

CHAIRMAN SUZUKI: Dr. Runner.

DR. RUNNER: I think Dr. Stromberg wanted to make one clarifying point about the Midlantic Study. There was some comment about exposure to other tissues, and we just wanted to make one comment about Regrainex. And Dr. Stromberg is from our Center for Bio Drug Evaluation and Research.

DR. STROMBERG: My name is Kurt Stromberg, and I was the Chairman of the committee which evaluated PDGF in diabetic ulcers. And we felt that it was important to exclude from incorporation in this labeling patients which had Grade 4 ulcers which extends down into bone and in ligament.

(Whereupon, the proceedings in the above-entitled matter went off the record at 2:24:33 p.m. and went back on the record at 2:30:28 p.m.)

CHAIRMAN SUZUKI: Okay. I'd like to call the meeting back to order. We will now hold the second open public hearing session. If there are any individuals wishing to address the panel, please raise your hands and identify yourselves at th is time. Okay. Before we proceed with the panel's recommendations, I'd like to invite the FDA and the sponsor to make brief closing remarks. Dr. Runner.

DR. RUNNER: My name is Susan Runner, and I just want to thank you for your input regarding the questions and concerns that we had regarding this application. And I feel that you have answered our questions, and hopefully you'll come to a conclusion.

CHAIRMAN SUZUKI: Okay. The sponsor.

DR. GENCO: Sam has asked me to make some comments, and I appreciate the opportunity.

CHAIRMAN SUZUKI: This is Dr. Genco.

DR. GENCO: Dr. Bob Genco. I, too, would like to thank the panel for the excellent questions, and I'm not being patronizing. I think you got to the heart of the issue. These are complex diseases and treatment is not simple, and experimental designs have to be very creative and very carefully done.

I very much appreciate the clinical trialists and statisticians viewpoint, and I think that they have addressed the important issues. They have forced all of us in science to think more carefully about clinical design. In fact, and I've seen clinical trials now for 25 years, they're much more sophisticated and give you much more information than they did 25 years ago. In fact, the whole science of clinical trial design and biostatistics is at a very high and sophisticated level as evidenced by the fact that many institutions now are developing Ph.D. programs in clinical trial design. It's a field unto its own, and it's very important.

Having said that, the FDA I think understands that, and I think there's a new era with the FDA, and that era is the interaction with the companies. And I think you're seeing that in this project. You're seeing multiple interactions, multiple meetings, multiple memos and letters going back and forth. And it leads to a complex situation where one forgets what was pre-authorized, what was prospective, and what was retrospective, so we have made this slide just to address that issue.

Our understanding, and I think we've checked with the FDA, that all of the end-points were pre-specified prior to the database lock that are listed here; CAL at three and six months, linear bone growth, percent bone fill, GR due to recession, pocket depth and wound healing in three weeks, and the CAL area under the curve.

Now the end-points analyzed -- now those end-points could be considered done before the data blinding was broken, so I think there's certain validity associated with doing those in a blinded fashion on unblinded data.

Now look at the end-points analyzed after the database lock, the composite analysis, the meta-analysis. These again were in discussion with the FDA an attempt to interpret, to get some idea of patient benefit. As I mentioned, particularly the composite analysis gives us an idea of the percent of the target population that benefitted from this treatment. And as we saw using those two composites of CAL and linear bone growth, CAL and bone fill, we came up with 60 to 70 percent of the target population reached our criteria of success which I think were quite rigid. And I remember discussions on the panel sitting where you are, where we had this argument, discussion, animated often - what is clinical significance? I think we have, with the FDA staff, I think have given you hopefully a little insight into a target population which we think is a reasonable representative of the general population. Sixty to seventy percent have actually benefitted from the treatment, so that, in my mind, would argue that this is a clinically significant result.

With respect to all of the results, I think you have to look at them en masse too, in total - even though strictly statistically speaking we have defined a primary outcome variable, and you're absolutely right - that is what we should be looking at. However, the reality is the path of biology dictates, particularly for complex diseases, to look more intensely at the data, look at the secondary outcomes and look at the composite variables. When you do this, you'll see that every single measure is positive, is in the direction of better effect on the patient. The patients are better for having had the treatment.

In summary then, as a clinician, and I've seen patients for 37 years, two years less than Dr. Nevins, but still 37 - I am very excited about this product. Yes, I have a connection with the company. It's an area of my research. I have a bias because I've been involved in the research, but I can tell you this is a very effective product. I would use it on my patients, and I hope that it's on the market for future patients. Thank you very much for listening.

CHAIRMAN SUZUKI: Okay. Before we proceed with the panel's recommendations, I'd like to invite the FDA and the Executive Secretary, Michael Adjodja, to proceed and give us some background information.

MR. ADJODJA: Thank you, Chairman Suzuki. The medical device amendments of the Federal Food, Drug and Cosmetic Act as amended by the Safe Medical Devices Act of 1990 allows the FDA to obtain a recommendation from an expert advisory panel on designated medical device, pre-market approval applications or PMAs that are filed with the agency.

The PMA just stand on its own merits, and your recommendation must be supported by safety and effectiveness data in the application, or by applicable publicly available information. There is reasonable assurance that a device is safe when it can be determined based on valid scientific evidence that the probable benefits to health under the conditions of use outweigh any probable risks. Valid scientific evidence shall adequately demonstrate the absence of a reasonable risk associated with the use of a device under the conditions of use.

There's reasonable assurance that a device is effective when it can be determined based on valid scientific evidence that in a significant portion of the target population, the use of a device for its intended uses and conditions of use when accompanied by adequate directions for use and warnings against unsafe use will provide clinically significant results.

Valid scientific evidence includes well-controlled clinical investigations, partially controlled studies, studies and objective trials without matched controls, well-documented case histories by qualified experts, and reports of significant human experience with the marketing device.

Your recommendation options for the vote are as follows; approvable if no conditions are attached, approvable with conditions. The panel may recommend that a PMA be found approvable subject to specified conditions, such as position of patient, education, labeling changes, or further analysis of existing data. Prior to voting, all the conditions should be discussed by the panel, or non-approvable.

The panel may recommend the PMA is not approvable if the data do not present a reasonable assurance the device is safe, or if reasonable assurance has not been given the device is effective under the conditions of use prescribed, recommended or suggested in the proposed label. If the vote is for not approvable, the panel should indicate what steps the sponsor may take to make the device approvable. Chairman Suzuki.

CHAIRMAN SUZUKI: Okay. Would anyone on the panel like to make a motion?

DR. SHARMA: Yes, I'll make a motion.

CHAIRMAN SUZUKI: Dr. Sharma.

DR. SHARMA: Based on all the data we have seen on safety and efficacy, the device is safe and effective. My condition is that in the labeling we shouldn't have a claim for superiority attributable to this device. I would recommend for approvability.

CHAIRMAN SUZUKI: Okay. Is there a second to the motion?

SPEAKER: I didn't hear the end of it. Can you repeat it?

DR. SHARMA: I will repeat it. The device is safe and effective. The only condition I'm putting is that in the labeling we need to make sure there is no labeling claim for superiority, and we go ahead and approve this device.

CHAIRMAN SUZUKI: Okay. We can discuss this motion if we have a second. We need a second before we can discuss it. And that is a main motion without conditions.

DR. SHARMA: That was about the labeling part, that labeling shouldn't have any superiority claims, has to be more like --

CHAIRMAN SUZUKI: So is that a motion as a condition?

DR. SHARMA: Yes.

CHAIRMAN SUZUKI: Okay. His motion was for approval. Can you rephrase your motion first?

DR. SHARMA: Sure. The motion is to approve the device as safe and effective device. That's the main motion. And the condition is that we should not have --

CHAIRMAN SUZUKI: We don't need the condition at this point.

DR. SHARMA: All right.

CHAIRMAN SUZUKI: Is there a second to the main motion?

DR. AMAR: I second the motion.

CHAIRMAN SUZUKI: Our main motion for approval first. And then we can proceed with the conditions. Is there a discussion on the main motion?

DR. COCHRAN: Jon, I think his proposal is approvable with conditions.

CHAIRMAN SUZUKI: Okay. Then may I have a condition.

DR. SHARMA: The condition is that the labeling should not have any superiority claim.

CHAIRMAN SUZUKI: Okay. Is there a second on the condition, that there's no superiority claim.

DR. AMAR: I second the motion, but I have other conditions.

CHAIRMAN SUZUKI: Okay. We'll vote on this first condition first, and then you can bring up another condition. Is there a discussion of the condition?

DR. ZERO: Domenick Zero. Is it necessary to make such a condition? Is the company asking for superiority claim?

CHAIRMAN SUZUKI: Dr. Runner, can you respond?

DR. RUNNER: I believe that's a labeling issue. I think you can have that as one of your instructions to FDA in terms of the labeling, so that could be considered a condition, that there would be no superiority claims.

CHAIRMAN SUZUKI: Okay. Then let's discuss and vote on this condition first, and then we will go to another condition, if necessary. Ms. Lawton.

MS. LAWTON: Alison Lawton. Let me just ask a clarifying question to that. Are you saying no superiority claims on anything, the control that was used in the study, as well as other products? What are you saying as far as superiority?

DR. SHARMA: No superiority based on the primary hypothesis, which was that this device is superior to the control in the study.

MS. LAWTON: So the CAL end-point only, which was the primary end-point. Is that correct?

DR. SHARMA: Yes, the primary end-point was to look at this device, was for the control device. And it's not about any other device approved or not approved, because that data is secondary data. And that's not from this study.

MS. LAWTON: So the superiority claim would relate specifically to the primary end-point of the study.

DR. SHARMA: That's right.

CHAIRMAN SUZUKI: And there was a second already on this condition. Any other discussion on this condition? Mr. Schechter.

MR. SCHECHTER: This is Dan Schechter. I just wanted to point out two things on this. One is the use of the word "retrospective" by the FDA or whomever may have given the impression that this is kind of created analysis afterwards, when in fact these were all pre-defined end-points, and as the sponsor pointed out, these end-points were chosen before the database lock.

In addition, if you look at the actual regulation for approving with conditions, it's contemplated that further analysis of existing data as it's stated on the slide and in the regulation is a reasonable activity for the sponsor to engage in. And, in fact, that's kind of what they've already done. They had existing data, and merely because they didn't meet the primary end-point, at least from some vantage points doesn't necessarily mean the product is not superior in other respects, so I just want to put that on the record for the FDA to consider when this condition is discussed.

CHAIRMAN SUZUKI: Okay. Other discussion? If not, I'll call the question on this condition. Each voting panel member will indicate yes, no, or abstain on this motion. There are four eligible voting members at this panel, and the Chair will vote only in the event of a tie. I'll begin with Dr. Amar.

DR. AMAR: Yes.

CHAIRMAN SUZUKI: Dr. Zero.

DR. ZERO: Approval.

CHAIRMAN SUZUKI: Only voting on the condition. Dr. Zuniga.

DR. ZUNIGA: Yes.

CHAIRMAN SUZUKI: Dr. Sharma.

DR. SHARMA: Yes.

CHAIRMAN SUZUKI: This condition passes. Dr. Sharma will restate the condition.

DR. SHARMA: The condition is that no labeling claim based on the data presented should be made in the labeling for superiority.

MS. LAWTON: Relating to the primary end-point is what I thought we agreed specifically, because there are secondary end-point results which I think do show superiority, which were important to have in the label, and so that's why I wanted to clarify that it's the primary end-point we're talking about.

CHAIRMAN SUZUKI: Okay. You meant the primary end-point.

DR. SHARMA: Primary end-point.

CHAIRMAN SUZUKI: That's correct, Ms. Lawton, the primary end-point. Okay. Are there other conditions? Dr. Amar, you had one.

DR. AMAR: Yes. I think we discussed this morning about the concentration dependency and I would like to have on the label - it's a label issue condition - there is concentration dependence or I leave it up to the FDA for better word smith as it would be in terms of concentration dependency, in addition to probably having another condition in light of the Regainex comment related to bone defect.

CHAIRMAN SUZUKI: Okay. Can you just make one condition at a time?

DR. AMAR: Concentration dependency.

CHAIRMAN SUZUKI: Okay. That was your motion on concentration dependency. Is there a second before we proceed with discussion? If there is no second, there is no discussion. Then we can have another condition. Okay. That won't be considered as a condition since there is no second. Is there another condition that you'd like to make a motion. Dr. Zuniga.

DR. ZUNIGA: My condition would be that the indications for the use be restricted to the treatment of periodontal and/or periodontal-related disorders. This is based on dental lack of safety regarding interactions with other non-periodontal tissues.

CHAIRMAN SUZUKI: Is there a second to this condition?

DR. SHARMA: Yes, I will second that.

CHAIRMAN SUZUKI: Okay. There is a second to this condition. Is there a discussion on this condition? A labeling issue restricting it primarily to periodontal defects. Is that correct?

DR. ZUNIGA: Periodontal or periodontal-related.

CHAIRMAN SUZUKI: And periodontal-related defects. Okay. Discussion? Okay. If there's no discussion, each voting member of the panel please indicate a yes, no, or abstention. Dr. Amar.

DR. AMAR: Yes.

CHAIRMAN SUZUKI: Dr. Zero.

DR. ZERO: Yes.

CHAIRMAN SUZUKI: Dr. Zuniga.

DR. ZUNIGA: Yes.

CHAIRMAN SUZUKI: Dr. Sharma.

DR. SHARMA: Yes.

CHAIRMAN SUZUKI: Okay. This passes. Any other conditions? Hearing none, we can go back to the main motion, which was Dr. Sharma's approvable with conditions. Is that correct?

DR. SHARMA: Correct.

CHAIRMAN SUZUKI: Okay. Let's review all the conditions then.

MR. ADJODJA: The conditions are no labeling claim of superiority using the primary end-points, and labeling restricted to perio-related defects.

CHAIRMAN SUZUKI: So Dr. Sharma's original motion was approvable with conditions. Mr. Adjodja has just reviewed those two conditions, and there was a second to the motion. We can now proceed with each voting member of the panel indicating a yes, no, or abstention. We are now voting on the main motion with conditions. Dr. Amar.

DR. AMAR: Yes.

CHAIRMAN SUZUKI: Dr. Zero.

DR. ZERO: Yes.

CHAIRMAN SUZUKI: Dr. Zuniga.

DR. ZUNIGA: Yes.

CHAIRMAN SUZUKI: Dr. Sharma.

DR. SHARMA: Yes.

CHAIRMAN SUZUKI: The motion passes. I'd like to ask if Dr. O'Brien or Dr. Cochran have any additional comments?

DR. O'BRIEN: Bill O'Brien. The bone regeneration materials are widely used due to the failure of preventive hygiene and creating periodontal disease. The progression of the periodontal disease can lead to loss of teeth, major problems, and the improvement or the improvement of bone regeneration materials has an important potential impact on periodontal therapy.

Since the clinical benefits have been shown to be effective and safe, it appears that this material will be very useful in periodontal practice. I would add one caveat, that watching the technique film or the video, it appears that it was up to the clinical judgment in terms of the application of the TCP, and with other materials the poracity of the final material, in this case the scaffold, is very dependent on the pressure exerted during the formation of the scaffold, that I would hope that this would be transferred or recommended in the clinical directions.

CHAIRMAN SUZUKI: Dr. Cochran.

DR. COCHRAN: Yes. I would like to commend the sponsor on an extremely well-designed and executed trial. It was one of the most blinded trials that we've seen come before the panel - well done and well executed, and outstanding. That's in large part due to the outstanding investigators that they used.

The proof of principal data, histological data is very solid. I think it's exciting, as Dr. Nevins pointed out, that we see these kinds of histological specimens. I think that's very encouraging.

The radiographic data are very convincing, as well, and I think that gives us a lot of excitement to be able to treat our patients. And I think especially, this is noteworthy when you consider that they used an active control which made the ability to detect any differences quite small.

My only concern, I'd like to raise for the FDA, is in the labeling where they have a discussion of the meta-analysis comparison. It's really not a meta-analysis, it's a comparison to meta-analysis data, and I don't feel that that's particularly necessary to be in the labeling or appropriate actually.

CHAIRMAN SUZUKI: Thank you, Dr. Cochran. Is that agreeable, Dr. Runner, that that can be a labeling issue.

DR. RUNNER: We take all of the comments of the panel into consideration in working out the labeling, yes.

CHAIRMAN SUZUKI: Thank you. At this time, I'd like to ask if either the industry representative or the consumer representative have any comments. Mr. Schechter.

MR. SCHECHTER: This is Dan Schechter. From information that I've gained from the sponsor and the panel, it seems that this particular study, it was almost a model of cooperation between the FDA and the sponsor. And, in particular, Dr. Runner and her staff are to be commended for their assistance to the sponsor. And I think other sponsors faced with future studies should not be afraid of coming to the FDA from day one.

CHAIRMAN SUZUKI: Okay. Thank you, Mr. Schechter. At this time I'd like to ask the four voting members of the panel to indicate their reasons for their decision, beginning with the first panel member who voted affirmative, Dr. Amar.

DR. AMAR: Thank you. I voted for the approval because I believe that this product is as effective as the predicament already in the market. It is safe. In addition to that, the data presented this morning were convincing. They did convince me. The clinical attachment remains the issue, and I've explained that in regard to that particular aspect I think that many products at this point in the market can improve clinical attachment level. What's at stake is the regeneration of the supporting structure, and I think that this product is now capable of doing that, and that's the reason I voted for the approval; approval for the condition. I'm sorry. With condition.

CHAIRMAN SUZUKI: I'd like to next ask Dr. Zero to justify his reason for the decision affirmative.

DR. ZERO: My affirmative vote with conditions was based on the strength of the safety data which was very convincing. In regards to the efficacy, I believe that the study, although not satisfying as primary outcome conditions, overall was convincing that there is the potential for regeneration, which was really the prime objective of the study.

In addition, there was the ancillary data that was presented from the various animal model systems, and the other clinical data was again in composite very convincing, and swayed my decision for an affirmative vote.

CHAIRMAN SUZUKI: Okay. I'd like to next ask Dr. Zuniga for his reasons for affirmative.

DR. ZUNIGA: I believe the health and safety issues were for -- indications for the treatment for using this product were very strong and solid, and based on two FDA approved materials with long history of use.

I think the effectiveness data was very well presented and very supportive in the presentations today and the discussions today, which led me to conclude that this will be effective for a very complex disorder that we're still trying to understand.

CHAIRMAN SUZUKI: Finally, I'd like to ask Dr. Inder Sharma for his reasons for affirmative.

DR. SHARMA: Based on all the data, my recommendation to recommend approval with conditions was based on very safe device, effectiveness even though it was not significant in a comparative way, but there was consistency. There were secondary end-points in the same direction as the primary end-point. And all the data looking at that, I see that it should be made available to patients and they should benefit out of it. Thank you.

CHAIRMAN SUZUKI: Okay. Thank you for your comments. At this time, I'd like to turn the program over to our Executive Secretary, Mr. Adjodja.

MR. ADJODJA: Thank you, Chairman Suzuki. I'd like to clarify that the motion just voted for was for approvable with conditions. The following conditions were voted on; that there should be no labeling claim of superiority using the primary end-point, and labeling should be restricted to periodontal-related defects. The vote was 4-0 approvable with conditions. Before we adjourn for the day, I'd like to remind the panel members that we are required to return -- they are required to return all the materials that were sent pertaining to the PMA itself. Materials that you have with you may be left at the table, any others that you have may be sent back to FDA as soon as possible. Chairman Suzuki.

CHAIRMAN SUZUKI: In closing, I'd like to thank the speakers and members of the panel for their preparation and participation of this meeting. I'd also like to extend a special appreciation to our reviewers, Drs. Amar and Sharma, for leading the discussion segment of this program, and I'd like to echo Mr. Schechter's comments regarding the partnership of FDA with industry, which is really highlighted I believe in the proceedings today.

Since there is no further business, I'd like to adjourn this meeting of the Dental Products Panel. Thank you.

(Whereupon, the proceedings in the above-entitled matter went off the record at 3:01 p.m.)