DEPARTMENT OF HEALTH AND HUMAN SERVICES PUBLIC HEALTH SERVICE FOOD AND DRUG ADMINISTRATION CIRCULATORY SYSTEM DEVICES PANEL 10:10 a.m. Monday, September 11, 2000 Gaithersburg Marriott Washingtonian Washingtonian Boulevard Gaithersburg, Maryland P A R T I C I P A N T S Panel Participants Cynthia M. Tracy, M.D., Acting Chairperson Megan Moynahan, Executive Secretary Voting Members Michael D. Crittenden, M.D. Julie Freischlag, M.D. Consultants (with Temporary Voting Status) Kent R. Bailey, Ph.D. Melvin L. Griem, M.D. Geoffrey Ibbott, M.D. Mitchell Krucoff, M.D. Kenneth E. Najarian, M.D. Tony W. Simmons, M.D. J. Frank Wilson, M.D. Industry Representative Gary Jarvis Guest Robert L. Ayers, Ph.D. FDA Participants James E. Dillard III Donna-Bea Tillman, Ph.D. Bram Zuckerman, M.D. Chris M. Sloan Kimberly B. Peters Henry T. Heaton II C O N T E N T S AGENDA ITEM PAGE Call to Order - Cynthia Tracy, M.D., Acting Chairperson 4 Conflict of Interest Statement - Megan Moynahan, Executive Secretary 4 Sponsor Presentation: Novoste Corporation, PMA P000018, Beta-cath Intravascular Brachytherapy System - Andrew Green, Director, Regulatory Affairs, Novoste Corporation 8 - Burton Speiser, M.D., Director, Radiation Oncology, St. Luke's Regional Medical Center 9, 29 - Jeffrey J. Popma, M.D., Principal Investigator, START Trial, Brigham and Women's Hospital, Harvard Medical School 15, 41 - Richard E. Kuntz, M.D., Chief, Clinical Biometrics Division, Brigham and Women's Hospital, Harvard Medical School 32 FDA Presentation - Kimberly Peters, Leader Reviewer, FDA 43 - Bram Zuckerman, M.D., Clinical Reviewer, FDA 49 Open Committee Discussion 56 Lunch 87 Open Committee Discussion (Continued) 88 Open Public Hearing -- P R O C E E D I N G S ACTING CHAIRPERSON TRACY: Good morning. While we're waiting for our panel to assemble, I'd like to call to order this meeting of the Circulatory System Devices Panel, and our Executive Secretary will read the conflict of interest statement. MS. MOYNAHAN: The following announcement addresses conflict of interest issues associated with this meeting and is made part of the record to preclude even the appearance of an impropriety. The agency reviewed the submitted agenda for this meeting and all financial interests reported by the committee participants to determine if any conflict exist. The conflict of interest statutes prohibit special government employees from participating in matters that could affect their or their employer's financial interest. However, the agency has determined that the participation of certain members and consultants, the need for whose services outweighs a potential conflict of interest involved, is in the best interest of the government. Therefore, a waiver has been granted for Dr. Mitchell Krucoff for his interest in a firm that could potentially be affected by the panel's recommendations. A copy of this waiver may be obtained from the agency's Freedom of Information Office, Room 12A15 of the Parklawn Building. We would like to note for the record that the agency also took into consideration other matters regarding Drs. Krucoff, Cynthia Tracy, Julie Freischlag, Frank Wilson, and Kenneth Najarian. These panelists reported interests in firms at issue, but in matters that are not related to today's agenda. The agency has determined, therefore, that they may participate fully in all discussions. In the event that the discussions involve any other products or firms not already on the agenda for which an Food, Drug and Cosmetic Act participant has a financial interest, the participant should excuse him- or herself from such involvement, and the exclusion will 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 with any firm whose products they may wish to comment upon. ACTING CHAIRPERSON TRACY: All right. At this time I'd like to ask the panel members if they could briefly introduce themselves. DR. BAILEY: I'm Kent Bailey, Biostatistics at Mayo Clinic. DR. CRITTENDEN: Michael Crittenden, cardiac surgeon, West Roxbury VA, Harvard Medical School. DR. SIMMONS: Tony Simmons, Cardiology, Wake Forest University. DR. IBBOTT: I'm Geoff Ibbott, medical physicist at the University of Kentucky in Lexington. MS. MOYNAHAN: I'm Megan Moynahan, Executive Secretary of the Circulatory System Devices Panel. ACTING CHAIRPERSON TRACY: I'm Cynthia Tracy. I'm from Georgetown Hospital, cardiology. DR. FREISCHLAG: I'm Julie Freischlag, a vascular surgeon from UCLA Medical Center. DR. KRUCOFF: I'm Mitch Krucoff. I'm a cardiologist at Duke University Medical Center and the director of device clinical trials at the Duke Clinical Research Institute. DR. WILSON: Frank Wilson. I'm a radiation oncologist, Medical College of Wisconsin, Milwaukee. DR. NAJARIAN: Ken Najarian, interventional radiologist, University of Vermont. DR. GRIEM: Mel Griem, University of Chicago, radiologist and radiation biologist. MR. DILLARD: Jim Dillard. I'm the Director of the Division of Cardiovascular and Respirator Devices at the Food and Drug Administration. MS. MOYNAHAN: I'd like to briefly mention that Robert Dacy, our consumer representative, won't be participating today. He was hospitalized late last week. And we attempted to find a replacement for him but were unsuccessful. So we'll be proceeding today without Robert Dacy. I'd like to also read the appointment to temporary voting status for today: Pursuant to the authority granted under the Medical Devices Advisory Committee Charter, dated October 27, 1990, as amended April 18, 1999, I appoint the following people as voting members of the Circulatory System Devices Panel for this meeting on September 11, 2000: Cynthia Tracy, Tony Simmons, Kent Bailey, Kenneth Najarian, Frank Wilson, Mitchell Krucoff, Melvin Griem, Geoffrey Ibbott. In addition, I appoint Dr. Cynthia Tracy to act as temporary Chair for the duration of this meeting. For the record, these people are special government employees and are consultants to the panel under the Medical Devices Advisory Committee. They have undergone the customary conflict of interest review and have reviewed the material to be considered at this meeting. It's signed by David W. Feigald, Director of the Center for Devices and Radiological Health. ACTING CHAIRPERSON TRACY: Okay. We'll move on to the open public hearing. Are there any parties present who would like to make a presentation at this time? [No response.] ACTING CHAIRPERSON TRACY: If not, then we'll begin with the sponsor's presentation, and I'd like to remind the speakers to introduce yourselves and to state any conflict of interest. MR. GREEN: Good morning. My name is Andrew Green, and I'm the Director of Regulatory Affairs for Novoste Corporation. Madam Chairman, panel members, representatives of the FDA, we are pleased today to present to you a review of the data that supports the safety and effectiveness of the Beta-cath system in the treatment of in-stent restenosis. Today's presentation will include: this overview; a device and procedure summary by Dr. Burton Speiser, a radiation oncologist and investigator in the START trial, as well as the principal investigator in the START 4020 trial; a review of clinical results by Dr. Jeffrey Popma, principal investigator for the START trial; a device performance and training review, again, by Dr. Speiser, investigator in the START trial; a discussion of specific topics by Dr. Kuntz, Director of the Cardiovascular Data Analysis Center, who did the analysis for the START trial; and some concluding remarks by Dr. Popma, the principal investigator. The Novoste PMA P000018 was submitted on April 17, 2000, and requests approval for the 30 millimeter Beta-cath system, specifically developed for intravascular brachytherapy in the cath lab, in the treatment of in-stent restenosis of native coronary arteries 2.7 to 4 millimeters in diameter. This was a randomized, multi-center, placebo, triple-masked trial for in-stent restenosis for the Beta-cath system. We believe that the data that will be reviewed today by Drs. Speiser, Popma, and Kuntz will show that the Beta-cath system has demonstrated effectiveness in significant reductions in all clinical and angiographic outcomes, demonstrated safety and significance reductions in major adverse cardiac events without increased risk of thrombosis to patients, and demonstrated ease of use, short treatment times, minimal exposure to patients and staff, allowing the clinicians to stay in close contact with the patients through the duration of the treatment. At this time I'd like to have Dr. Speiser come up and present the device and procedure summary. He is, again, a radiation oncologist and an investigator in the START trial. DR. SPEISER: Thank you. Novoste has reimbursed normal travel expenses and as well as paid an honorarium for my meeting attendance. I'd like to first discuss the use of Strontium-90. Overall, radiation in any form has a fairly long history for proliferative diseases. External radiation has been used to prevent keloid formation and heterotrophic bone formation. Brachytherapy, specifically Strontium-90, has approximately a 50-year history for the treatment of pterygia, a proliferative disorder of the eyes. The Strontium-90 has been used, as I mentioned, for benign proliferative conditions in the past. The primary mechanism of in-stent restenosis is a proliferative problem in intimal hyperplasia, and the Strontium-90 has an excellent therapeutic ratio, that is, a dose-to-target which is much higher than the dose-to-non-target tissue. Now, while I'm talking about Strontium-90, in effect, Strontium-90 decays to Yttrium-90, and in the decay of Yttrium-90 to Zirconium, we have an energetic beta particle of 2.27 MeV, and that in effect is the issue(?) that we're using. However, for simplicity, I will be referring only to Strontium-90. Strontium-90 has some very advantageous features. It has a dose rate which is quite high, providing very short treatment times, in the range of 3 to 5 minutes, a long half-life of 28.8 years, which eliminates the problem associated with frequent source replacement. Its dose penetration is limited, which matches the dose profile needed in the coronary arteries. It then also means that there's minimal exposure to non-target tissues, as defined as tissues greater than 1 cm from the source axis, and the physician is able to stay with the patient during the entire procedure. This is a graph showing the ISO dose curve with--iridium-192 is the isotope that I most frequently use for brachytherapy and which I have the greatest experience, and in red is the Strontium-90. The bottom is in centimeters, not in millimeters. And what it shows in the shaded area is that in the range of the arteries that we'll be treating, both isotopes provide excellent delivery. The added value for the Strontium-90 is that there's a very rapid dose fall-off, providing extra safety. Now, the exposure on Strontium-90 to the patient is approximately three-tenths of a millirem per procedure, and that's contrasted to an average dose that they received during the fluoroscopy during the procedure of 350 millirem. The radiation oncologist and the interventional radiologist or interventional cardiologist receives approximately two-tenths a millirem and 4 to 16 millirem for the fluoroscopy. In addition, the radiation oncologist receives another 4 millirem per procedure hand dose by handling the device both pre- and post-procedure. Now, the radiation exposure for the patient is quite low. The Strontium-90 component is less than one-tenth of 1 percent of the total dose received during the procedure. The radiation oncologist and interventionalist receive an extremely low dose of the yearly maximum allowable, and the cath lab personnel receive a very tiny dose of their total yearly maximum allowable. I'd like to go into the Beta-cath system and the procedure as it's used. Now, the system is an integrated system that contains a source train of Strontium/Yttrium-90 within a transport device which is then mated to a specific delivery catheter, the Beta-cath catheter, and is complemented with different system accessories. Now, the features of the system is that it's a completely closed system, which allows for controlled delivery and return of the source train such that the sources never make contact with the patient's blood or tissues. In addition, the device has safety interlocks to prevent the sources from inadvertently being discharged from the device unless everything is hooked up completing the closed system. Once again, the treatment time is very short, 3 to 5 minutes, and that allows the physician to remain with the patient during the entire procedure. Now, the State of Georgia has performed a safety evaluation of the Beta-cath system and has issued a sealed source and device registration certificate August 4th of this year, and this certificate has been included in the Nuclear Regulatory Commission sealed source and device registry. Now, the team consists of radiation oncologists, the interventionalist, either the radiologist or cardiologist, a medical physicist, and the complementary cath lab staff. Now, while the interventionalist is completing the angioplasty, the radiation oncologist, with the aid of the physicist, will prepare the Beta-cath system. And this includes over here putting the device in a sterile bag and attaching the syringe--this is a hydraulic system--then next attaching the delivery catheter, and you prime the system to ensure that there's sufficient fluid in the system before the start of the procedure. Next is the prescription of the dose and treatment time based on individual assessment of the reference vessel diameter. Now, the dose is prescribed at 2 millimeters from the center of the source axis, and this is based on individual assessment of the reference vessel diameter. So for an RVD that's equal to or greater than 2.7 millimeters or less than or equal to 3.3, the dose delivered was 18.4 Gray and for the larger vessel it was 23 Gray. Next, the delivery catheter is placed across the injury site, and this shows the marker bands here and here that on the delivery catheter to aid the interventionalist to place it across the appropriate area. And this is done with the aid of fluoroscopy. And next is the delivery of radiation, and this is an animation showing the sources coming out hydraulically. Now they're here, there are markers here and here to further assess the placement of the sources. And the source train remains for approximately 3 to 5 minutes, and the placement is checked periodically with fluoroscopy, so you can see the marker here and the marker band on the catheter just outside of that area to ensure that the source train is at the proper position. In addition, what I'd like to do is, if you can see this arrow over here, there's a light that indicates the amount of pressure. As long as that light or any of the two lights above it are lit, there's proper pressure to maintain the source train at the right position without any source drift. Following the completion, the sources are hydraulically removed back into the transport device. At that point the radiation oncologist removes the system, which is both the transport device and catheter, and the interventionalist completes the procedure. I'd like to introduce Dr. Jeffrey Popma, who is the principal investigator for the START trial, who will discuss the clinical results of the trial. DR. POPMA: Dr. Tracy, panel members--next slide, please?--I have no equity interest in Novoste. I will receive travel expenses but no honorarium for today's meeting. Next slide, please? Again, I apologize to Dr. Krucoff, but what I want to do for the non-interventionalist group is to put a little bit of a perspective on where we stand with in-stent restenosis. We estimate this year that over 725,000 procedures, coronary interventional procedures will be completed in the United States, and in these 725,000 procedures, 80 percent of the patients will receive one or more stents. Now, stents have been very useful for us in the cath lab to prevent restenosis, but, nevertheless, clinical restenosis still occurs in 10 to 20 percent of patients. And what that means is that over 100,000 patients will develop recurrent symptoms due to in-stent restenosis in the United States this year. We have several existing treatment options. Most commonly we perform balloon angioplasty, repeat dilatation within the segment. We tried for a while using a stent within a stent, and we were disappointed that that did not prevent recurrence for in-stent restenosis. We've tried rotational atherectomy, directional atherectomy, ex-(?) or angioplasty, but to date we've had no randomized trials that have demonstrated that this lowers the frequency for recurrence. And, of course, oftentimes, the patient is left with the option of bypass surgery or in many cases a repeat coronary artery bypass operation because the first one has had some limitations. Next slide, please? Now, we have also learned that, depending upon the pattern of restenosis, the recurrence rate will vary. In very focal, discrete lesions, we know from data from Roxanna Mayron (ph) at the Washington Hospital Center that the recurrence rate is 20 percent. However, restenosis often recurs in a much more proliferative pattern, and when the lesion recurs with length more than 10 millimeters, when it's proliferative, or when it's totally occluded, the recurrence rates after treatment of in-stent restenosis range from 35 to 83 percent. It's still a problem. Now, we have some good randomized trial data that has been done. This is from the ARTIST trial that was reported at the ESC last year, and this was a randomized trial of patients coming into these European investigators' clinical practices where patients either received balloon angioplasty, conventional way of treating patients, or a debulking device, rotational atherectomy. And what was found in this I think was relatively important. The restenosis rates ranged between 51 percent to 64 percent, a slight increase in the restenosis rate with rotational atherectomy, but target vessel revascularizations occurred between a third and a half of patients. So it's still a clinical problem, and we have not fixed this with aggressive debulking therapies. So the purpose of the START trial was to assess the safety and efficacy of intracoronary beta radiation using a Strontium-90 source train following successful coronary interventions in patients with in-stent restenosis. I should also emphasize that this is a unique trial design for device trials in that it is a large-scale trial, which was prospectively constructed, including 50 centers, was triple-masked so that the patients, the investigators, as well as the core laboratories did not know which treatment strategy the patient received, and this randomized trial included 476 patients with successfully treated in-stent restenosis. Two hundred and forty-four of these patients were randomized to treatment with Strontium-90, and 232 of these patients were randomized to treatment with placebo. We'll talk a lot about endpoints, and I'm going to give you some definitions for these endpoints in just a moment. But the primary efficacy endpoint of this trial was 8-month target vessel failure. The secondary efficacy endpoint was 8-month angiographic restenosis, the occurrence of in-stent--the in-stent minimal lumen diameter, and the degree of the late lumen loss within the vessel. The safety endpoint was the 8-month major adverse cardiac event rate, and the occurrence of a new aneurysm formation. Let's talk about these definitions, and they will be presented as the primary and secondary endpoints. So I'll ask your patience to move through these. First of all, the primary endpoint will be target vessel failure, and that's defined as target vessel revascularization, the clinical need for a repeat revascularization procedure, myocardial infarction or death that could not be clearly attributed to a vessel other than the target vessel. Major adverse cardiac events were defined as death, Q wave and non-Q wave myocardial infarction, emergency CABG, and target vessel revascularization. For this study and all the studies that we do that evaluate restenosis, we used target vessel revascularization as an endpoint, which was defined as any clinically driven repeat percutaneous intervention of the target vessel or bypass surgery of the target vessel. What this means for the patient is the following: They've had a successful treatment. They go home from the hospital. Somewhere between 3 and 6, maybe 8 months later, they develop recurrent symptoms. They have an exercise test that shows that there's a problem with the distribution in the area that was treated. They come back in and they have an angiogram, and the angiogram shows that there's been a significant renarrowing within the area of either the vessel for target vessel revascularization or of the lesion itself when we use the endpoint of target lesion revascularization. Again, clinically driven, exercise test recurrent symptoms associated with the appropriate anatomy that shows that this is, in fact, clinical restenosis. This trial was supported by the following individuals: Rick Kuntz ran the data coordinating center; Alexander Lansky ran the core laboratory; Peter Fitzgerald served as the IVUS intravascular core laboratory director; Peter Zimentbaum did the EKG core laboratory. I think it's very important to emphasize that this study was watched over very carefully with Tom Ryan as the head of the Data and Safety Monitoring Committee. Tom Ryan is a professor of medicine at Boston University and has chaired up some very important Data and Safety Monitoring Committees. I think it's fair to say that he really--with this study and with the Beta-cath trial--has done a very good job, I think, overseeing this very important committee. The Clinical Events Committee was chaired by Dave Cohen at the Beth Israel Hospital in Boston. Now, let's put this talk about which patients are going to be included in the study. This study included single lesion, single intervention, where there was a greater than 50 percent narrowing within the previously placed stent. The target vessel diameter needed to be between 2.7 and 4 millimeters in diameter. And we'll talk about this in a bit, but the target lesion length was treatable with a 20-millimeter balloon, in which case we used a 30-millimeter source train, or treatable with a 30-millimeter balloon, in which case we used a 40-millimeter source train, in order to have adequate margins outside the injured area. The major exclusion criteria were multi-vessel coronary intervention, a target lesion residual stenosis of greater than 30 percent. The patients needed to have a successful procedure before they were entered into this study. Other exclusion criteria included unprotected left main disease and prior chest radiotherapy. Dr. Speiser has reviewed the dose prescription. It was done visually, and it was 18.4 Gray in reference vessel diameters between 2.4 and 3.3, and 23 Gray in reference vessel diameters between 3.3 and 4 millimeters. We'll talk about antiplatelet therapy at the initiation of this trial. We felt and left the adjunct antiplatelet therapy to the discretion of the physician. Now, I also want to stop and say this is adjunct antiplatelet therapy we're talking about. Aspirin was standard therapy in all patients after their intervention. The adjunct antiplatelet therapy that we'll be discussing are drugs like tyclopadine(ph) or clopitogril(ph), and they're given in addition to aspirin therapy. And at the initiation of this protocol, we left that decision to the investigators. We then learned some important information from Tom Ryan and the Date and Safety Monitoring Committee from the Beta-cath trial and suggested that there might be a benefit in extended antiplatelet therapy in those patients who received a stent. So on March the 19th, we modified the adjunct antiplatelet regimen, and we recommended at that time that there be a minimum of 90 days of adjunct antiplatelet therapy with the placement of a new stent. The results that you will see today with the 8-month clinical follow-up will comprise 96 percent of patients that were included in the study. This is an updated report from what you have seen in your panel pack. Angiographic follow-up was obtained in over 80 percent of patients, and this is superb for a clinical device trial. I'll just go through some very basic clinical demographics. Let me just say at the outset that they were balanced in the two groups: age, gender, the presence of diabetes, prior myocardial infarction, prior bypass surgery, all balanced without differences between the two groups. The reference vessel diameter was 2.77 millimeters and 2.76 millimeters in the two groups, and there was no significant difference in minimal lumen diameter, the pre-procedural percent diameter stenosis, the lesion length, or the percent of vessels that were treated in the left anterior descending artery. We did use debulking devices frequently to obtain a successful procedure in the study, and rotational atherectomy was used most commonly in approximately 40 percent of patients. We also used stents relatively infrequently. Only 20 percent of patients received a new stent. I think it's important to emphasize that we reserved the new stent use for bail-out indications, and those bail-out indications were for a severe residual stenosis or a dissection that occurred after radiation therapy. So the stents that were placed in the study were placed after radiation therapy had been delivered. Again, you got into the study because you were felt to have a successful procedure. Now, I'll just go over just very briefly what antiplatelet therapy the patients actually received if there was not a new stent placed. Forty percent of the patients did not receive antiplatelet therapy or that information was not available to us, which means that 60 percent of patients who did not receive a new stent received subduration of antiplatelet therapy. In patients who had a new stent placed, we did not know or there was no antiplatelet therapy given in 8 percent, but the vast majority of patients received antiplatelet therapy most commonly between 1 to 30 days, 11 percent received 30 to 60 days of antiplatelet therapy; another quarter of patients received more than 60 days of antiplatelet therapy. I'm going to have to take a moment and explain this slide. You're going to see this slide several times, and I think it's going to be important that I just take a moment and try to set the stage. When restenosis occurs in patients who have a stent, it occurs because there's intimal hyperplasia, tissue growth within the stent. And our conventional method of analyzing that tissue growth is by analyzing the stented segment itself. If tissue grows within the stent, it most commonly occurs within the axial length of the stent, and this is where we would determine our stented segment recurrence rate. So that when we do our conventional analyses, we look at the stented segment itself, and we'll talk about those results, and you'll see percent stenosis and restenosis rates that are specifically confined to the axial length of where the initial stent was placed. You'll also see a total analysis segment, and that total analysis segment includes a lot of things. It includes along the axial length of the vessel where the stent was. In your panel pack, you'll see that we also have the numbers for where the injury was with the balloon, where the radiation was delivered, but the analysis segment that you'll see will include the stented segment, a little bit longer length with the tissue that's injured, a little bit longer length that had the radiation, and then 5 millimeters both proximal and distal to that will be included then in this longer axial length. And we'll talk more about this in just a minute. The restenosis rates were significantly reduced within the stented segment, from 41.2 percent to 14.2 percent, a 66-percent reduction. Within the analysis segment, there was also a significant reduction from 45.2 percent to 28.8 percent. Now, these are the clinical outcome measures within the study. We'll talk about the primary endpoint, target vessel failure, about major adverse clinical event rates, and then the clinical indices of restenosis of target vessel failure and target lesion revascularization. Those are what we use. There was a 31-percent reduction in target vessel failure, a 31-percent reduction in major adverse clinical event rates, a 34-percent reduction in target vessel revascularization, a 42-percent reduction in target lesion revascularization--again, all the clinical endpoint parameters that we use to demonstrate efficacy in reduction of clinical restenosis. This is a major adverse event-free survival curve. It shows that the two curves are superimposable out to 90 days. And then at 90 days, which is the typical time course when restenosis occurs, then there's a separation of these curves showing a benefit of Strontium-90 therapy over conventional--placebo-treated patients, and this is a significant difference out to 360 days. We also need to talk about subacute stent thrombosis, and I want to take a moment and do that. There was one patient in the placebo group who developed subacute stent thrombosis within the first 30 days of the procedure. There was no patient in the treated group that had that event. Between 31 and 240 days, which is the time endpoint, there were no occurrences of subacute stent thrombosis in either group. You will hear in just a moment about one patient who had a clinically adjudicated subacute stent thrombosis that occurred at 244 days, and we'll talk about the details of that patient in just a moment. The angiographic total occlusion rate between the two groups was identical, 3.3 percent versus 3 percent in those treated with placebo. So no differences in the occurrence of late total occlusion at follow-up and no differences between the occurrences of the clinical event of subacute stent thrombosis. Let's go into the details about the patient who developed the event at 244 days. On March 2, 1999, the patient's mid right coronary artery was successfully treated in the radiation group. Following radiation treatment, a new stent was placed because the clinical investigator identified a dissection. Despite the fact that a new stent was placed, there was a 48-percent residual stenosis within the treated area, within the stented segment by the core angiographical laboratory, and that suggests to us that this was a suboptimal initial treatment. On 11/1/99, which was 244 days after the treatment, the patient presented with chest pain and EKG changes which were new in posterior-lateral Q waves. The angiogram showed a total occlusion of the mid right coronary artery. The proximal mid right coronary artery then received a balloon angioplasty and additional stent. Now, there are a lot of issues to something like this, but we were most conservative in how we reported this. And it is not clear to us whether or not this represented progression of disease or a new stent thrombosis, but we're going to classify it in our presentation to you most conservatively as a new stent thrombosis. And to summarize, that means that there was one stent thrombosis in the placebo group that occurred within the first 30 days, and even after the 360 days where we follow up patients now, there's only one additional event that occurred in the Strontium-90-treated group. Next slide. Let's talk about the 8-month safety results. There were four deaths in the study--one death in the placebo group and three deaths in the Strontium-90 group. The overall incidence of about 1 percent death rate is what we'd expect for a randomized clinical trial of this size. The occurrence of myocardial infarction, there were seven in the placebo group, four in the Strontium-90-treated group. There was one patient in your panel pack that was classified as having had an aneurysm. When this was reanalyzed by Dr. Lansky in the angiographic core laboratory, it was found that this aneurysm was present at baseline, and it did not significantly change during the follow-up period. Let's just review very briefly a description of the deaths in this study. The first patient was a 77-year-old patient who was successfully treated with radiation on 12/7/98. He died 193 days later of complications following surgical resection of a colonic polyp. The causes of death included his longstanding coronary disease, congestive heart failure, and multi-organ system dysfunction. The next patient was an 83-year-old man who was successfully treated with radiation on 3/4/99. He died 225 days after treatment. The cause of death was metastatic prostate and rectal cancer. The third patient in the radiation group was an 83-year-old patient who was successfully treated with radiation on 3/5/99. He died 160 days later, two days following a left upper lobectomy for lung cancer. The death was reported as a post-operative myocardial infarction. The final patient was a 69-year-old patient successfully treated in the placebo group on 1/22/99. He died 102 days after treatment with the official cause of death reported as cardiac arrest. We'll talk more about some of the specifics of this in just a moment with Dr. Speiser and Dr. Kuntz's next presentations, but I think what we can take home from the START trial is the following: The 8-month clinical outcome summary shows significant reductions in all outcome parameters, which include target vessel failure, major adverse cardiac event rates, target vessel revascularization, target lesion revascularization, and angiographic restenosis. We do not feel that there was an increased risk of thrombosis. There was one subacute thrombosis in the placebo group and conservatively classified one in the Strontium-90-treated patient that occurred at 244 days, and there were no new aneurysm formations found. What I'd like to do is to turn this back over to Dr. Speiser who will discuss device performance and training programs based on what we found in the START trial. DR. SPEISER: Thank you. First, I'd like to briefly go over the device performance. In the START trial, there were 476 patients enrolled. Successful treatment occurred in 467, or 98.1 percent. The causes for unsuccessful treatment were the catheter did not optimally cross the lesion in six patients, or 1.3 percent, and the sources could not be sent in three patients, 0.6 percent. The minor device malformations I will refer to as MDMs. There were 89 patients that had successful treatment with MDMs. Now, the total of these different reported MDMs here equals greater than 89 because some of the patients had more than one observation. If, however, you look at the source transit time of greater than 5 seconds and the source marker drift, this accounted for approximately 90 percent of the cases. One of the cases correctly listed here had in your packet been inadvertently categorized as a reported event in the Beta-cath study when, in fact, it was in the START trial. That patient had a reported event but not a mis-administration. Now, with the observations of these MDMs and experience gleaned from the START trial, Novoste has made device modifications to the Beta-cath system, and those modifications are what was submitted to the FDA in this PMA that is being reviewed. In addition, they were able to create an in-depth training program that incorporates experience specifically from the START trial and have modified the user's manual to include detailed instructions on component connections, pressure tests and monitoring, as well as the manual removal procedure, which is a mandatory procedure for any brachytherapy type of treatment. Next, briefly I'd like to discuss the training program, and this will consist of regional training where the individuals and team will train on the device, procedures, both the treatment and safety, and the roles and responsibility of each team member. Specifically, there will be a hands-on session with the devices to familiarize everybody with the device and detailed instructions for the individuals and team with specific experience from the various trials. There's also cross-training of team members on terminology and professional fields so that there's no confusion in the cath lab and radiation safety training for everybody involved. This is then followed by on-site--which means the facility where the device will be used--training and, once again, first will be reinforced the training on the device, the procedures, and roles and responsibilities. Again, detailed instructions will be given a second time. The procedure will be demonstrated, and that in turn will be followed by a mock procedure conducted in the cath lab; and last, but not least, reinforcement of the radiation safety training for all members of the team. And the last phase of the training program will be a proctored program with an estimate of 3 to 5 procedures that will be proctored to assess the team proficiency with the procedure and system and to advise team and individuals on device use and handling. I have one slide I'd like to put in for long-term safety. The BERT trial now has four-year follow-up, and in this particular trial, what I'd like to do is to show that approximately from 12 months to 48 months, the curve is flattened out, which would indicate that there are no long-term problems that are currently being seen with the radiation arm. Thank you. I'd like to next introduce Dr. Kuntz, who is the Director of CDAC, who will be discussing specific clinical topics relevant to the START trial. DR. KUNTZ: My name is Rick Kuntz. I'm an interventional cardiologist at Brigham and Women's Hospital. I'm also the Director of the Academic Contract Research Organization at Harvard Medical School which conducted this trial. I don't have any equity in this company or any other medical device or drug company, and I'm not being paid for my presentation today. Next slide? I'd like to focus on two issues here that I think require further explanation. One is the clinical impact of the minor device malfunctions, and the second is an analysis of the edge effect. Next slide? Starting with the MDM analysis, as Dr. Speiser has shown, 87 percent of those patients classified with an MDM had a radiation-related subcategory of either source drift or a prolonged transit time greater than 5 seconds. The remainder of the MDMs, such as inability to deliver the delivery catheters and others, did not deal with radiation issues. Next slide? We attempted to try to determine whether the classification of an MDM, as written in the protocol, had any substantial clinical consequences. So the first thing that we did was look at one of the primary endpoints, that is, major adverse cardiac event rates at 240 days, and compare them between those patients classified as having either drift or increased transit time compared to those patients who fit within the guidelines. This is an analysis of the placebo patients, and remember, nobody knew if they were placebo or radiation, so drift or increased transit time could occur on both sides. If we look at the placebo group, there was no substantial or significant difference between the incidence of MACE or between those patients classified as having drift or transit time compared to those without. We then evaluated the same endpoint for patients assigned to the active arm. Again, we found no significant difference in the incidence of MACE between those with drift and transit time MDMs versus those without. However, you'll notice there that there may be trend, that is, that there was a tendency for an increased estimate rate associated with patients with drift or transit time MDMs. The difference that was seen here did not reach statistical significance, but we tried to understand what might explain the trend. The first thing we evaluated was the component of the MACE that we felt was the actual safety issue, that is, the occurrence of myocardial infarction or death. What we found was that there was absolutely no evidence of an increased risk of MI or death for patients who had been classified as having drift or increased transit time. There were zero events in those classifications compared to those without drift or those that fit the criteria. So 3 percent of these MIs and deaths overall occurred in patients who met the criteria and protocol and zero occurred in those who were classified as having an MDM. Clearly, there is no significant difference here and no evidence that the drift or transit time issue increased the risk of myocardial infarction or deaths. Next slide? So what explains the difference that we see in this? And the difference is explained by the incidence of target vessel revascularization or the other component of MACE. Now, whether that difference is going to be substantial or clinically important or not leads us to conclude that possibly if there is a difference, it was a difference in the efficacy endpoint of restenosis, and that's it, not in the safety issue of death or myocardial infarction. We can postulate that possibly drift or increased transit time might reduce the overall deliverable radiation effect. It may be--it's something that you can imagine, and maybe that explained the difference in reduction in efficacy. Then, again, there may be no difference at all because the p-value for this is 0.11. But the important point is that when we look at the incidence of major adverse cardiac event rates between these two subsequent applications, we found no evidence that we exposed the patients to an increased risk of a safety endpoint of death and myocardial infarction. Next slide? In order to evaluate whether restenosis differences could have occurred because of the tendency for increased target vessel revascularization was evident in that initial analysis, we looked at another measure of restenosis using quantitative angiography. And in this case, we compared the overall restenosis rates of patients using quantitative angiography between those classified without a source drift or transit increase and those with the MDM. We found absolutely no difference in that measure of restenosis overall. So our conclusion, therefore, is that source drift and source transit greater than 5 seconds were prospectively collected and possibly could have been very conservative measures and potentially ambitious goals, but were still collected as events and identified as primary minor device malfunctions. The clinical impact of this classification of MDMs demonstrated no statistical difference in any safety measure of the Beta-cath system in the treatment of in-stent restenosis; and, moreover, the sponsor has proposed minor device modifications and training measures to reduce the occurrence of these sort of drifts and transit increases. The next issue I'd like to address is the issue of edge effect. As you recall, Dr. Popma's presentation demonstrated that all measures of restenosis showed a significant benefit for patients exposed to radiation therapy compared to those exposed to placebo. But the impact was greatest in the analysis that was confined to the stent area compared to the analysis that was liberated to the wide analysis segment. One of the questions is that if we start to see more restenosis on a wider measure of restenosis compared to one confined within the stent, is there any activity occurring at the areas outside the stent that are measured by the analysis? And this has been brought up by many investigators in the past as a potential problem associated with radiation therapy called edge effect. And there have been a lot of postulated ideas that the radiation therapy can cause narrowing itself in areas that aren't treated, and there may be areas termed geographic miss in which the balloon injures the artery and there's a fall-off of radiation that may cause increased narrowing. So we were very curious to understand whether this analysis was, in fact, accurately depicting edge problems or it was an artifactual result from the limitations of this analysis and may represent nothing at all. Next slide? So in order to approach this, we have to review again how the initial analysis is done. As Dr. Popma showed you earlier, this is a conventional analysis that's been used for all coronary treatments in the last 15 or 20 years; that is, we tend to measure restenosis based on the location of the minimum lumen diameter, and that's how conventional angioplasty has been evaluated. So that if we look at patients and measure restenosis defined by narrowing within the stent segment, a quantitative angiographic technique used by Dr. Lansky at the core laboratory would identify the area of most narrowing and would tell us what that narrowing is and tell us where it's located. If that narrowing is greater than 50 percent of some reference value, we call that restenosis. The other analysis that we can do is to actually measure the minimum lumen diameter across a very wide area. In this case, we called that the analysis segment. Again, we would detect just one single minimum lumen, and we would tell where the location is. What we found is that if you measure the minimum lumen diameter, it tended to be located in the stent for the vast majority of patients assigned to placebo. If we looked at the minimum lumen diameter for those assigned to active arm, a fair substantial minimum lumens were located outside the stent and not in the stent. So we attempted to understand whether there was an increased propensity for radiation therapy to cause more narrowings outside or if, in fact, we were just unmasking disease that was already there. Next slide? So Dr. Lansky re-evaluated the data set with a specific analysis looking at edge effect, and what was done here was that the measurement of restenosis was confined to just the edges of the analysis segment and not the stent itself to specifically address this issue. So at the source end, both proximal and distal, an analysis centered on that end going 5 millimeters on each side was performed in which the minimum lumen diameter was defined, both proximal and distal, and the incidence of restenosis was measured between the two groups. And what she found was that there was no significant difference between restenosis measured both proximally or distally, depending on the assignment of patients to placebo or active arm. And I think this definitively shows that there was no significant increase in edge narrowing seen in patients assigned to radiation therapy. Next slide? So the question is how did we get that disparity between our initial analysis, which showed a 66-percent reduction when we measured restenosis within the stent, compared to 36 percent when we measured restenosis within the analysis segment. We think we can understand how that artifact might have occurred. In a normal vessel that gets treated, most of the narrowing occurs initially within the stent, hence in-stent restenosis. After treatment, we generally clear out the entire lumen with either aggressive balloon angioplasty or debulking device. And then the patient is assigned to either placebo or active therapy with the radiation source train, and then they're followed up six months later to see what happens. Next slide? In patients assigned to placebo, what we found was that there was narrowing that occurred normally on the edges and in the middle, but the vast majority of narrowing and the pattern of narrowing occurred mainly within the stent. And this has been seen in other in-stent trials and even in de novo stent lesions; that is, the response of restenosis is a little bit on the edges, and the majority of it occurs within the stent. So when we do analysis of where the minimum lumen diameter is, it tends to occur mainly within the stent for patients assigned to placebo. In the radiation therapy group, we get the same degree of narrowing on the edges, but a profound reduction in restenosis in the middle because that's where the targeted therapy was. This may be familiar to radiation oncologists with the concept of central control, where once you take care of the central control of a tumor, you may start to realize peripheral disease starts to show up. Very similar concept here in that what we see here is that an effective therapy may unmask the already present narrowing that has occurred is identical between the two groups. So as Dr. Lansky showed us, the two narrowings and the size were the same, but if you have an effective therapy in the middle, you may unmask the occasional case of patients who have 50-percent narrowing on the sides. So in that analysis you will see that some of the restenosis occurs in the stent and some of it occurs on the side when you have an effective therapy in the middle. Next slide? So the conclusion of the edge analysis is that significant treatment effects seen between the analysis and stent segments was potentially due in part to injury of the radiation therapy or the masking of progression of normal restenosis seen with any cardiac treatment unmasked by effective radiation therapy. Now, given both of those two possibilities, the data really supports the masking issue of progression of disease rather than an induction of narrowing occurring by radiation therapy. I think now I'd like to turn the podium over to Dr. Popma to make some conclusionary statements. DR. POPMA: Go to the next slide. I think in very rapid sequence you've heard that there is a medical need to treat in-stent restenosis. It is a problem for patients. It's a problem in all of our clinical practices. What we've demonstrated in this trial are the following: First of all, the START trial was the largest trial of its type that used randomization, used triple-masking, used placebo-controlled to demonstrate its conclusions. And the summary of what we've heard so far was that the pre-specified hypotheses were all achieved with statistical significance. Target vessel failure was reduced by 31 percent. Major adverse clinical events were reduced by 31 percent. Target vessel revascularization was reduced by 34 percent, and target lesion revascularization was reduced by 42 percent. On the angiographic analysis, we saw that the pre-specified restenosis hypotheses were achieved with statistical significance. Within the stented segment, there was a 66-percent reduction in angiographic restenosis. And with the analysis segment, there was a 36-percent reduction in angiographic restenosis, both highly significant. We also learned that the treatment with Strontium-90 in the START trial was safe and that there were no differences in the occurrence of death or myocardial infarction between Strontium-90-treated patients and placebo-treated patients. There were no differences in late thromboses, there were no differences in total occlusions, and there were no differences in the occurrence of new aneurysm formation. So what we would conclude from this trial is that the Beta-cath system has been shown to be safe and effective for the treatment of in-stent restenosis. MR. GREEN: That concludes the presentation by Novoste Corporation for the Beta-cath system. ACTING CHAIRPERSON TRACY: Thank you. I'd like to ask the panel members if they have any brief clarifying questions they want to ask. We'll have much more time later for discussion. Anybody? Any brief question? [No response.] ACTING CHAIRPERSON TRACY: Then we'll move on to the FDA presentation. MS. MOYNAHAN: While FDA is setting up, let me mention that we've been joined by Dr. Robert Ayers of the Nuclear Regulatory Commission. Dr. Ayers will be participating as a guest today, and he is identified as a coregulator of the use and licensing of the Novoste Beta-cath system. MS. PETERS: Good morning. My name is Kim Peters, and I'm a biomedical engineer in the Interventional Cardiology Branch of the Office of Device Evaluation. I am also the leader reviewer for the Novoste Beta-cath system PMA submission P000018. Today, Dr. Bram Zuckerman, the medical officer for this submission, and I will present the FDA summary for the Beta-cath system. This presentation will identify the FDA review team members, provide a brief summary of the device description, provide a summary of the non-clinical tests conducted on the Beta-cath system, provide a summary of the clinical investigation of the Beta-cath system, and identify the FDA questions for the panel. Members of the FDA review team include Dr. Sabu Subramanian and Dr. Bram Zuckerman, both from the Office of Device Evaluation; Mr. Tom Heaton from the Office of Science and Technology; Mr. Gary Kamer from the Office of Surveillance and Biometrics; and Ms. Marianne Linde from the Office of Compliance. As described during the sponsor's presentation, the Beta-cath system is comprised of the Beta-cath delivery catheter, the transfer device, the source train, and system accessories. The Beta-cath deliver catheter is a sterile single-use catheter that provides the path through which the source train is delivered to and retrieved from the treatment site. The catheter includes three lumens to allow for the passage of the guide wire, source train, and hydraulic fluid. At the distal end of the catheter, the source train and hydraulic fluid lumens are closed, while the guide wire lumen remains open to the vasculature. The distal end of the catheter also features two radiopaque markers that define the treatment zone of the catheter. The transfer device stores and shields the source train when not in use and controls the hydraulic delivery and return of the source train during the treatment procedure. The transfer device features a series of components intended to protect the health care workers and patient from unnecessary radiation exposure, either by shielding the source train or maintaining proper position of the source train. The transfer device also features a series of components intended to regulate, direct, and manage the hydraulic fluid that controls the delivery and return of the source train. The source train consists of a series of individual, cylindrical, sealed radioactive sources with an inactive gold marker at each end of the train. The radioactive sources are Strontium-90 seeds encapsulated in stainless steel. The system accessories include a procedure accessory pack, an emergency storage container, a response kit, and a medical physicist kit. These components are intended to facilitate the operation of the system during the clinical procedure, permit temporary storage of the system in the event of a disrupted procedure, and facilitate handling of the source train if located outside the system. The clinical investigation for the Beta-cath system, the START trial, was conducted with the Alpha III and Alpha IV models of the transfer device. Approximately 83 percent of the clinical data was obtained using the Alpha II model, with the remainder of the data being obtained using the Alpha IV model. Modifications were made to the transfer device in response to reports of device malfunctions during the clinical investigation and to improve the system function. The main difference between the Alpha III and Alpha IV models is the addition of the LED pressure indicators. The pressure indicators provide feedback to the user regarding the pressure necessary to remain the source train at the treatment zone and to send or return the source train to and from the transfer device. The indicators also advise the user when excessive pressure is being administered. The Alpha IV revision two model of the transfer device is the subject of the PMA submission. No clinical data was obtained using this model. As noted in the FDA's summary, the Alpha IV revision two model mainly includes refinement to some of the electronic circuitry and indicators. The Alpha IV revision two model also includes a modification to prevent the transfer device gate from inadvertently being locked prior to the delivery of the source train. FDA believes that these modifications can be evaluated through bench testing. Optional accessories of the Beta-cath system include an introducer sheath and an extension tubing set. Excessive hemostasis valve tightening can restrict the movement of the sources in the Beta-cath system. An optional component, the arrow introducer sheath, may be used to increase the resistance of the catheter to collapse when compressed with the hemostasis valve. The optional extension tubing set provides an additional fluid management system for use during the clinical procedure by allowing two control syringes to be connected to the Beta-cath system. During the START trial, both the 30-millimeter and 40-millimeter delivery catheters and source trains were used. The difference between the 30-millimeter and 40-millimeter delivery catheters is the marker spacing at the distal end of the catheters. The spacing identifies the treatment zone of the catheter. The difference between the 30-millimeter and 40-millimeter source train is the number of active source seeds. The 30-millimeter source train includes 12 seeds, and the 40-millimeter source train includes 16 seeds. Due to the limited clinical data available for the 40-millimeter model, only the 30-millimeter delivery catheter and source train are subject of the PMA. A series of in vitro tests were performed to evaluate the mechanical integrity and function of the Beta-cath system and each of the individual components. FDA is working with the sponsor to resolve questions regarding this testing information. The delivery catheter is the only patient-contacting component of the Beta-cath system. Biocompatibility testing completed in accordance with the ISO Standard 10993 demonstrated that the catheter is non-toxic and non-hemolytic. Electrical safety, battery, and electrode magnetic compatibility tests were conducted in accordance with applicable voluntary standards. All test requirements were met. As discussed in the FDA summary, the sponsor has conducted two animal studies using both oversize stent injury and balloon overstretch injury pig models. The results of the animal tests show no difference in restenosis between the Beta-cath system and the control. With regard to the source dosimetry, FDA is working with the sponsor to resolve questions with the dosimetry information and dosimetry labeling recommendations. The sponsor has provided data from three clinical studies: the Beta Energy Restenosis Trial, the Beta Radiation in Europe Trial, and the Stents in Radiation Therapy Trial. The Beta Energy Restenosis Trial was a U.s. feasibility study evaluating the use of beta radiation following PTCA and de novo lesions. Eighty-three subjects were enrolled in the study. The Beta Radiation in Europe Trial is a multi-center, non-randomized registry that is studying the use of the Beta-cath system following PTCA or stenting of de novo lesions. One hundred fifty patients were enrolled in the study. Summaries of these two trials are provided in the panel pack. The Stents in Radiation Therapy Trial is the pivotal study for the evaluation of safety and effectiveness of the Beta-cath system. Dr. Zuckerman will provide an overview of the trial design and a summary of the results. DR. ZUCKERMAN: Good morning. My name is Bram Zuckerman. I'm a medical officer cardiologist with the Food and Drug Administration. The START trial is the key data set for consideration of this PMA. The sponsor has previously outlined the major elements of the protocol and shown key results. The agency would like to discuss several aspects of this trial prior to presentation of the panel questions. Next slide, please? The START trial was a well-designed trial. A large number of patients were randomized to beta radiation or placebo treatment. Patients, investigators, and core labs were blinded to treatment assignment. Next slide, please? Inclusion criteria indicated that a patient needed a symptomatic case of in-stent restenosis with a reference vessel diameter between 2.7 and 4 millimeters. Visual estimation was performed of reference vessel diameter because this mimics real-world clinical practice. There are few sites in the United States that routinely use online quantitative coronary angiography, QCA, or intravascular ultrasound. But as expected, we saw the discrepancy between our visual reference vessel diameter results and the QCA results. For example, the mean QCA result reported from the core lab for all vessels was 2.76 millimeters. The other point in interpretation of this trial is as noted by Ms. Peters: 95 percent of the data pertains to the 30-millimeter source train. Next slide, please? Vascular brachytherapy represents a new technology for treatment of in-stent restenosis with an unclear risk/benefit profile. As such, a superiority hypothesis was chosen with a primary clinical endpoint: 8-month target vessel failure. Target vessel failure is a conservative endpoint that includes death, non-fatal myocardial infarction, and target vessel revascularization. The angiographic and ultrasound data provided in this panel report should, therefore, be viewed as supporting data that helps to mechanistically explain the effects of vascular brachytherapy. Next slide, please? On this slide, we have acute results presented. You've seen much of this before presented by the sponsor. The only difference with these slides and the next two are that we will have also the 95-percent confidence interval of the differences presented. The key point for acute results was that a respectable post-procedure percent diameter stenosis was obtained, and high rates were reported for both device success and procedure success. Next slide, please? There were nine cases of device failure. These cases have been individually reviewed in your panel pack report. Next slide, please? Please note, however, the definitions used for device failure and procedure success--I'm sorry, for device success and procedure success. Device success was defined as successful placement of the Beta-cath system, and procedure success was defined as a post-procedure percent diameter stenosis less than 50 percent without the occurrence of major adverse cardiac events during the hospitalization. Hence, such problems as initial device failure, minor device malfunction, in all cases where the bail-out box was used emergency, would not necessarily be captured in those two preceding definitions. A balanced assessment of device performance needs to include these variables and results as well as device success and procedure success results. Next slide, please? Eight-month safety results are shown on this slide. At 8 months there was no difference in the rates of death, myocardial infarction, stent thrombosis, site thrombosis, total occlusions, or aneurysms. Two points need to be noted. Firstly, a minority of patients, 20 percent, were restented during this trial. The restented population may be the population at greatest risk for long-term safety problems. The second point is that these are 8-month safety results. As previously noted and as noted by the asterisk on the bottom of the slide, there was already one stent thrombosis reported at greater than 240 days in the beta radiation arm. Next slide, please? Eight-month effectiveness results are reported here. The primary endpoint, target vessel failure, was reduced at 8 months by beta radiation treatment. This was a robust result. Multiple other clinical and angiographic markers of restenosis were reduced by beta radiation treatment, as noted in your panel pack and on this slide. Next slide, please? So, in conclusion, the primary endpoint, target vessel failure, as well as selected clinical and angiographic endpoints, were all reduced by beta radiation treatment. There was no difference at 8 months in the incidence of death, myocardial infarction, stent thrombosis or total occlusion. Device-related malfunctions were observed. MS. PETERS: FDA would like to obtain panel input on the following questions: The original START protocol suggested that the institutional standard of care for antiplatelet therapy after source treatment be utilized for patients who were restented or received PTCA. This regimen was modified based on recommendations from the Data Safety Monitoring Board. A report of the antiplatelet therapy usage during the START trial is provided in the addendum to the START clinical report on page 3. No incidents of stent thrombosis were reported during the START trial. Question 1: Based on this information, please discuss your recommendations for the antiplatelet therapy for patients who receive a new stent and for patients who do not receive a new stent. Table 31 of the START clinical report and the addendum to the START clinical report on pages 13 through 35 identify the device failures and malfunctions that occurred during this study. Question 2: Please discuss the clinical importance of the device failure and malfunction events and the evaluation of the safety and effectiveness of the Beta-cath system. As demonstrated by the results included in Table 1 of the START clinical report, the incidence of the primary endpoint, target vessel failure, was significantly lower at 8 months for the treatment arm compared to the placebo. The incidence of target vessel revascularization, target lesion revascularization, and major cardiac adverse events were also significantly lower over the 8-month follow-up period for the treatment arm compared to the placebo. No incidents of stent thrombosis were detected in the treatment arm, and the frequency of total occlusions was comparable between the treatment and placebo arms. Question 3: Please discuss whether you believe the probable clinical benefit of the radiation treatment outweighs the probable risk of death, myocardial infarction, late total occlusion, and late stent thrombosis posed by the device in the intended patient population. One aspect of the premarket evaluation of a new product is the review of its labeling. The labeling must indicate which patients are appropriate for treatment, identify the product's potential adverse events, and explain how the product should be used to maximize benefits and minimize adverse effects. Please address the following questions regarding the product labeling. Question 4-A: Please comment on the indications for use section as to whether it identifies the appropriate patient population for the treatment with the device. Question 4-B: Please comment on the contraindications section as to whether it identifies all conditions under which the device should not be used because the risk of use clearly outweighs any possible benefit. Question 4-C: Please comment on the warnings and precautions section as to whether it identifies all potential hazards regarding device use. Question 4-D: Please discuss whether any improvements could be made to the labeling to help minimize the occurrence of device failures and malfunctions as discussed under Question 2. Question 4-E: Please comment on the remainder of the device labeling as to whether it adequately describes how the device should be used to maximize benefits and minimize adverse events. Question 4-F: Does the panel have any other recommendations regarding the labeling of the device? A summary of the physician training program has been provided in Section E of the panel pack and in the addendum to the START clinical report on pages 18 through 25. Question 5-A: Please discuss any improvements that could be made to the training program to help minimize the occurrence of device failures and malfunctions as discussed under Question 2. Question 5-B: Please identify any other important elements that should be contained in a physicians' training program for this device. The panel pack includes the available one-year data from the START trial, the available one- to four-year data from the BERT feasibility trial, and the available data from the BRE European trial. Question 6: Based on the clinical data provided in the panel pack, do you believe that additional clinical follow-up data or post-market studies are necessary to evaluate the chronic effects of intravascular radiation administration? If so, how long should patients be followed, and what endpoints and adverse events should be measured? This concludes the FDA's summary presentation. ACTING CHAIRPERSON TRACY: Thank you. At this point we'll move to the open committee discussion, and I'd like to ask Dr. Simmons to begin the discussion with his review, and we'll go around the table after that. And I'd just remind the panel members to restate their names and speak into the microphone when they're asking their questions or making their comments. Dr. Simmons? DR. SIMMONS: Thanks. Well, it's a very nice presentation by the sponsor and the FDA. I have a few questions. The data that the sponsor presented I think presented a better clinical outcome than the data that's in the panel pack or that the FDA presented. I mean, the 8-month stent segment restenosis rate was 14 percent versus 14 percent, which is a 27-percent reduction. However, the target lesion revascularization at 240 days was 86 versus 76, and you were presenting 31 percent. So it's actually about a 9- to 10-percent reduction in total vessel failure at the 8 months--is that right?--as opposed to the 30 percent that I saw on your slides? There's quite a difference between 30 percent versus 9 percent. DR. BAILEY: Is it possible he's talking about a relative reduction? DR. SIMMONS: That's what I'm interested to see. DR. POPMA: I wonder if we could just go back to our slides and the presentation just very briefly, if that will help find out where the discrepancies are. I should also note that it is difficult to take numbers out of the event-free survival curves and then put them back into the rates that are measured. The 240 days is absolutely accurate, but some of the data that's in the pack is extended out further than that and before that. So I want to make sure of the 240-day endpoint for event-free survival curves. Let's go back and discuss those. If you can put up, Richard, just the graph that has the four reductions of TVF, MACE, TVR, and TLR, the one slide with four graphs. ACTING CHAIRPERSON TRACY: Could you also introduce yourself for the-- DR. POPMA: Sorry. I'm Jeffrey Popma. Now, we can go by these one by one, if you like. I think that these absolute rates have been relatively consistent and should be consistent to the rates that are reported in the panel pack. DR. SIMMONS: Well, they're not, actually. Why don't you go to page--let's just get on the same page here. Go to your panel pack, page 414, and you've got TVR-free at 240 days, 81.4 percent in the treated group versus 72 percent in the placebo group, so the difference is only 9 percent. DR. POPMA: I got it. I'm going to defer to Rick Kuntz, who did the statistical analysis for this. DR. SIMMONS: And that's different than 32 percent. DR. KUNTZ: The numbers that you're referring to is the absolute difference. This is a relative difference. If you look here, the difference in TVF there is 26 minus 18, which is 8 percent. And then you're seeing a difference in the MACE-free or TVF-free of 9 percent, which is consistent with the differences between sensoring and non-sensoring survival analysis versus a discrete analysis. Maybe--Dr. Bailey, do you understand what I'm talking about there? DR. BAILEY: In other words, 9 percent is 31 percent of 26. DR. KUNTZ: Right. So the difference of 9 percent in the TVF and 8 percent in the event-free survival or vice versa is pretty typical when you're using a sensored analysis for survival versus one that is a discrete analysis at 240 days. To me that's pretty clear. I'm not quite sure--so 31 percent refers to the relative difference you see here, but the 9 percent you're referring to is the absolute difference. The absolute difference here is 8 percent as well. DR. SIMMONS: Okay. So what we're actually talking about is an 8-percent improvement clinically. DR. KUNTZ: Right. DR. SIMMONS: A 9-percent improvement clinically. DR. KUNTZ: Right. Again, if we refer to the MACE or the TVF here, the absolute differences there are in the 8 to 9 percent range. DR. SIMMONS: Okay. DR. KUNTZ: Which is similar to the event-free survival we're seeing there. DR. SIMMONS: All right. You know, I understand that most of the device failures did not result in any complications to the patient and any real damage. But, you know, the physicians doing your clinical trials, they're more skilled physicians. They've got more back-up. They've got more interest in what's going on. But we're still talking, I think, a fairly amazing, almost 20-percent incidence of some device failure, either minor or major. And even though it didn't appear to have any clinical impact in the physicians that were performing the study, I'm just wondering what happens when physicians who do two angioplasties a month and don't have any company representative around, whether that's really going to translate into no complications to the patient. In addition, if we look at page 221, where Dr. Zuckerman analyzed the differences between your number three and your number four revision, the incidence of drift and everything else didn't really seem to be affected by putting the lights on the box and changing your connector. So I'm just wondering if one of the clinicians might address this issue, my concerns. DR. SPEISER: Burton Speiser, radiation oncologist. I think the primary problem with both the drift and the transit time is the lack of observation of the observer. It's a very simple process, either by feel before the LED lights were added, or by watching the LED lights. The primary problem is the training of the individuals such that they pay attention to that, and I don't want to put down my colleagues too much, but it isn't a very hard job really to keep the pressure in there. Part of the training process is to use the device and ensure that they know how to keep the pressure constant, which is, in effect, a very easy process. DR. SIMMONS: I don't know. I mean, these are very motivated, highly skilled people. If they can't do it, do you really expect people in other cath labs who aren't as motivated to be able to do it? DR. SPEISER: I think the problem is primarily with the radiation oncologist who is in the cath lab for the first time, feels in a foreign territory, and I think what is necessary is the feeling that they're comfortable there and they know how to use the device. And that will take training. Most radiation oncologists probably have never stepped into a cath lab before, and I think that's why the training and doing the mock procedures is quite important to make sure that each radiation oncologist feels comfortable. If you play with the device, you'll be surprised how simple it really is, so that I do have a little difficulty trying to explain a 20-percent rate of transit problems and drift, when, in fact, if you use the device in a mock session, it's very easy to send the sources out and keep them in station or in place. Now, I know that may not be answering the question, but I do have difficulty in understanding why many of my colleagues had difficulty with it. MR. GREEN: I'd just like to also add to that that the trial did actually include 50 centers, and we believe that was--we tried at least to develop that trial with those centers so that we included both, if you will, the normal usual suspects or the normal usual trial centers that would be seen in these type of trials so we had that component of an understanding of clinical trials, as well as the regional and, if you will, everyday hospitals that would use such a system. We did learn that if you look in your panel pack at the section on the device observations, for instance, with regard to what we call the manual removal procedure or what we term there the bail-out procedure, that as time went on and as enrollment went up, there was a decrease in the rate of the use of that procedure. So, therefore, as people became accustomed to the cath labs, people became--we had improved training from experiences in the trial, we were able to modify not only the device but the user's understanding of the procedure and how they could apply that device and that procedure to obtain the results that we found in the trial. DR. POPMA: If I could just maybe help this a little bit with just a clinical perspective, as the cardiologist, of course, we're not responsible for moving the radiation source, but we actually do have a lot of experience with catheters. And some of the issues with respect to the transit time may well have related to very simple things like having the touie borst (ph) too tightly ratcheted down so that they couldn't move back and forth. We learned these things as we went through, and I can tell you the procedure that I performed today is very different than the procedure that I performed a year ago because I have much more attention to having the touie borst open, it's loose, the catheter being straight, and helping the radiation oncologist deliver the sources more quickly, as well as the fact that we stay on fluoro a lot more to make sure that there's no source drift. All of this is covered in training, and I think it's very important to emphasize that a lot of the things that you're discussing in 20 percent, none of us would want to have that shown prospectively. But they are covered in training, and I do think that we all need--we've all learned from this, and I think that we perform a very much better procedure now than we did before. We just have to, you know, show that, I think, with data. DR. SIMMONS: So even though the radiation oncologist delivers the seeds, the cardiologist should be able to see whether this thing is drifting or not, right? DR. POPMA: Absolutely. And we will do--we do that and have done that, and we're very much aware of that now, by stepping frequently on the fluoro pedal. There's nothing that prevents the-- DR. SIMMONS: Well, there is something in the manual about how many times, how often it should be observed, and-- DR. POPMA: Exactly. DR. SIMMONS: Do you know what those are? DR. POPMA: Well, right now we--you know, as I say our clinical perspective is--I'll let Drew address what the IFU is, but we do that very, very frequently now in the catheterization lab with a much more heightened awareness about the importance of drift. And by doing that, we actually can catch events. But I'll let Drew--Dr. Green discuss what's in the IFU. MR. GREEN: During the clinical procedure, the protocol described using fluoroscopy, I believe, every 10 to 15 seconds to observe to see position and location of the source train and correct--the instructions for use included in your panel pack also reflect that. One thing I would like to also point out is that we did--you know, when we went and designed a clinical trial, we didn't have some of the experience, of course, you go to gain from a clinical trial. So we applied what we believed to be conservative estimates based on bench testing, which, as you know, when you go into a clinical trial is your first available information. We prospectively defined what we believed would be some measures we would want to look at, things like source drift, source transit times, et cetera. The panel pack does now describe its modifications to the instructions for use reflecting source transit, for instance, which also goes into the train. Instead of just having recommendations based on the bench testing, we have recommendations that are based on really three things: one, the bench testing that was conducted over the expected possible pressure ranges by different users; two, experience from the START trial, what was actually reported in terms of time to send or return of source train. If you look in the panel pack there, I think you'll find that the actual--when it was reported, it was reported between 5 and 14 seconds. And then, third, we went to our oncologists and our medical physicists that participated in the trial, and we asked them what would be clinically acceptable based on what the device can do and what is necessary to occur during a clinical procedure. And we put all these things together, and we have a recommendation, which is in the panel pack, instructions for use for 15 seconds for source transit. If you don't see the sources arrive where they're expected to arrive within 15 seconds, you should then perform your manual removal procedure, which I think, again, as we pointed out just a minute ago, as time goes by, as enrollment goes up, as people become more familiar with the system, that is on the decline, indicating that the additional training and the experience they gain in the cath lab as a team--because it is a team approach--is beneficial in changing that. DR. SIMMONS: You know, I've got some questions about the training program, which is actually very much at the end of the thing. Should we--I mean, because I think this addresses part of this problem, but maybe we could put that at the end and I'll just keep going with the clinical stuff now. Okay? ACTING CHAIRPERSON TRACY: Okay. DR. SIMMONS: All right. On page 413 of the submission here, it's interesting--and I think you brought this up--that only 69 out of your 476 patients had--which is 15 percent--got 60 to 90 days of antiplatelet therapy, and only 13 of 476 patients, 2.7 percent, got anywhere near 90 days of antiplatelet therapy. And yet your recommendations in the labeling are going to be for greater than 90 days of antiplatelet therapy. I mean, I realize that a lot of cardiologists would want their patients on antiplatelet therapy for maybe other reasons. A lot of us keep them on them anyway. But, I mean, is it really necessary that we put in the labeling--I mean, you didn't do it, and you didn't have a problem. So where did the 90 days come from? MR. GREEN: As Dr. Popma presented in his presentation earlier, initially at the beginning of the trial, the initiation of the trial, we had the recommendation for physician discretion. As the trial went on, we had information from the Data Safety Monitoring Board, from the Beta-cath system trial, a de novo and restenotic lesion trial that suggested that it may be beneficial to patients to have extended antiplatelet therapy if they receive a new stent. And the recommendation was made by them and submitted to the FDA and the IEE to make that change to the protocol, minimum 90 days antiplatelet therapy for patients receiving a new stent. Therefore, at the end of the trial, we carried that recommendation over into our labeling because it did define, if you will, the experience that we attempted or we implemented to investigate in the trial, and it was a recommendation of the trial. So it's approved protocol and we carried it over. Dr. Benot (ph) is our medical director at Novoste Corporation and an interventional cardiologist at Montreal Heart Institute, and I think he'd like to add something about antiplatelet therapy. DR. BENOT: I think, as was presented either by us or by the FDA, the data that we have concern that type of antiplatelet therapy, that type of adjunct antiplatelet therapy. We have not studied anything else. Why we have the movement from one type of antiplatelet therapy was left to the discretion of the physician at the beginning, because when we start that study in September '98, we don't know, we have not the knowledge of any event related to late stent thrombosis and radiation. At this time we have already the Beta-cath trial in process, which is a different indication, is a treatment of de novo lesion with better radiation. This trial, the Beta-cath trial, has two arms: a PTCA arm, a balloon-only arm, and a stent arm. In the stent arm, we finally find out by the end of October '98--the DSM Committee chaired by Tom Ryan come to us and report some of the complications related to the late stent thrombosis. At this time we started to apply longer adjunct antiplatelet therapy, and at this time we propose two months. By March '99 we were secure, we are sure that the problem was still there, and at this time was implemented the minimum of three months of adjunct antiplatelet therapy. That's the detail we have from the Beta-cath trial. We have never had a problem as reported in the START trial for that as medical officials, and we discussed that, we translate the data we learn from de novo stent and radiation to in-stent restenosis and radiation and apply to ask our investigator to prescribe a minimum of three months when they implant a new stent. DR. SIMMONS: It's just it's interesting that you didn't have any problems in this study and only 2.7 percent of your patients had, you know, anything close to ninety--but I guess there's no harm in putting-- DR. BENOT: As the data that we have--and I can have the report from the statistician from the Beta-cath trial, Stuart Pocock, we put on the letter from Dr. Stuart Pocock, and on the Beta-cath system trial, again, different education, de novo lesion, but using a stent as the arm differentiating with the balloon-only. And based on the analysis and listing of the current interim data, the incidence rate of late stent thrombosis, Q wave and non-Q wave MI are all satisfactorily much lower in the patient first randomized and treated in the provisional stent branch following the protocol amendment, which is a minimum of 90 days of adjunct antiplatelet therapy (?) . These findings are based on 492 patients randomized in the provisional stent branch before the protocol amendment with a median(?) follow-up of 18 months, and, further, 449 patients randomized in the provisional stent branch after the protocol amendment, which is the minimum 3 months of antiplatelet therapy, and that with a median follow-up of 9 months. That's the detail we have. We have no other data than that to explain why the level of the protocol of adjunct antiplatelet therapy. DR. POPMA: I appreciate your letting my colleagues address the background behind that, but as the principal investigator of the trial, I'm comfortable with the statement of at least 90 days, because we do feel that we only had 50 patients in the study who received new stents and radiation therapy, approximately. So to make a definitive statement that there were no subacute stent thrombosis within the first 242 days, there may be some broad confidence intervals to that statement. In addition, we do have this patient at 244 days that had an episode that could be very consistent with a subacute thrombosis event. So I think I am comfortable for those reasons in saying a minimum of 90 days of antiplatelet therapy. DR. SIMMONS: Now that you are up there-- [Laughter.] DR. SIMMONS: So the new stents were discouraged. DR. POPMA: That's correct. DR. SIMMONS: But 20-some percent got new stents. DR. POPMA: I think that's a good point. Let me just again walk through very simplistically exactly how you got into the study and how you got a new stent. First of all, in order to be randomized in the trial, one had to have a successful result. You had to have a 30 percent or lower residual stenosis. So the concept was that in order for you to make the decision for randomization, you didn't want to have a new stent in place. At least we didn't have a new stent in place. So the radiation was then delivered. And time passes during that period of time, and we know some things about stent restenosis. One of the things we've learned from work done at the Washington Hospital Center, Roxanna Mayron and Gary Mintz, is that there is an early recoil that sometimes occurs. The mechanism of treatment for in-stent restenosis, you extrude the tissue outside the stent struts, and then within the first 30 minutes or 40 minutes, there's actually a collapse of that and the tissue comes back within the stent. And there's time that passes as we're delivering the radiation effect. So it's understandable that in some patients the residual stenosis would look a little higher after a time delay than it would be if you just ended the procedure and we called it a successful result. So the new stents that went in went in for two reasons: one was there was a dissection that had to be treated; or, secondly, because there was a residual stenosis that was within the lumen. DR. SIMMONS: Well, wasn't a dissection a contraindication to giving the radiation therapy in the protocol? DR. POPMA: Yes, that's correct. In order to get into the procedure, one had to have a successful result, which was the absence of dissections. But then, as I say, there is a dynamic change that can occur within the lumen. Sometimes the recoil that occurs from re-extrusion of stent plaque into the vessel wall looks like a dissection. It's a flap that can fall back in. Angiographically it's somewhat difficult to tell those. But what we did is we really lowered what was a prevailing rate of 80 percent new stent use for in-stent restenosis down to 20 percent. And if the truth be known, what we know about the START data now, we'd like to get that even lower. And so some of the things that we would really like to say with this study is that we really want to reserve the use of new stents in the study for bail-out circumstances, some circumstance that happens during the procedure that, after radiation therapy, one has to treat with a stent. That can be a new lesion at a new site or within the site that you're treating initially. DR. SIMMONS: So you would say that these mostly were done--the new stents were put in after the radiation was already given? DR. POPMA: That's correct. DR. SIMMONS: Or where the placebo was given. DR. POPMA: They all were, yes. DR. SIMMONS: And so did--I mean, I guess it could significantly affect the results. Did you look and see were they equally divided on both sides for stents, for the placebo versus the active-- DR. POPMA: There was no deleterious effect of radiation on causing more stent in group than another group. DR. SIMMONS: But did one group have more stents than the other group? DR. POPMA: No. They were equally balanced between the two. DR. SIMMONS: How about the breakdown on the diabetic patients? I was especially curious. Did you look at that as far as-- DR. POPMA: Diabetic subset? I'm going to let my colleague, Dr. Kuntz, address the subset for diabetes. DR. KUNTZ: Diabetics were evenly distributed, and there was no effect of diabetes on the instance of restenosis in the trial. DR. SIMMONS: Okay. So there was no beneficial effect either. DR. KUNTZ: There was no differential effect. Both groups benefited. It just was no differential effect. That is, the interaction between diabetes and radiation therapy was not positive. DR. SIMMONS: Okay. You know, I guess I'm just going to ask the radiation oncologist--this is exposing my naivete here, I guess, but I just have to ask this. I guess I'm a pessimist by nature and a therapeutic nihilist to a certain extent, and I guess I just don't believe that you can put radiation inside a coronary artery and have a beneficial effect without also having a risk of a negative effect. I mean, there's got to be some downside to this as far as creating aneurysms or scar carcinomas or something. There has got to be a downside. I guess I just want your opinion. You know, what's the downside and how long do we have to wait until we see it? DR. SPEISER: Probably the easiest to dispel is the incidence of cancer from this as an overall problem. About one-tenth of a percent of the total dose is delivered from the Strontium, so that by itself is very insignificant compared to the fluoroscopic dose, which would be a greater concern. The only concern I would have is that it's a very high point dose, so it would be more likely to be a concern as far as increasing fibrosis or aneurysmal formation. At the present time, the data that we do have is the BERT trial that shows neither of those two effects have increased. So the answer is that with the available data, which is very scant, there has not been a late deleterious effect. However, I think most radiation oncologists would agree, because radiation effects are delayed, that we would like to continue looking for it for a longer period of time. DR. SIMMONS: What are talking here? I know some radiation effects, like for lymphomas, can even occur ten years later. Are we looking at something that may all of a sudden show up five or ten years from now with severe scarring in that area? DR. SPEISER: Most late effects, such as scarring or vascular effects, usually occur between 6 and 24 months after completion of radiation. So that I would expect that most of them will show up in that time period. The later effects, as you mentioned, for instance, carcinogenesis, is delayed. Lymphomas are the earliest cancers, about 10 to 20 years. Sarcoma is 20 to 30. So that for those we'd have to wait a much longer time. However, I'm not, as a radiation oncologist, concerned about the carcinogenesis, but just the direct immediate effects of the high dose on the vessel wall, and that I anticipate that we should see for the most part between 6 and 24 months from the completion of the procedure. DR. SIMMONS: Okay. Maybe our radiation people will have something more to comment on that. Just one more and then I'll-- MR. GREEN: Perhaps, if you'd like, we could have maybe one more opinion from one of our radiation oncologist? DR. SIMMONS: Maybe not. Maybe we'll let our radiation oncologist ask some more questions on that issue since I'm not... As far as your warnings and your contraindications section on your labeling, this study actually did eliminate people with ejection fractions less than 30 percent and it did eliminate people with myocardial infarctions within the last 72 hours. Shouldn't those be--I mean, since those patients weren't studied, I just want to know shouldn't we put some warning or at least some contraindication. I'd like to have your opinion before we discuss it when you're not available to comment. MR. GREEN: When we developed the protocol and we put it together to study this trial, as you do in many trials, there are a lot of things that you put in the trial to try to either limit bias or try to determine what the effect on the patient was. For instance, the one recommendation in the protocol you talked about was myocardial infarction within 72 hours, and that was to be able to delineate the baseline factors for a patient. We are recommending that the patients--the instructions for use reflect what we did in the clinical trial. However, like I said, they were recommendations in the--or exclusion criteria in the protocol that were specifically limited to the ability to evaluate the patients in the follow-up to see if the therapy was effective. DR. SIMMONS: I think that's fine, but for right now what I'd have to say is at least when go to discuss this later on, I would have to say those are things that would have to be added, at least a warning if not a contraindication, if they aren't there now. DR. POPMA: Respectfully, I'm not a labeling expert, but as a clinician, I don't see that there's any reason to suspect that in a patient who has recurrent refractory in-stent restenosis and ejection fraction plus the 30 percent that this therapy should be contraindicated. And I'd only hope that the trial design construct could be described in the labeling, and then a very careful construction of what was done in the trial and the inclusion sets. But I think at this point in time, I would say from a clinical perspective that I wouldn't know that there'd be data suggesting it should be contraindicated in a patient with an ejection fraction plus the 30 percent. DR. SIMMONS: I've got some other issues on the training session, but maybe we can put those off until everybody else has had a chance. ACTING CHAIRPERSON TRACY: Okay. I'd like to ask our statistician if he has any particular questions, and then we'll go--we'll probably break for lunch before we go around the rest, but if we could have Dr. Bailey ask any questions. DR. BAILEY: Why don't I just list a few questions? Because I tend to get confused when I hear the answers to them. So, in no particular order, well, first of all, I'd like to echo I thought this was a nice study and well reported. With respect to the analysis of these minor--what are they called?--MDMs. That's all I can remember. Device malfunctions. I notice that there was an analysis of the clinical impact, and this probably will just reveal my ignorance. I noted that you pooled the drift and the long transit time. And I was wondering if that's based on a priori considerations that those would have the same impact or just--I would have thought, I guess naively, that drift would be a relatively more important issue and perhaps should be analyzed separately. And in that same vein, if you're trying to understand the impact, it would be useful to look at the edge effect with respect to those cases that had drift. I thought that was a nice analysis of the edge effect where you had lots of power for quantitative analysis. That's sort of one set of questions. The second one, actually sort of related: Did you actually look at the cases with these malfunctions to see if there were any patient differences? In hearing the discussion, it sounds like this is more or less a random occurrence, but I didn't know whether you looked at whether there were patient differences in those that had the drift problem. Relative to the recommendation of length of antiplatelet therapy, do these results in the START trial, are they consistent with the earlier--the Beta-cath results? In other words--I'm sort of following up on your comment. If you looked at the results, there's a certain number of patients that did not get what would have been considered the desirable length of antiplatelet therapy and, nevertheless, no events, no thrombosis occurred. And I wondered if it's just too small a sample or if those results could be compared to the earlier results and see if there's anything different about these data. That's question two. The third question has to do with the--I think I saw among the analyses that were in the packet that wasn't presented some modeling done of the effect of lesion length and treatment and an interaction term, which I thought was very interesting and showed that the treatment effect was more pronounced at longer lesion lengths. And I guess I would just ask if this has something to say about the risk/benefit ratio in terms of the labeling aspects, and I thought that analysis should be made more accessible to the user to determine if there's a lesion length that's less than optimal. And then my last question has to do with the heterogeneity between sites, and I saw some analyses in the packet, but I didn't really understand what was being done. And in particular, were analyses done to suggest that sites had different overall restenosis rates, or was there also differences in efficacy rates? So I guess it was more just ignorance that I didn't know what was being presented. ACTING CHAIRPERSON TRACY: I'm not sure how you want to approach that. Maybe one question at a time? You can identify which aspect you're dealing with and also identify yourselves. DR. KUNTZ: Rick Kuntz. I'm a cardiologist and a part-time statistician, I guess, although I'm a little intimidated by Dr. Bailey. Why don't we just start from the top there? DR. BAILEY: Drift and--pooling drift and transit time. DR. KUNTZ: Right. We had a variety of different MDMs. There were four. There was the drift, transit time issues, inability to deliver the catheter, and some other issues that were like one person per category. They represented a variety of different fields that were prospectively collected in the case report form, all classified as potential device malfunctions. In looking at that overall data set, in order to reduce multiplicity and try to deal with, you know, diminishing the alpha to zero, we focused just on the radiation issues. So the decision to pool transit time and drift was an issue of power and reduction of multiplicity. So we haven't looked at the individual events themselves because they were evenly distributed. I think there were something like 80 cases overall between the two groups that were MDMs, and there were about 40 and 40 on each one, drift versus that. So my expectation is that since we generally found no difference in the adverse events with the pooled group that we probably would be completely underpowered to look at the individual groups themselves, and it probably wasn't worth the analysis. DR. BAILEY: What I was thinking, though, is if you looked at the edge effect in a very quantitative way, looking at the delta minimum luminal diameter in that region and separated specifically the drift ones, you might have some power to look at it. DR. KUNTZ: Right. Your second part of the question about did we look at drift effect with the edge analysis specifically, we have not done that, and I agree that would be an interesting analysis because we may have enough power there because they're both continuous measures. And I think that would be an interesting analysis to do. DR. BAILEY: Patient differences in terms of predicting who--was that just a random event? DR. KUNTZ: Yeah, we spent--the question was could we look at--were there any anatomical patient factors that explained patients who had these MDMs. So we spent a lot of time looking at those factors, and we couldn't find, that is, by generally exploratory analysis, that there were any indicators of increased tortuosity, that there were indicators of the distribution of the vessels. For example, was a right coronary artery more likely to drift than a left coronary artery? The amount of calcium that was in the vessel, the age of the patient, I think a variety of different things. I'm thinking off the top of my head. We tried to evaluate whether we could predict who was going to have a drift of source, and we couldn't find them. Practically speaking, the issues of drift to start with and the intransitude to some degree were really issues of the radiation oncologist, an issue about the touie--the hemostatic device more than they were issues of patient factors; that is, the device itself is designed not to have any kinks and is bulky enough that it generally won't be delivered down very, very tortuous vessels to allow the catheter itself to impede delivery. So delivery impedance were issues of maintaining pressure and issues of the hemostatic valve. So we think that those were the things that explained the differences, not issues of patient characteristics, where we can say that this patient is at more risk of a drift than the other patient on initial exploratory analysis. We couldn't identify patient factors. DR. BAILEY: Can you compare the thrombosis rates between patients who in the START trial had new stents placed and did not get 90 days of therapy to the earlier data that were the basis for-- DR. KUNTZ: In the Beta-cath trial, we're looking at a 1,500-patient trial compared to a 476-patient trial, the START trial. So the history which was reviewed is important to understand. In that trial, patients were treated initially with balloon angioplasty. Then depending on the result, the physician decided whether they would go down a PTCA branch based on a very, very good result for which the patient was randomized to placebo versus active therapy blinded with no further stent placement, or if the result from balloon angioplasty was suboptimal, they were arbitrarily decided to go down a stent branch and then randomize after that. So we had a fairly large volume of patients initially, as you can imagine, because of the stent interest at that time of patients who had new stents placed, on the order of five or six hundred patients, as opposed to 50 patients with new stents in this study. So the opportunity to observe stent thrombosis was greater in the Beta-cath trial than the opportunity to observe stent thrombosis in this trial. So, initially and early on, when the Data Safety Monitoring Committee with its blinded review of the data identified that there were some problems going on when new stents were placed and patients exposed to radiation therapy and expected--and declared that they wanted to extend antiplatelet therapy--this, by the way, was reviewed with the FDA and the protocol was changed. We anticipated that this also might be an issue in the START trial where new stents were placed. However, at the end of the START trial, the 476-patient trial, only 50 patients received a new stent. So we are extrapolating the potential for stent thrombosis, even though we had excellent results in this study, to our experience with over 500 patients early on where there was a higher incidence of stent thrombosis; hence, the interest in potentially having 90 days or more of antiplatelet therapy. DR. BAILEY: The interaction between lesion length and treatment effect. DR. KUNTZ: It's very interesting interaction. We mainly saw it in the restenosis defined by the analysis segment, not by the stent segment. And what we see is that, in general, lesion lengths are associated with a higher risk of restenosis. That's been true with multiple data sets, especially in non-radiation areas. That is, patients who have longer lesions tend to have a higher risk of restenosis than patients with shorter lesions. When we look at the analysis segment, which actually lets us have the opportunity of measuring the minimum lumen over a wide area, we start to see that radiation therapy had an extra effect on patients with longer lesions, and that made sense; that basically the increased risk the patient was exposed to with a longer lesion afforded a more profound treatment effect from radiation therapy than those who had shorter lesions. So the interaction term of longer lesion lengths and radiation therapy made sense to us, understanding the underlying risk the patient had with longer lesions. DR. BAILEY: In fact, based on the coefficients, if you have a lesion length of 8 millimeters, you're at dead even. DR. KUNTZ: Well, right. It's hard to go back and say where the breakpoint is of radiation therapy being ineffective at some level. All we can say is that the continuum shows that longer lesions have more potential for effect than shorter lesions. But these lesions have been linearized in a linear model. We didn't do a lot of non-linear models to see where the breakpoint is. And so I think that it's an interesting extrapolation, maybe the basis of a hypothesis for a new study. DR. BAILEY: I agree it's not very exact, but I think it points to at least an issue if you're a user whether you want to embark on radiation therapy in a shorter lesion. DR. KUNTZ: That's a good point. DR. BAILEY: And, finally, the site heterogeneity. DR. KUNTZ: Right. The site heterogeneity we thought was typical in most of the studies, that is, the overall distribution of treatment effects for a 50-patient trial showed--a majority of patients showed a similar result as the mean effect overall. A couple sites out of the 50 had the opposite results, which you typically see in a normally distributed trial. The other heterogeneity issue dealt with--there was one or two sites that tended to use new stents more often than others per se. However, in the overall analysis, the restenosis rates didn't differ, so we didn't see a profound effect on the site. We performed the typical boilerplate pooling analysis for the FDA looking for interactions between treatment site and the overall main effect, and albeit that's another powered analysis usually, we didn't see any deviation from the normal studies that we saw. DR. BAILEY: Thank you. ACTING CHAIRPERSON TRACY: Okay. At this point I think we'll break for lunch, and if we could resume at 1:15. And I would like to remind the panel members not to discuss the contents of this meeting. [Luncheon recess.] AFTERNOON SESSION [1:30 p.m.] ACTING CHAIRPERSON TRACY: We'll resume the open committee discussion, and we'll resume with the panel questions, and I think we'll start down at that end, please. Again, just to remind everybody to identify yourself and to speak into the microphones. DR. AYERS: Okay. I moved up a little in the order, but I have a couple of questions. One, your presentation indicated that you were giving 18.4 Gray and 23 Gray, depending on the vessel size, but actually since this is non-centered sources, an asymmetric lumen, shadowing by guide wire and stent, what really was the dose range for these studies, minimum, maximum, your estimates to the (?) . MR. GREEN: I think we'll let Dr. Crocker, radiation oncologist and investigator in some of the BERT feasibility studies and who helped in the START trials, answer this question. DR. CROCKER: My name is Ian Crocker. I'm a radiation oncologist at Emory University. I'm a consultant and a shareholder in Novoste, and, in addition, Novoste has licensed intellectual property from Emory University, and I am co-owner of that intellectual property. Within the vessel wall, there is a wide range of doses that are delivered, and that's really true of any brachytherapy source. We did prescribe a dose at 2 millimeters from the center of the source, and that initial prescription represented a small incremental increase in dose over what was prescribed in the Beta Energy Restenosis Trial based on anticipated shadowing of the source by the stent struts. We had done some measurements which had shown that there was approximately a 10-percent decrement in dose immediately underneath the stent struts, and as a result of that, we recommended increasing the dose that was delivered in the START trial by 2 Gray, which represented an 11- to 14-percent increase in dose over what was delivered in the original BERT trial. DR. AYERS: Do you know how much effect the dose varied from the fact you had a 6-millimeter variation in vessel size for the same dose, so I guess that would be 0 to 3 millimeters in variation from the vessel wall to the prescription point, and also the fact that it wasn't centered. DR. CROCKER: Right. With the cohort of patients who were treated in this trial, we undertook a retrospective analysis of dosing using intravascular ultrasound images, and information on that has been submitted to the FDA as part of this submission. Basically, the catheter assumes a relatively centered position within the lumen based on these IVUS ultrasound images, and really there are only minor differences in the doses, you know, that are received to the vessel wall with active centering of the catheter within the lumen compared to non-centering of the catheter. DR. AYERS: One other one, I guess just for clarification. It wasn't clear to me. When you added new stents to about--what, 20 percent of the patient population, as I recall. DR. CROCKER: Correct. DR. AYERS: Was that done before or after the radiation therapy or a mixture? DR. CROCKER: Those new stents were added after the radiation therapy was delivered, so that the protocol specifically excluded patients with stent sandwiches or stent within a stent, so that we didn't anticipate that there would be any areas in which there would be stent overlap and more than approximately a 10-percent decrement in dose due to the shadowing effect. You know, I should say that that decrement in dose becomes less important as you get further away from the stent. In other words, there's a relative filling-in of dose at increasing depths beneath the stent. DR. AYERS: Okay. One other thing I noticed, you spent a lot of time on indicating how small the dose to the patient was from--or incremental dose to the patients from the beta therapy, particularly whole-body, which is certainly true. But nowhere in there I saw addressed is how much increase in the dose was due to the fairly substantial additional fluoroscopy in sign, particularly, you know, monitoring the source position every 30 seconds. Do you have any value for the added skin dose for that additional fluoro? DR. CROCKER: I'm not sure that I have any additional information regarding the fluoroscopic dose. Maybe Dr. Popma might want to comment on this. DR. POPMA: These are very short pops of fluoroscopy and not a long length. You really just have a second or less, just to check the position of it, which you can review on your video replay. DR. AYERS: Okay. But every 15 seconds, that would be, what, over a 2-minute treatment time? DR. POPMA: Again, just a second or so each time. DR. AYERS: And I was curious--and the last item I have for right now. I guess I forgot--I'm Robert Ayers, NRC. I didn't identify myself starting this. Novoste introduced later in the study, when problems were uncovered and we investigated some of these and, in fact, generated an information notice on source transport difficulties, particularly through the introducer or touie borst valve or whatever was used for that. It's known that if that's overtightened, it can block the sources going either in or out, and I think well less recognized perhaps by the panel, if you overtighten it too far and go past the elastic limits of the catheter, that blockage stays there even if you loosen the valve. And they introduced this introducer sheath as a corrective measure for that but don't require it, and I wonder how come. Our experience has been that the cardiologists don't use it because it's an extra step, in the one incident we looked at. DR. POPMA: Drew? MR. GREEN: What we found when we--first, of course, you're correct. We did qualify an arrow sheath introducer as an additional accessory that the clinician could use in the procedure. They would place the catheter through the introducer sheath, which is its labeled indication for use for introduction of percutaneous catheters. And, therefore, when they would tighten the touie borst, the hemostasis valve down onto that sheath, it would protect the catheter. What was found in talking with the centers and looking at what was happening, especially the center that you were talking about, was that this had, if you will, a learning--it was part of a learning curve, you know, part of the learning of using the device, and that the interventional cardiologist places the catheter and, you know, it's their job every day to maintain placement of that catheter, angioplasty catheter, guide catheter, what have you. And so that's their job. So now they have a new player, a new team member who's also involved with the interaction of that catheter with the transfer device. So it's part of the training, you know, of the team working together, about moving and then learning when to tighten and how much to tighten on the catheter to allow for passage of the source and, you know, to compensate for another person being in the team. So they wanted the ability to have this as a tool if it was necessary to use in their practice, and if not, or if they felt that they were at--or had demonstrated they were at the learning curve to where they didn't need this tool, they didn't have to use it. And this becomes very important because in the training section of the panel pack, we actually talk about, you know, going through experiences from the trial or all the trials and evaluating what the proper--you know, how a user would use the system. And as part of the hands-on training and the mock training and these pieces of the training, it's important for the clinicians to determine, you know, of the optional accessories such as the arrow sheath and the fluid management system, how they would apply that in their practice; in other words, what works best for them so that they can use this system the most effectively to gain the results that were seen in the START trial. DR. SPEISER: At our institution, the radiation oncologist uses the arrow flex sheath, will flush it and place it over the delivery catheter so that there is no time delay for the cardiologist. And it is my intent in the training program to train the radiation oncologist to do this and to use it all the time, unless the cardiologist specifically says they do not want to use it. DR. AYERS: Well, we're strongly considering making that a mandatory requirement. That's why I wanted to ask the question, at least at our regulatory agency. And one last one was with our upcoming change to our medical regulations--and I'm assuming--I'm not saying that that's, in fact, the way it will work out, but most cases for brachytherapy and particularly high dose rate, we have a mandatory requirement coming now that the licensee, user, medical physicist, you know, the medical institution, however you want to characterize it, is solely responsible for the calibration of the brachytherapy source dose rate. And going over your submission, particularly in Section 2, it is not clear to me that you provide the tools with your system to allow the medical physicist at your customer site to perform proper dose rate calibrations on these sources. MR. GREEN: We understand that some sites do at this time have requirements at a site level, possibly there will be other requirements later for site verification of dose rate, et cetera. And we do have proposals on how to handle that. And I think that Dr. John Lobdel, our Director of Radiation Management, can speak to that and let you--you know, what the proposals we are planning to do at the sites to be able to address that are. DR. LOBDEL: John Lobdel, employee of Novoste. We have a source train that was calibrated at NIST to determine the dose rate at a half millimeter--I'm sorry, at 2 millimeters inside the--in water--I'm sorry. Let me go back. We have a source train that was calibrated by NIST to measure the dose rate at 2 millimeters from the center line of the source train in water. This train is our transfer standard. This train is used to calibrate all the trains we send to the clinical sites. Now, during the clinical trials, we had two sites that asked to verify our dose rate. We worked with them on this. The tool we have is a solid water block that positions the center line of the source train at 2 millimeters from a film plane. We went to the site, irradiated the trains for the hospitals, then from this film they analyzed the film, determined the dose and dose rate, and also the homogeneity of the train. We actually published a paper on the results. The hospitals were very happy with the results, and it came out in the literature about a year ago. Now, we're also looking into a different method, and that different method is there's another source train in this being calibrated for dose rate and activity. That train will be sent to at least one and probably two accredited dosimetry calibration laboratories, or ADCLs. The ADCL will in turn calibrate their equipment on this train. Then when a hospital wants to know what the dose rate and activity of our trains are, they can simply send their well chamber to the ADCL. It will be calibrated there and returned. And then as often as they wish, they can simply take a source train and put it into the well train to determine activity and dose rate. So we have here one system that has been proven and has been published in the literature. We have another one that we are working on that should be available quite soon. DR. AYERS: The latter was what I was really looking for since our anticipated regulations require our licensees to go to an ADCL or NIST for this type of calibration that you just concluded with. So it sounds good. DR. LOBDEL: Thank you. DR. AYERS: That's all I have. ACTING CHAIRPERSON TRACY: That's it? Okay. Dr. Crittenden, any questions? DR. CRITTENDEN: Yes, I have several. The first question I'm going to direct to Dr. Speiser and Mr. Green. What was the impetus behind the device change--the impetus behind the change in the design for the device going from the Alpha III to Alpha IV? Was it to minimize the problems with source delivery? If so, was there a comparison made between these two devices to see if there was a difference. And then, finally, for this first question, Dr. Speiser stated that the radiation oncologist might feel uncomfortable in the cath lab given that this is a new setting for them and that this may have been a source for some of the source drift or source transit time problems that we saw with the devices. Is it your position, because the analysis showed that there may be no difference when you look at placebo versus the Strontium-treated groups in terms of outcome, whether you have an MDM or not, is it your position that there are no untoward sequelae for source drift or source transit time? MR. GREEN: I'll go first, kind of go in order of the questions. The first question I understood to be, you know, what was the reasoning behind going from the Alpha III transfer device to the Alpha IV device. Basically, we started the clinical trial with the Alpha III transfer device. We were having that transfer device manufactured or built by a subcontractor. We moved to another subcontractor, qualified that subcontractor, and part of the qualification of that subcontractor went through for us was to do an evaluation of the Alpha III transfer device and to propose some improvements in the device that may make it more user-friendly, the user interface a little easier to use. And there were several minor things besides the LED. For instance, the shape of the housing was changed a little better to fit the hand. Some of the graphics were made a little clearer, and the LEDs were added. And the LEDs were added because the subcontractor here determined that they believed that that gave a more accurate feedback than did the mechanism of the Alpha III, which was simply an open window that showed to where you could visualize a spring as part of the pressure relief valve. So they believed this would be a more accurate and calibratable method of providing feedback to the user. And that was the reason we went to that change. So when we implemented that change, one of the things we learned was how people used it, and that's kind of where we've gotten with the training now, taking those experiences with that and how can we make that what is actually a very well calibrated and accurate gauge and make it something that the user can use most beneficially in the treatment here. The other piece of that would be, just to segue into the question you asked Dr. Speiser, would be with regard to the source transit time. I think--this is the second time we've covered this. I think the first time I forgot to mention that when we prospectively designed the trial to try to figure out--you know, one of the objectives of the trial was to determine device--you know, what the device could do and how it would be used, and that was built into the protocol, and you can see that in the protocol as one of the objectives. So we developed on the case report forms multiple pages to collect some of these pieces of information. The best information we had without clinical data was that on the bench source transit could be accomplished in five seconds. So we put one of the items on the case report form as being, you know, was source transit greater than five seconds. And in most cases it was a check box with very little information to be added or able to be added. So we incorporated that. And as I said, now having had the clinical experience, having had the feedback in cases where it was reported how long it actually took send or deliver, given the time to evaluate the clinical situation, and having the feedback from actually specifically asking the oncologists and physicists in the trial what is clinically acceptable with regard to how long is this taking, given the anatomy where this would be located, the fact it's in a guide catheter, et cetera, and that's where we came with the source transit recommendation of 15 seconds, which does kind of put things a little more in perspective when you look at the minor device malfunctions in that about fifty--I believe 54 was the number of my device malfunctions, with source transit greater than 5 seconds. You know, if you apply what would be the clinical reality and the practicality in the cath lab of doing this procedure to that of 15 seconds, you actually see a great reduction in the number of those reports. Dr. Speiser? DR. SPEISER: I'd like to reiterate that the 5 seconds was chosen before the clinicians were involved, and it was based on an engineering bench testing. What we've done with the radiation oncologists and physicists in the study is look at different parameters and decide that 15 seconds would be achievable in the real world clinically, and that it was a safe dose for transit. So that the transit itself being more than 5 seconds I do not see as causing a problem to the patient. The drift was a bigger problem, because if the drift truly means that the source is outside of the target area, then we'll have a diminished effect. We've done some bench testing recently to show that the gold marker, which is very dense, moves a long time before any of the sources move. So that while I can't say whether the sources moved or not because they're very difficult to see on fluoroscopy, we still have to assume that if the gold moves, that eventually the sources will move, and if the sources move, we'd have a decreased benefit to the patient. However, I don't really see with the very slight movement an increased deleterious effect for those treatment times, just the opposite, a decrease of the beneficial effect. DR. CRITTENDEN: Now, when I went over the protocol, I didn't realize this--and this is my fault--that the radiation oncologist actually injects the source train after the cardiologist places the delivery catheter at the appropriate spot? DR. SPEISER: That's correct. DR. CRITTENDEN: I realize that the radiation oncologist brings a lot of expertise to this that is absolutely necessary in the cath lab. But it seems to me, just on face value, if I want someone injecting something in my coronary artery. I'd prefer a cardiologist to do it. Why do we need--and I'm not trying to be funny. I just--why do we need someone who, by your admission, hasn't been the cath lab in years now injecting a catheter, which the interventional cardiologist does--I mean, they can do that not literally--but you know what I'm saying. DR. SPEISER: Yes. The catheter is placed by the interventionalist, not the radiation oncologist. The radiation oncologist sends the sources within a closed system. So there's no effect on the coronary system. The primary role of the radiation oncologist is not mechanically to do it, but for the other attributes of making sure the device is prepared correctly, the dose as well as any problems that might come up. So that right now at the present time it's an NRC-mandated rule that this amount of radiation be handled by somebody who is licensed through NRC or through an agreement state to handle the radiation. DR. CRITTENDEN: And that includes--because if I understand that correctly, you're squeezing the syringe and you're pushing the source train-- DR. SPEISER: That's correct. DR. CRITTENDEN: And then aspirating it. But the NRC-- DR. SPEISER: Well, it's not really--it's a hydraulic system, but it works that way where I'm putting pressure on the syringe to move hydraulically the sources up to position, keep that in position, and then with the change of the valve, the same syringe will return my sources back into the device. MR. GREEN: I think the key--one of the key--the two things here--there's actually two points to be made. One is that there's a regulatory requirement. Within the facility, the radiation oncologist is the authorized user or the person licensed to actually not only handle but actually treat with or apply the treatment of radiation, the second piece being the cardiologist is actually controlling the delivery catheter, if you will, the patient contact portion for the coronary anatomy and that the oncologist is delivering the therapy, as they are licensed to do. It's also important to point out that the team approach that comes from the oncologist, the cardiologist, and the physicist I think puts the right emphasis on both the dynamic cardiac responsibilities and care that need to be taken for the patient as well as the radiation safety, and then the radiation protection type issues that the medical physicist and the oncologist bring from their arena I think is very important. DR. POPMA: To put this in perspective, usually in a case, as you've probably seen in the cath lab, there's a first operator position and a second operator position. The cardiologist is always in the first operator position. And my responsibility as part of the case is to make sure the catheter position is, in fact, appropriate through the whole period of time. The second operator position is actually the radiation oncologist who's injecting the source train. So it's not that--the catheter position is clearly our responsibility. MR. GREEN: Just one subtle note. When you send the source train, you apply a positive pressure on the syringe. In order to return the source train, you still apply a positive pressure. There's a fluid control valve. You simply change direction and apply a positive pressure again and it comes back. So it's continuous. DR. CRITTENDEN: The next question is for Dr. Popma. In this trial, there were a few cases of late thrombosis, at least out to 8 months, but historically there has been a question of whether brachytherapy leads to late thrombosis. There is data that has come from the Washington Hospital Center and it was just kind of a meta analysis of some of their trials that suggested that late thrombosis was related to new stent placement. So, with that preamble, why do you think the rate was so low in this study? Is this from a reduced radiation dose? In the trials before, they had higher radiation doses and different sources; some of it was gamma radiation as well. So does this represent a difference between beta and gamma radiation? Or is it due to this adjunctive platelet therapy? Then, finally, is the sponsor going to make a claim in this regard vis-a-vis late thrombosis? DR. POPMA: I am academically aware of the data from the Washington Hospital Center suggesting that there are cases of late subacute stent thrombosis that have been associated with other forms of therapy. It's very difficult to make comparisons, and I don't feel comfortable making a comparison of gamma and beta about incidence of late subacute stent thrombosis rate because in this trial we didn't use very many stents. And I think that the reason that this was such a safe trial, which it was, and we didn't really observe much in the way of stent thrombosis, is that we only used stents in 100 patients, and only approximately 50 of those got radiation therapy. Having said that, we don't think it's a very frequent event. We have this one case that we have been, you know, in due diligence, I think, reporting, but that was outside the 8-month time frame. Other than that, we saw no occurrences within there. Personally, I was a little surprised about how many patients did not receive extended antiplatelet therapy. It was my impression that the clinical investigators, knowing the data that had been coming out, would have put patients on their own on extended antiplatelet therapy, but in this trial they did not. And you saw the vast majority of the patients had relative short antiplatelet therapy durations. That just comes back to one thing, and I think that that is we didn't use stents very often. And we used them specifically for bail-out indications. It wasn't always perfect. Sometimes they were cosmetic, but the majority of time they were for bail-out indications. And I think for us as operators that is a very important lesson. When you're treating in-stent restenosis, it's important that you try to get the best you can mechanically without adding a new stent. It means sometimes debulking devices, sometimes balloon angioplasty, being patient with it, but then avoiding using a new stent if possible. I think following those rules we should anticipate that we're going to have continually lower subacute stent thrombosis rates. I do think it's premature to make any comparisons between one isotope versus another isotope. DR. CRITTENDEN: The next question is for Dr. Speiser. I realize that emergency coronary bypass was rare, but what recommendations would you make for surgeons or other OR personnel if we had to do an emergent case on someone who just had this therapy? Not worry about it? Should we worry about it? I know it's an extreme example, but it's been talked about kind of at the water cooler at conventions. If you had to remove a stent, how should it be disposed of? DR. SPEISER: The stents in this case are not radioactive, so the minute that I withdraw the device, there is no radiation in the patient. So to answer the question, you do not have to take any precautions vis-a-vis the radiation. DR. CRITTENDEN: That's all I have. DR. IBBOTT: Thank you. I'm Geoff Ibbott. I'm a medical physicist, and I have a couple of questions about radiation safety issues and dosimetry. My first question is: In the description of the device, the device is described as having a quartz chamber where the sources are housed which attenuates the beta radiation from the sources. But in the labeling there are warnings about avoiding holding the device with your hand over the quartz chamber because of the risk of exposure. And so I'm wondering what the exposure rate or dose rate at the surface of that quartz chamber is and what the exposure might be to an individual. MR. GREEN: You are correct that there is a quartz in the device that does provide the shielding for the beta radiation from the source train. There's a slight rem strong component. There is a stainless steel capsule to the source. I think Dr. Lobdel, if he could come back, I think he could answer the question, just provide to you what the numbers in the instructions for use mean and what the component that you're seeing and he's talking about there is. DR. LOBDEL: Again, John Lobdel, employee of Novoste. We don't know what the dose rate is at the surface of the quartz. That has never been measured. The quartz is sufficient to stop all of the beta. Obviously, there's a rem strong output. We have not quantitated that. We do know what the dose rate is on various surfaces of the device. Specifically over the lens you could see roughly 100 mRAD per hour at contact with the lens. Again, that's due to the rem. But there is no reason for the operator to open up the device. It is sealed and our instructions clearly say don't open up the device. DR. IBBOTT: Perhaps I misunderstood the description of the device. So the lens that you mentioned is the surface that the operator would be in contact with. DR. LOBDEL: Right. MR. GREEN: Just for orientation, yes. The transfer device--the quartz is actually enclosed inside the device, and when you look through the device--if you refer to your picture, we can find it in the panel pack. When you look through a lens, a magnifying lens so you can see in the device, the quartz chamber is actually below that lens internal to the device. So measurements that are in the instructions for use would be described at different surfaces on the outside of the transfer device, the part that you could actually come in contact with. As Dr. Lobdel said, the quartz--the viewing lens is on top. Of course, we (?) device on the bottom. DR. IBBOTT: Yes, I see. Okay. Thank you. I'd like to follow up on Dr. Ayers' questions about the doses to the target site. Presumably there were variations from one patient to another treated with the device because of differences perhaps in the timing of the source placement and in the drift of the sources that's already been mentioned. Have you estimated the range of doses that were received by the target lesions over the course of the number of treatments you delivered? MR. GREEN: There's really two pieces, I think, to answer the question you asked. One is how you handle variations in the trial. Another is how do you try to look at that more, I'd say, mechanistically or from trying to see what was going on. We had a randomized trial, so what we tried to do, of course, in the randomized trial was look at what would be the realistic occurrences in actual use. So use, you know, visual estimate of reference vessel diameter to choose your dose, the catheters are placed after the standard treatments that you would expect, debulking, et cetera. So that you would get the expected type of outcomes you would expect to see in the real clinical application, and then we look to see how did those--were those effective were those safe when we did that. Now, the second part to that is to actually look at the individual patients in as many cases as we could to see what was happening. I think we--we did an IVUS-- retrospective IVUS dosimetry was done by Dr. Crocker and Dr. Fox at Emory University on 28 IVUS patients that we did have come back, and Dr. Crocker could touch on what was found in that first part of variations. DR. CROCKER: The maximum surface dose based on that retrospective IVUS dosimetric analysis was 75 Gray. That, of course, is, you know, a high dose, but I think you need to keep in mind that that is a dose that's received by an extremely small volume, an extremely small portion of the vessel wall and the dose, you know, falls off rapidly from that point. When considering tolerance doses, I think you also need to keep in mind that this is a dose--or that the volumes that we're treating are about 1/30,000th of what one might normally treat with external radiation treatment. So, yes, there were high doses received on the luminal surface of the vessel. Those doses do fall off fairly rapidly. The average dose received to the vessel wall is much lower than that, and we haven't see any adverse effects from these localized high doses in either the START or in the BERT trial. I guess I should also mention that, you know, the vessel is supported by a stent in this situation as well. ACTING CHAIRPERSON TRACY: Could I just ask you to state your name? This is being recorded, so just remember to state your name. DR. CROCKER: I'm sorry. Ian Crocker, Emory University. DR. IBBOTT: Dr. Crocker, before you go, I guess I'm understanding that the variations in the dose due to positioning of the catheter within the vessel and the dimensions of the vessel are much more important in determining the final dose to the target lesion than any variations in the length of time that the sources are in position. Those variations-- DR. CROCKER: Yeah, I think that-- DR. IBBOTT: --are much smaller. DR. CROCKER: You know, there is a substantial variation in the thickness of the vessel wall, which is then reflected in a heterogeneity of dose, and that is a much greater determinant of the heterogeneity of dose than, as you say, the estimate of the vessel size that we used in order to determine the dose prescription in this study. DR. IBBOTT: Thank you. You have mentioned the minor device malfunctions, and I wondered if there was a correlation between the frequency of these malfunctions and the institution performing--participating in the study, if you looked at the incidence of those malfunctions from one institution to another. MR. GREEN: One of the things that makes this a very interesting analysis to look at is that when you start to go into it, we were running, of course, two other trials at the same time. We had the Beta-cath system trial, and we had the START trial, and we had the START 4020 trial. So there were more than one trial going on at the site. So to begin with, it kind of made it difficult to look at a site's experience based on that. Was this the first patient they had treated in this trial and all trials, et cetera? Another piece would be that we would train--you could train multiple--we could train multiple teams of users at a site, so collection of the information was not always made on the user; therefore, we couldn't determine if it was the users at first. So what we do know is that there was a difference in the reporting at some sites, i.e., when you look at the data there was some, if you will, clumping of reports. In some cases, there was more reports from one institution than, say, the other, many of the others. DR. IBBOTT: Then that makes me wonder if there was a--if this is a training issue, if you see that as an indication of inadequacy of the training at some institutions as compared to others. MR. GREEN: I think that in all cases that we can--you know, we can look at these, and we believe that the training is very important here. And we learned with a new technology and a new device here, we've learned a lot of things about the device. And I think that that is the most important thing from the interaction with the company, figuring out what we need the companies to do with the sites and we can put that in the training program. And I do believe that that would make a difference. Again, we said as you look at things like manual removal or the bail-out procedure, as sites became more experienced and did more treatments, those numbers--the rates went down of those incidents, and it was clear that the training and learning curve was a portion of that. DR. SPEISER: I definitely feel that training is important, and there was variation from institution to institution. I think at this point training will include physicians involved in the procedure, which will be a major advantage for the future. MR. GREEN: I guess one other thing to point out would be that it's also possible--I mean, we have to show both sides of that angle. It's also possible to--some sites were--if you could be overtrained in the reporting of incidents, and maybe they--they were doing exactly what we wanted, trying to give the feedback to our design development system, and they did a very good job of that. DR. IBBOTT: One last question, and this may be something I missed in your package or my own naivete, but I'm wondering why you assigned a maximum usage of six months or 250 treatments to the device. MR. GREEN: It's quite simply based on the data, the bench data and the testing data that we have to support the number of uses of the device. We both collected information from the clinical trials on the number of uses and time of use of the devices, and we've also done reliability and other engineering tests that have been submitted to the agency. And that is what that testing supports, and that really implies a safety margin above what could actually be done. That's what we recommend. DR. IBBOTT: Thank you. ACTING CHAIRPERSON TRACY: Thanks. I just have one brief question and one comment to make. What exactly is the recommendation for antiplatelet therapy in patients who did not receive a stent, a new stent? MR. GREEN: In the instructions for use, we basically made a recommendation that reflects what we did in the clinical trial. Patients received--we recommended that patients be treated after a successful result. If they did receive a new stent, they got a minimum of 90 days. We did not make a recommendation in the clinical trials for patients without new stents. And those patients, as you saw, received what reflected probably the clinical practice of about 14 to 30 days. ACTING CHAIRPERSON TRACY: So you're not making a specific recommendation on that. MR. GREEN: We're not. We are not making a recommendation. ACTING CHAIRPERSON TRACY: There's a discordance between what you have in the labeling and what you have in your training recommendations in terms of the frequency of fluoroscopic evaluation to prevent drift. One says 15 seconds and one says 15 to 30 seconds. I would just encourage you to make sure that there aren't other discrepancies like that. MR. GREEN: Thank you. DR. FREISCHLAG: Julie Freischlag. I read your animal study summary that's in the booklet, and I guess I was impressed that it doesn't seem to be working in animals. Usually with our studies we find perhaps it may work in an animal model and not in humans. It seems that perhaps there isn't any animal proof that it worked, but it looks like it works in humans. Can you explain that? MR. GREEN: Yes. Actually, we looked at--the progression through the FDA process was actually earlier on, which was not reviewed in the memo provided by Dr. Subramanian. There were five additional studies--two 2-week studies, two 4-weeks studies, and a 6-month study--looking at the device safety and performance and the reduction of proliferation. That was provided in the PMA. It was not reviewed in the memo. Those studies were the studies that provided the safety device performance and the initial information on whether there seemed to be feasibility of radiation including proliferative tissues to go into clinical studies, and that's what we did. In conjunction or in parallel with the clinical studies we ran, the studies we've done now, the model there did not show concurrence with what we've seen in the BERT data out to four years and the START data. It did, however, provide similar information on device safety, device performance, and the initial feasibility of doing the study. DR. FREISCHLAG: So you think--if you look at why you think the radiation works, why is it preventing intimal hyperplasia in the stent, it sounds a little bit better than on the edge of the stent, what do you think the reason is that this works at all? Most of us haven't been able to figure out even what causes intimal hyperplasia. It looks like you may be able to prevent it in the stent. What is it doing to those cells in the human model that perhaps it didn't do in the animal? MR. GREEN: I think I'm going to have to defer that question probably to one of our cardiologists, Dr. King, if he could step up and provide some insight on some of the more biological issues. DR. KING: I'm Spencer King from Emory University and Atlanta Cardiovascular Research Institute. As a developer of the technology, I have a licensing agreement with the sponsor and also research support. The information in the packet seems to indicate that it didn't work in animals, but, in fact, the reason that we went into patient trials had to do with animal research. For the last 15 years, we've been using the pig model, tried to look at not restenosis model but a model of vascular injury. You blow a balloon up, you damage the pig coronary artery, and you get a healing response. It's not restenosis, of course, because the pig never had stenosis in the first place. All our work was short term, two-week to a month work, with radiation, and it had been the same with other things, lipid-lowering agents and a lot of things over the past 15 years. What we saw in the animal lab, with the work of Dr. Waxman and Dr. Robinson in our lab, was that the animal proliferation was dramatically inhibited in that model, two weeks, at four weeks, and we had some six-month data at Emory that was not included in the packet. That encouraged us to apply to FDA for the feasibility trial, and that's what got us to that level. I would emphasize that this model--we don't know from radiation or other therapies what the long-term pig effect in growing young pigs is in terms of modifying a healing response after vascular injury. So the model I think was a good model to look at acute proliferation, migration of cells. We know things about the effect on measures such as bromodeoxyuridine measures to show that cell proliferation is being reduced and that sort of thing acutely. And that led us into the clinical trials. But in terms of the pig having a sustained long-term benefit, the pig model, particularly with stenting, is a very difficult model. A lot of thrombosis, we've had a lot of deaths with pig models using all kinds of agents. And so I don't know why or even if we should expect that it has this good long-term effect in the pig. DR. FREISCHLAG: If you were to hypothesize how this works, you think it is because it decreases proliferation of the smooth muscle cells? Do you think anything else might be involved besides that? DR. KING: Yes, well, we think that in terms of the lesion that develops in the pig model, that is, cellular, that those cells are inhibited. We've seen that with, as I say, the proliferation part with BrDu staining. There's work on apoptosis that may be a part of this formula as well. There's extracellular matrix elaboration that may be modified. There's also a modification of the vessel contracture, we believe, that occurs, not as well demonstrated in the pig as in the patient with IVUS examination, but we think it affects not only the smooth muscle cells but also fiberblast and periadventitial scarring that occurs with contracture of vessels. So we think all these mechanisms are operative in what we're seeing. Now, with the stent model, the contracture part is less active because the stent is holding everything open. So in the stent model it's almost all animal proliferation and extracellular matrix elaboration. DR. FREISCHLAG: Did you ever treat normal vessels that hadn't been dilated or injured with the radiation? DR. KING: Yes. Normal vessels have been treated, and that's been reported, and there's been little observable effect in normal vessels. DR. FREISCHLAG: Did you check for apoptosis in normal vessels? DR. KING: I might have to defer to my colleague. I don't know if Ron is still here. Ron, did you do apoptosis in the--I can't remember in the normal vessels if you looked at that? ACTING CHAIRPERSON TRACY: Can you come up to the microphone and introduce yourself? And also state whether you have any relation with the company. DR. WAXMAN: I'm Ron Waxman. I'm from the Washington Hospital Center. I do have an interest in the device, and I'm entitled to royalties through Emory, an agreement, but I'm not(?) serving a consultant to the company. We have performed studies in which we have seen apoptosis in normal vessels and injured vessels as a result of the radiation. There were different time points as compared to the placebo, but definitely apoptosis is one of the mechanisms. There is also clearly a reduction of acting as smooth muscle cells--staining to smooth muscle cells, (?) staining. So we have seen a reduction of the smooth muscle cell population, and I think literally what we're doing with the radiation is we're killing the cells. Now, as opposed to the models, at two weeks and four weeks it's very suggestive that you reduce the amount of new intima formation within the stent. When it comes to six months, there is actually no studies except one that we have performed before with gamma radiation that in others that showed that you do have a benefit on the long term. But in terms of the answer to the apoptosis, it's probably part of the mechanisms, but we don't know to what extent it's explained the entire phenomenon. DR. FREISCHLAG: One more follow-up. Have you followed normal vessels treated with the radiation longer than two weeks? DR. WAXMAN: We have performed the examinations, and there were others that reported on that. If you take a source in a normal artery without injury, you do see some effects, deleterious effects on this artery. Now, it depends on the dose. This is only coming from animal data, and it was shown on the rabbit and in the pig. However, in our human data from gamma radiation, when we looked by ultrasound on the areas that were exposed to radiation and were not injured, we have not seen any deleterious effect at six months with intravascular ultrasound on these segments. Again, these are segments that were not injured and were treated with radiation. So there could be some differences between the animal models and the human, which we appreciate, and we can summarize in this respect that the animal data does show some effects. The human data so far on arteries that were not exposed to injury does not show any deleterious effects. It comes from our lab in the Washington Hospital Center, and also from the Toric (ph) Center there was a paper that was published and a paper that's about to be published. DR. FREISCHLAG: I had another question concerning the edge effect, and perhaps this is best to Dr. Kuntz since he talked to us about that. Is 5 millimeters enough to look at on either side to see whether or not your edge effect may actually be further away from your treated site? DR. KUNTZ: Right. Rick Kuntz, cardiologist in Boston. There are a variety of ways to look at the edge effect. The first analysis--these are very labor-intensive analyses to be done, and maybe if Dr. Lansky can talk about the more technical aspects of this, it would be helpful as well. But as an opening statement, we chose this analysis or, more precisely, Dr. Lansky chose this analysis because the biggest concern about edge effect was the point of the end of the source; that is, if there was any theoretical problem associated with the radiation therapy on normal vessel, or on a vessel causing so-called edge effects, that is, effects not targeted outside the target lesion, that is, effects not anticipated outside the target area, it would be at the area where the patient was exposed to injury and potentially exposed to reductions in radiation therapy. So because of that, the edge of the source was the center of the analysis, and we went 5 millimeters on each side to bracket a 10-millimeter segment. Now, since the source is 30 millimeters long in the majority of these cases and the stents are on average between 15 and 20 millimeters in length, that 10-millimeter bracket covered almost everything up to the edge of the stent in the majority of cases. But there was possibly a few millimeters on each side of the stent that wasn't included in this analysis that was still outside the stent. And that would require a second analysis to go forward. I feel pretty confident that the analysis, because of its broad range in this 30-millimeter train for 20-millimeter lesions, was pretty comprehensive in looking at the edge effect as it was done up front. But there are multiple different ways to look at edge effect, each of which will require different labor-intensive analysis to go forward. So I would say that as an initial look specifically at edge analysis, which has never been done in any other study yet, that this was a very good first pass and probably covered most of the territory which we could see the problem of theoretical edge effect if, in fact, it did occur. And I do want to leave the podium open for Dr. Lansky if she wants to make any further comments. DR. FREISCHLAG: Was there any effect in other vessels that may be in close proximity to the treated vessel with effects that may be related to the use of the device? DR. KUNTZ: That's a very interesting question, and we have not specifically looked at that question, at least I haven't specifically. I do know that there are indicators indirectly that suggest that there was no untoward effect. That is, we do have information about the frequency of repeat revascularization in other vessels that looks very similar to other trials that we've done without radiation therapy. But we haven't done a specific analysis looking at the incidence of new events at other vessels in the adjacent area. I think theoretically it might be tough to imagine that vessels that may be separated by 4 or 5 centimeters might be affected by beta radiation in the heart. But at first blush, looking at the distribution of revascularizations, we didn't see anything that looked like there was an increased frequency of new lesions or restenosis at the other vessels. But I can't say that we specifically looked at that question. DR. SPEISER: Can I make a comment on that? Burton Speiser, radiation oncologist. Along the source axis linearly, 5 millimeters away within the vessel, the dose falls down to about 1 percent. So it's a very low dose. At a right angle, the dose falls down well below 1 percent at 10 millimeters, so that any of the vessels close by will get a very low dose. DR. FREISCHLAG: Dr. Speiser, you mentioned when you spoke that the radiation effects in vessels you felt from this device would be seen at 6 to 24 months, is the range I think I heard you say, and you followed these patients for 8 months. Can you hypothesize what might happen to these vessels up to 24 months, or have you seen any vessel treated with your device at 24 months after it's been treated? DR. SPEISER: The only information is from the BERT study where there's 4-year follow-up. Is there adequate anatomical material? No. But I stand by the statement that most radiation effects usually have a time course of 6 to 24 months after delivery when we see that. The higher the dose, the sooner we see the effect. And what type of effects? It would be decreased proliferation so that if you gave a very high dose of radiation well above what we're giving, you would completely stop all healing and lead to other problems such as aneurysmal formation. Another effect, which would be less likely, would be usually destruction of the endothelial lining can lead to a problem, but the endothelial lining is already destroyed during the process. So that the other possibility would be fibrosis in the future. However, because of the small area, radiation also has an effect that's volume related, so that the smaller the volume or length of artery, the less the effect for the same dose. DR. FREISCHLAG: Is the vasovasorum preserved with treatment of these vessels when you've looked at them after treatment with your device? DR. SPEISER: Well, the vasovasorum are usually in the wall of the major arteries, and those would be affected if we were to treat them. For the most part, we're in much smaller vessels, usually about 2.7 millimeters, so that it's less of a concern. And those vessels have an intermediate sensitivity between the very small and the larger vessels. But specifically I know the effects of the vasovasorum, but not with this device. DR. FREISCHLAG: And my last question has to do with the 20 percent of patients in each group which got that stent. Did you look at those cohorts separately compared to head to head? Because couldn't they--the reason they got better was the stent and it had nothing to do with the device or anything else we did to those patients. And was the stent the reason that those patients at separate cohorts in both groups did better? And did you compare those two 20 percent head to head even though I know it's only 100 patients, but was that the group that did better in both groups, and, therefore, we really should be just putting more stents in and not doing something new? DR. KUNTZ: Right. That's a very good question, and we did do the analysis, and what we found was that treatment effect was diminished in patients who received stents, that is, the cohort who received stents, the difference provided by radiation therapy was diminished compared to patients who did not receive stents. However, the ultimate restenosis rate for both groups was intermediary between the actual treatment effect on patients who didn't receive stents and the placebo arm. So that, on average, we may say--and I'll just use this as an arbitrary example. It's not the real numbers. Say, for example, we had a 15-percent restenosis rate with active patients treated without stents. We may have had a 25-percent treatment rate on average between the group that received stents, a little bit better for radiation compared to political, and a 40-percent restenosis rate for patients who didn't receive stents and had placebo. So the answer is that when stents were placed, the radiation therapy was diminished, the effect was diminished; however, the performance of antirestenosis effect observed for both the radiation arm and the placebo arm were not a substitute for the largest effect seen in patients who didn't receive stents overall. So stenting was better than placebo, but not as good as radiation therapy for patients who did not receive stents. DR. FREISCHLAG: Okay. Thanks. ACTING CHAIRPERSON TRACY: Dr. Krucoff? DR. KRUCOFF: Mitch Krucoff from Duke. I also want to thank everybody for the clarity of presentations and what's obviously a strong fundamentally designed clinical trial to look at the safety and efficacy of their device approaching a very tough and complicated clinical problem. I do have some questions, and a couple relate to pieces of data that you guys showed that, at least to my eyes, were not in the panel pack. And, Jeff, if you could go back to them, they were on classifications, which you mentioned. We have seen data elsewhere as affecting significantly what we would expect from a procedural outcome. I didn't see anything in the START data that actually characterized the patients on inflow, the patients enrolled in the study, relative to the nature or the class of in-stent restenosis. Were these all truly in-stent restenosis or were these all classifications from diffuse restenosis and in between? DR. POPMA: There are some lesion characteristics that we know for comparative purposes. The average lesion length was 16 millimeters, which is longer than we typically see for a focal de novo lesion. I think that your point is well taken, and I wouldn't want the introductory slides that we have to confuse the picture at all. I think the summary statement that we're trying to make about showing the Mayron data was that there is some heterogeneity with respect to outcome based on what the pattern morphology was. Because this was a randomized trial, there was equal distribution of the complexity of the lesions in both groups. The investigators at the clinical site were blinded as to whether the patient was going to receive radiation therapy or not. So it would be very unlikely, and we have not seen any of the data analysis so far, that there was a misreputation of more complex lesions in one group and less complex lesions in the other, because they were--the investigators were blinded at the time of randomization. One thing about these trials as you look at the literature is that the restenosis rates in the literature are fairly heterogeneous. Some trials will have very low restenosis rates. Other trials will have very high restenosis rates, which makes registry type comparisons very, very different. What was done in this trial was a randomized clinical trial, randomly assigning the two different complexities of the lesion subsets to treatment or to placebo. And I think that's the real strength of doing a blinded randomized clinical trial in this fashion because the appropriate issues that you're raising about the fact that morphology of the baseline lesion can affect result is absolutely true, and may likely be the source of bias in a trial that's not randomized and blinded. So we know at least in our trial that the restenosis rate in placebo was ranging between 41 and 45 percent in the placebo group, with equal distribution of pre-procedural lesion morphology. And we know that the treatment associated with that was significant resulting in a 36- to 66-percent reduction. DR. KRUCOFF: Okay. My question may be a little different or you may have answered it, so let me just find out. Lesion morphology, for instance, an 18-millimeter lesion, slightly eccentric, inside of--that is the result of having placed a 9-millimeter stent would be classified differently than an 18-millimeter lesion that is in the middle of a 30-millimeter previously placed stent. So I guess what I'm asking is the lesion complexity, as it was analyzed, at least in my appreciation, was not the Mayron type of classifications relative to the previously implanted stent. DR. POPMA: Correct. That's correct. DR. KRUCOFF: And I just wonder--and I think a theme that we're going to come back to--because I'm literally sitting here thinking about if I were to use this device selecting patients for whom it was most appropriate, what would go through my head? DR. POPMA: Right. DR. KRUCOFF: And one of the things that I think we've talked about, as was mentioned earlier, with the longer lesions possibly having a greater benefit, is where is the real benefit relative to classifications that we know are-- DR. POPMA: Right. DR. KRUCOFF: So did you guys actually look at whether the Mayron type classifications randomized equally? DR. POPMA: I can let Rick answer that as well, but oftentimes the initial stent length itself was difficult, and I'm not aware that we did any analyses looking at lesion length compared to the relative stent length at the start. So your point is well taken. A lot of the retrospective data that came from the Hospital Center for that was when we had 15-millimeter stent available or a 20-millimeter stent available. DR. KRUCOFF: Right. DR. POPMA: But now with the tremendous heterogeneity of stent lengths ranging from 9 to, you know, 34, absolutely there can be some relationship that may require refinement of the classification system. For the purposes of the study, which is a clinical and angiographic study, we know that the complexity of the lesions was the same, and there's no reason to think that the stent length pre-procedurally wasn't the same in the two groups, because that was not--that was also blinded to the investigators. So I think the bias was introduced, but I'll let Rick comment about that. DR. KUNTZ: Rick Kuntz. I just want to maybe clarify that for you, Mitch. Following up on Jeff's comment about the fact that there were longer stents available for this study compared to when the analysis of Washington Hospital Center was done, almost all these stents, if not all these stents, contained a lesion; that is, there was a requirement on entry. So these lesions tended to be within the stents in almost all cases, and I think that's probably true, Dr. Lansky. The second is that we do know the distribution of the lesion lengths. The average lesion length was 16 millimeters, but we do have a breakdown of other lesion lengths by minimum amounts. For example, one-third of the cases had lesion lengths of at least 20 millimeters long, and when we averaged those patients, it actually was 25 millimeters. So we did have a fairly large cohort of individuals that had long lesions in this study. The other issue is that we do know that this study verified the fact that the longer lesions made the placebo group at risk of having higher restenosis. But we also know from the interaction terms that the radiation therapy had its biggest effect in the longest lesions. So your notion that there's a cutoff by which radiation therapy is effective is actually true. Where that cutoff exists is hard to say, but most likely--and I don't know if this will be reflected in the label or whatever. This is probably not appropriate for short discrete lesions. It's probably more appropriate for lesions that are moderate lengths and longer based on the data we see so far. DR. KRUCOFF: I guess part of my other concern is whether we know or whether we can tell or whether you can tell us from the data you have whether lesions that are restenotic and extend beyond the margins of a previously implanted stent would be effectively treated. DR. POPMA: Right. I know that you know this, but sometimes it's hard to see the stents, and even as an investigator in the trial, I try to guess where the stent is based on what I see angiographically. But sometimes it's still difficult to do. DR. KRUCOFF: Okay. Another, I think, new piece of information, Rich, was the data from Dr. Lansky's QCA lab on the edge effect, and that to me is very important, and I'm not sure I really understood what you said. What I think I heard you say was that when you looked at the edges per se, the edge areas, the placebo-treated group and the Sr-90 treated group were actually not very different, and that some of that may be the result of the healing of the inner segment being better as a relative artifact or illusion. And yet the data--I guess my question is: Are you saying that the analytic segment binary restenosis rates that are reported in the panel pack are, in fact, not hemodynamically significant restenoses? DR. KUNTZ: No. I'm glad you brought that up, Mitch. I think from my perspective as a cardiologist, the analysis segment restenosis rate is the actual restenosis rate that we should be quoting, because that's what the patient cares about. The patient doesn't care whether their narrowing is in the stent or somewhere else. So the overall--this treatment effect was 36 percent. That's what this trial shows. We showed that within the stent it was profound, but that doesn't matter if it turns out there's a lesion on the outside. So I don't think anybody is trying to say that the actual treatment effect can be nullified because the analysis segment is more. I think what we found, though, is that after you treat a segment, that is, what we call the analysis segment or from--you know, wherever the radiation therapy and a little bit outside that was, then almost all patients had an opening that was less than 50-percent residual. What Dr. Lansky found was that when we look just at the edges, there was about a 12-percent rate of narrowings that actually tripped the 50-percent threshold that in and of themselves could be called a restenotic lesion. Okay? If you have a lot of failure in the middle of the stent, you never get to see those 50-percent lesions show up because they're always the second or third MLD, not the first. If you have an effective therapy in the middle, all of a sudden they become the minimum lumen diameter. So that's why the radiation therapy has such a jump from stent to analysis segment and the placebo had not much jump at all. That's the artifact as to why it looks like there may be an edge effect, but, in fact, the edge narrowing portion, which represented 10 to 12 percent of the cases, was identical for both placebo and for active arm, suggesting this is just the typical carrier restenosis effect associated with dilating a stent on the outside. DR. KRUCOFF: Okay. But we are on the same page that ultimately the binary analytic segment or target vessel restenosis rate is a real clinically meaningful-- DR. KUNTZ: And that more reflects the level of degree of effect seen in the clinical restenosis rates of 30 and 40 percent as well. DR. KRUCOFF: Okay. A couple of quick questions. Is it the notion to go forward that the way to deal with the touie borst clamping down too hard is to put in an arrow sheath introducer? I haven't heard of any other either existing or planned engineering designs that would--that seemed to me to be a significant, if perhaps not a cause of patient harm but a significant operator nightmare to clamp down a little too hard on the touie and then not be able to deliver or retrieve the train. Is the plan simply to go forward with an arrow sheath introducer as an optional part of the instructions for use? MR. GREEN: That's the current plan for this device as you go forward, the arrow sheath as well as the training on not only the use of the arrow sheath but also on the use of, you know, the touie borst, the catheter, the entire system, as well as going through training on what could occur in a clinical procedure and how to avoid that. DR. KRUCOFF: Okay. And am I correct in reading the routine use of the instrument involves over-the-wire both insertion and retrieval, while the bail-out strategy for the instrument includes removing the guide wire from the coronary? MR. GREEN: The manual removal procedure, the bail-out procedure could be done over the wire or as an entire catheter and guide wire. DR. KRUCOFF: Now, as I read the instructions for use, it's pretty clear that you say yank the whole thing. MR. GREEN: Removing the catheter and the guide wire at the same time would be the most expeditious way to remove the entire system and limit the exposure to all personnel and patient. DR. KRUCOFF: And the reason for an expeditious removal is an exposure issue rather than anything else? MR. GREEN: Yes. DR. KRUCOFF: I wonder about just ischemic tolerance of this device. I don't see data reported to it. Obviously your patients are pre-selected to already have a fairly well dilated segment, so presumably that gives a much more forgiving environment to cross a lesion. But dwelling for 5 minutes in any coronary, as we know, is not always well tolerated even with IVUS sometimes shorter passes. Can you give either data or a flavor of what the ischemic tolerance was or whether there was a non-exposure reason to remove the device encountered during this trial? MR. GREEN: This was actually a question asked to us by the FDA, and we did provide it. It's provided in the panel pack in the addendum to the START report, in the START section on page 12, page 12 after the START clinical report. It's called average dwell time and patient tolerance. It provides the average--again, the average dwell time and describes only one of the patients who was unable to tolerate the dwell time necessary to deliver the dose. DR. KRUCOFF: Thank you. And in the instructions for use, it also talks about what to do in case of a breach of the system. Either in the Beta-cath or START experiences, I didn't see any description. Obviously that wouldn't be a minor sort of event. Have you actually encountered that in the human application? MR. GREEN: No, we have not encountered that in the human application or in animal application. DR. KRUCOFF: Okay. I'm going to pass over my questions about dosing, although I think they've already been addressed, but hopefully in the future we'll continue to learn about eccentricity and curvature and clearly I think the dosimetry involved here are some of the real open-ended--still issues, but I think they've been largely addressed. In the START study, I was very struck that 40 percent of the patients had rotational atherectomy applied to debulk the stenosis, even though, as you showed in European and other data, there's been the suggestion made that this may be more a stimulator than a solution to in-stent restenosis. Can you share some of your thoughts about how you think that falls out? Were there more detailed looks? I'm sure you took more detailed looks at whether this was an interactive factor. I didn't see it as a feature in either the univariable or multivariable models. DR. KUNTZ: Yes, Rick Kuntz. We did look at the effect of different pre-treatments on outcomes, and we found that there was no significant effect, which essentially validated some of the other debulking studies in the past and showed no significant differences if you used rotational atherectomy versus not. I can share my interventional cardiology kind of qualitative feelings. I think Dr. Popma may want to chime in as well. I think the study, the ARTIST study, which is actually the only good randomized study to look at the effect of rotational atherectomy as a debulking agent, may have had some limitations; that is, the burr sizes used were small and the tolerance for an acceptable result is different in Europe than what we do in America. So I still feel in my own practice that there's a role for debulking, especially in big beefy lesions in large vessels. And if we can get a big burr in there, it's very helpful because it just reduces the amount of plaque that we have to push through the struts. So I think that this data validates the fact that rotational atherectomy was not detrimental, and I think for selected individual cases, the discretion of the operator who is versed and comfortable with rotational atherectomy I think is still compatible with the use of this device. Jeff, I don't know if you want to comment. DR. POPMA: I was hoping that we'd have some data comparing those that got rotational atherectomy with those that didn't, because amongst the 50 investigators cited, it becomes a little bit religious about whether you use it or you don't use it. We didn't see that there was any beneficial effect to rotational atherectomy and yet another trial. I agree with Rick that we still want to do something for very diffuse lesions. That involves debulking. But I have to say as a clinician scientist that we sure haven't proven that so far. We didn't see that there was an interactive effect with rotational atherectomy and radiation. It appeared that radiation worked in those patients who got rotational atherectomy, and it also worked in those patients who didn't get rotational atherectomy. I think that's the most we're going to be able to say about debulking from the study. DR. KRUCOFF: I'm sorry to hear that because, you know, I think really and truly what you'd love to know is if this had just been balloon with only provisional stenting, whether beta radiation would be enough. And as everybody knows, in the practice at this point the relative costs of adding these technologies together--rotational atherectomy with angioplasty, with beta radiation--is not a negligible increment. So I guess maybe an unfair question, but if this had just been a balloon angioplasty study, do you have any basis for saying whether you think the effects we are seeing would be the same or larger or-- DR. KUNTZ: Sure. We do have indirect evidence of that. We did multivariate modeling to look at the predictors of restenosis, and after adjustment for those predictors we added the variable of rotational atherectomy. It was not effective. This study supports the fact that this works just as well with balloon angioplasty as it does with rotational atherectomy. We also did interaction terms to look at any interaction between the devices and the outcomes and found none. So, actually this device is more supportive by the fact that the plain old balloon angioplasty is probably just fine and less supportive for the use of rotational atherectomy, especially if it comes down to a cost issue. DR. KRUCOFF: Okay. So you know my next question, which is Reapro (ph). What was the instance of usage, and did you examine it again? It isn't indicated in the multivariate or univariate analyses, but did you look and can you share with us what it looked like? DR. KUNTZ: Yes, the Reapro was discouraged in this study because when the study was started, there was a feeling that--or at least a disseminated feeling that Reapro reduced restenosis. So because of that we didn't want to have a potential confounding effect of an unbalanced Reapro effect explaining differences in restenosis. So given the fact that in-stent restenosis is generally associated with a low complication rate, anyway, and that Reapro is used mainly to prevent complications and the fact that Reapro may have had a potential to reduce restenosis based on when this trial was designed, we tried to discourage the active use of Reapro in the study. So there was Reapro used, but the frequency was low enough that it didn't allow us to do any meaningful analysis. DR. KRUCOFF: What was the frequency? Twelve percent, I think--right around 12 to 15 percent? DR. KUNTZ: Right. Very low. DR. KRUCOFF: Learning curves, you know, this I think has been touched on several times, but it seems to me despite this semi-disclaimer earlier that this was too complicated to look at because people are in multiple trials going on at the same time that actually it's probably n to all that complicated to look at, and whether particularly your minor device failures were in earlier phases of operator experience or not would I think be a very meaningful piece of information, particularly if you're going to mandate training or some other kind of approval-oriented condition. So have you guys actually looked at how cases 1 through 5 and operators went into double digits fared with minor problems relative to later cases? MR. GREEN: Again, what we tried to do is go back and look at that, but the problem is that--and I'm not saying we're not still trying to gain this information, but the problem becomes that it wasn't something that was captured, the individual user wasn't captured on the case report form. So it's something you'd have to go back to the patient forms back at the individual institutions in the clinical trial and collect. So it's not something we have available, so it was not an analysis we were able to do. DR. KRUCOFF: Okay. Patients who were enrolled clearly by the structure of the study must have been a subset of patients who were consented, since you had to have a successful angiographic result in order to go on and be randomized. Can you give us a sense of what percentage of patients who were consented for these procedures actually had a successful enough procedure to enroll? DR. POPMA: I can tell you what happened at our site, but, Rick, do you have any specific-- DR. KUNTZ: We didn't look at a universal log to look at that percentage, which I think would have been very valuable to see how many patients came in front. I can tell you at our center the vast majority of patients who came back with chest pain after stent placement we felt had clinical in-stent restenosis who were consented before they received the pre-treatment enzyolitic(?) therapy were the general catchment area. Among those patients were people who we verified had in-stent restenosis by angiography, and because of the practice patterns of actually treating patients after their diagnostic angiogram at the same setting, there were patients who would be treated and then qualified for the study. It's my experience that that probably represented close to 90 percent of the patients, that the vast majority of individuals with standard therapy qualified for this analysis, but we don't have any hard data to show that because it would be very, very hard to start from the consent portion and go through, and we prospectively didn't capture each of those logs before the randomization actually occurred. DR. POPMA: In our institution now, either with the trial and with the confines of the compassionate use trial, we have been able to treat 90 percent of patients without putting a stent in before we give radiation therapy. So it's frequent. The answer is we can often get them in. DR. KRUCOFF: Okay. And, Jeff, I heard you respond earlier--if I heard you correctly--the 21 percent of stents that were placed were placed after the Beta-cath-- DR. POPMA: That's correct. DR. KRUCOFF: --had been positioned, either mock or real treatment was administered, and then the catheter was removed. DR. POPMA: That's correct. DR. KRUCOFF: Then there was an angiographic deterioration that was of concern enough to do something in addition. Can you tell me how, then, we would understand whether that 21-percent deterioration was just from natural recoil or how we would know that it was or was not a result of sticking the delivery system through the previously dilated artery? DR. POPMA: That's an excellent question. It's an excellent question because I'm not sure that we have definitive data that would tell us the exact circumstances before and after the stent implantation. This was recorded on the angiographic core laboratory sheets, but oftentimes, as you know, in the cath lab the exact occurrence of the events are not actually filmed. The biggest fear would be that there was a suboptimal result before the radiation was done with the investigator knowing that they were going to try to get the radiation in and then see how things looked. I would hope that wouldn't have been a frequent occurrence in the trial and that they were done for real legitimate reasons with respect to the bail-out stenting. The majority of the cases were done for residual stenoses. The minority were done for new dissections. And I think it is reasonable to suggest that there is some reintrusion of plaque within the vessel wall. The message that we'd like to give from this trial is avoid stenting if possible. We want to make sure that we have a successful result first, make sure that we really do have a less than 30-percent residual stenosis and there really are not significant dissections; and if that's the case, then to deliver the radiation therapy as prescribed in the IFU. If after that the patient's in trouble and has a greater than 40 percent stenosis or has a dissection that develops, then to not be afraid to use stenting, use it because you're correcting a complication. My personal feeling is that that number will be actually much less than 20 percent. At least it is in our practice, that we're actually able to not stent often in the vast majority of cases. DR. KRUCOFF: Well, the issue--and I'm sure you can see it--is that while--as we look, for instance, at the acute outcomes with radiation or non-radiation, they're spot on, they're identical. On the other hand, if putting the delivery system across an adequately dilated lesion itself engenders a 21-percent complication rate regardless of whether or not you radiate, then we have a whole different kind of dilemma. DR. POPMA: Right. DR. KRUCOFF: And, you know, I'll say up front this is a Catch-22 because you guys did what I think was a very reasonable study direction to blind this. To really get at the effect of radiation, you got to put the device in or you can't blind it. DR. POPMA: Right. DR. KRUCOFF: But I really just wonder what your comments are, whether 21-percent deterioration of lesions in the cath lab from before putting the device in to after putting the device in, I think you have to at least examine the question: Does it relate to putting a new piece of hardware across an adequately dilated stenosis? DR. KUNTZ: We actually did some analysis on that which I think will put you at ease a little bit. One is that 65 percent of the cases were done for a suboptimal result. And so when we looked at the pre-stenosis, the stenosis after initial treatment but before stent placement, compared to those people who didn't get a stent, there really wasn't much of a difference. What we really saw in this study was a different belief pattern of the use of stents to improve a cosmetic result, and that was--as you recall, when this study was started, there was a huge interest in restenting in-stent restenosis lesions, much more than there is now. So because of that we had a heterogeneity of some sites who tended to use stents more often for the same residual stenosis that other sites would leave alone. So the vast majority, 60 to 70 percent of the cases, were not done because of inducible dissections by the delivery catheter, but were done at different belief patterns as to whether or not you needed to optimize a stent--an in-stent restenosis result with a second stent. And when we analyzed these data, the overall rate was 20 percent. When we actually went back and looked at residual stenoses and dissections, only about 20 to 25 percent of the cases could be justified as a dissection. And we're looking at 20 to 25 percent of cases of the 80 or 90 patients who got stented. We're talking about 18 or 19 patients out of 470 that had evidence of a dissection induced by the device. DR. KRUCOFF: Okay. Two last quick questions, and I'll be quiet. One of the concerns, I think, again, thinking about the use of this technology, if it came to market, is the later picture, what does happen afterwards, and I have two questions related to that. One is from an animal model. Does anybody--or can anybody describe to us at what point in time, one year, three years, five years, tissue either is normal or at least in a stable, healed, scarred, whatever state, following radiation therapy? My second question is: In human beings who underwent, as in the START study, radiation who still had restenosis, again, after receiving treatment, what do we know, have they been redilated? Is the tissue more friable? Does it dilate well? What happens when you have someone who has a radiated segment that's still restenosis and you approach it again with intervention? Do we have any data on those individuals? MR. GREEN: Maybe we could come back--as far as what the--in the animal model, as you asked, what the breakpoint is for when endothelialization occurs, I don't believe that we can answer that from our animal studies. There are some suggested data out there from other animal models and studies, but I'm not sure that's a clear answer. We can bring someone else up to talk to that as well. I don't know if Dr. King maybe has something to add to that. DR. KING: A short answer. No, I don't think we know the time when everything is normal. Animal models--we don't have any animal models that far out. There is experience longer than four years with radiation, intravascular radiation, the longest being ten years from the experience of Lehrman (ph) in the superficial femoral artery. But I don't think we have any information that tells us exactly what the nature of the arteries are years later. ACTING CHAIRPERSON TRACY: Actually, I think we better move along. DR. KRUCOFF: Okay. Can I just ask the last question? The patients in START who received beta radiation therapy who restenosed within the eight-month follow-up, what happens to them? MR. GREEN: The patients in the START trial who did need reintervention later went through the standard revascularizations--angioplasty, CABG, et cetera--that were available to the placebo patients that failed. They were all available for this. DR. KRUCOFF: Is this an observation? Are there data? DR. POPMA: Maybe Dr. Lansky can speak to this, but the lesion length at follow-up was somewhat shorter in the treated patients than it was in the placebo patients, suggesting that the restenosis that did occur was more proliferative. It was less proliferative with the Strontium-90-treated patients. What that means to me as a clinician, those that we've treated, is it's a simple percutaneous therapy if the lesion length is not quite so diffuse. That's the structure from the data, but if you're kind of asking the question about how we would manage it, if the lesion recurs and it's relatively focal, not diffuse, that allows me then to treat it one more time percutaneously. I don't think we have any of the data so far on the outcomes of the re-retreatments for in-stent restenosis. ACTING CHAIRPERSON TRACY: Dr. Wilson? DR. WILSON: Thank you. Frank Wilson, radiation oncology. I also appreciated the clarity of the information that was in the briefing book. I do have some questions, though, that probably I'm the only person in the room that doesn't know the answer to them, and I apologize for that if that's true. But I think my questions mostly relate to the device itself, if I could--as I think about it. This PMA specifically is asking for approval of the 30-millimeter, 12-seed source train as part of the device. But during the time period of the START trial, there was also a 40-millimeter, 16-seed source train that was utilized, and my questions about that are the following: Is the data that's in this book pertinent only to the 30-millimeter seed train experience? And if not, is the 40-millimeter seed train experience analyzed separately? My question is obviously related to the fact that radiation effects are going to relate not just to the dose and fractionation, but also to the volume, in this case length of tissue that is irradiated, whether there are favorable or unfavorable effects. And it isn't clear to me that we're not talking about the experience with both of these source trains in this information. MR. GREEN: You're correct. We're asking for approval here for the 30-millimeter Beta-cath system. The trial did allow for the use of both. There's a very small, about 5-percent use of the 40-millimeter system. Table 3A on page 27 of the clinical report in the START section does look at that, and actually, I think I'm going to let Dr. Kuntz talk about the analysis. DR. KUNTZ: The bottom line is the estimates look about the same, but they're so underpowered it's hard to say if the 40 is going to be all right. The estimates, though, are consistent with the 30. My guess is that 40 is going to be just fine, but I think it's probably safe to say that the company's not going for a 40-millimeter label on this analysis. DR. WILSON: Does the sponsor feel that all eligible patients who might benefit from the procedure can be treated with the 30-millimeter train, or is there going to be the temptation on the part of users to treat lesions that are not necessarily satisfactorily treated with other than the 40-millimeter seed train? MR. GREEN: Of course, we'll only be providing the 30-millimeter train, and the instructions for use and the training we'll be providing will be on how to treat with that, and the recommendation for that is that treatment with--lesions that can be treated with up to a 20-millimeter balloon. You can't--because we're looking actually at injury created here that can be covered by the source train, and that will be what the training and the instructions for use recommends, and that's what we will advocate because it's all we can support with the clinical data. DR. WILSON: Along this same line, the Alpha IV device is what's being specifically requested for approval, I believe. But it's unclear to me whether the bench testing of the improvements in--what is it?--the touie borst valve and the LED pressure indicator system. Is bench testing still going on with those, or was bench testing completed? Were any problems related to that valve related to pressure maintenance? Are those now corrected fully to the sponsor's satisfaction with the Alpha IV device? MR. GREEN: We have completed the testing that we submitted in the PMA. The FDA is currently reviewing that. Now, they may have additional questions as the review goes on, as they had suggested. We qualified the arrow sheath. We did the testing to show and support that the pressure LED does, in fact, show or reflect the pressures necessary to do in individual increments what's supposed to be occurring with the device at that time. But, again, it's under review with the FDA currently, and if they have more or additional testing, we will provide it as well. DR. WILSON: Well, also, concerning what's been called the minor device malfunctions, which is on the order of 20 percent, as I recall, most of those seem to relate to source transfer. You can either--and hydraulics is just one way of transferring sources. When you're transferring high dose rate brachytherapy sources, you can do it manually or with projection cables and, in fact, both of those approaches are more common than hydraulics. But if the transfer is interrupted as the sources proceed into the patient, that's one thing, creating no radiation hazard for the patient or users, by and large. But it's another if the transfer coming back is not expeditiously possible. None of these MDMs were of that latter type, were they? There was no transfer problem where the source could not come back home? DR. SPEISER: They were not included in the MDMs. That would be in the removal portion. DR. WILSON: It does say seven where it was aborted. DR. SPEISER: Well, basically what would happen is if you had a small increase in transit because you didn't apply enough pressure or there was a micro kink at the delivery catheter within the touie borst valve. You can slow down the return, the transit time to more than 5 seconds. If, however, there was any concern at all, then the radiation oncologist was obligated to remove the entire system, whether there was truly a problem or not, to be on the safe side. DR. WILSON: And the failsafe is separate from extraction of the entire system; is that right? Can you explain-- DR. POPMA: The failsafe is in the device to prevent the sources from exiting unless the proper catheter is docked within the device correctly. DR. WILSON: Okay. I think those are the only questions I have. ACTING CHAIRPERSON TRACY: Dr. Najarian? DR. NAJARIAN: Ken Najarian from the University of Vermont. Most of my questions have been answered, but I did have a question about--I was surprised by how many stents were placed, particularly when the study criteria specified that stent placement really would exclude a patient. And I understand you use it for a bail-out, and I, you know, kind of agreed with your explanation. But it does seem quite high, and I have a hard time believing that there were 101 cases where bail-out was realized only after the radiation was given. But since you do have the data and you have split it out somewhat, I was wondering, have you looked at the clinical difference between those patients treated with radiation without stents and those patients treated with radiation and with stents? Now, you've broken out the data with restenosis, and it looks like the restenosis rate is higher in patients who were treated with stents compared with those treated without stents. But how about the clinical outcome? Was that different at all? And then my second question is: Why not just drop those patients from the study or at least segregate them and look at only the 375 patients who were treated--I believe it was 375--treated without stents? DR. KUNTZ: We did do analysis of the stent use, and we found basically that when stents were used, the radiation effect was not as good as when the stents weren't used, and that's clear. That went across all the different outcomes. The sample size of 100 out of 476 was not big enough for us to see any significant difference, especially with the clinical results. But there's no question that when the stents were used, the effect of radiation was diminished substantially. However, the comparison of a stent using placebo was still not as good as radiation without stent, so I think ultimately the--why do we have 100 patients? Because there were a couple sites that liked to use stents. That's what happened. And so despite having it in the protocol, they tried to optimize the results of otherwise tolerable small residual stenoses in the 23-percent range that most of the investigators allowed to have in there, and this was during a very controversial time when in-stent restenosis was the focus of a lot of interest and there were two camps. There were camps out that stented these and camps out there that just did balloon angioplasty. And despite the admonition of the study to try to avoid stents up front, some people's tolerances were different levels. And so the vast majority of these new stents used were for cosmetic improvement of residual stenosis, not for dissections. And we can show that by analysis of what the stenosis was before stent placement use compared to those without, and they were very, very similar. So we have to say that a lot of them were due to just belief patterns of the individuals. What wasn't anticipated was that when we looked at patients who received stents, their result actually did do better than when you don't have a stent used. Therefore, it diminished the amount of effect the radiation therapy had to offer for those cases. However, the use of a stent in a placebo patient still was not as good as radiation therapy overall. So without going through too much overanalysis and getting to multiplicity issues of multiple comparisons, I think it's probably safe to say that we can expect the stent use to be low based on the true dissections that we saw and that the admonition should be to avoid stent use except in the cases of dissection that may be threatening about closure and to try not to use stents for improvement of cosmetic results or any other type of suboptimal result because we have data to suggest that when those suboptimal results exist, radiation therapy is very, very effective. DR. NAJARIAN: One other question, kind of a minor question. But, again, inclusion criteria are arteries 2.7 millimeters to 4 millimeters, and stenosis or degree of stenosis 50 percent or greater. Yet you do this by visual inspection. Was any measurement performed at all, or was any measurement performed subsequently? Were the angiograms reviewed and actually measured? DR. POPMA: This is a point--and I noticed the FDA asked the same question as well, but this has been something from the core laboratory thing we've been struggling with for 20 years. That is, when one compares the clinical site assessment for reference vessel diameter with what we obtain using conventional quantitative angiographic techniques, there is always between a 0.3 and a 0.5 difference in reference vessel diameter. We know that from selecting balloon sizes. We know that from NACEY (ph), from a variety of different comparative analyses where the site has even done calipers on vessels and then compared to the quantitative angiographic result. And so this is typically seen in every interventional trial that we do. One could say that the clinical sites are wrong, but we know actually by intravascular ultrasound that sometimes we underestimate the size of the vessel using conventional angiographic techniques, and that the reasons why are a little bit complex, but, nevertheless, it happens. So most of what we wanted to do for this trial was to let clinicians use the tools for vessel sizing that they were comfortable with, and clinicians are pretty good about picking out a 3-millimeter balloon for a 3-millimeter artery. We know that because when we do quantitative analysis of the balloon, the balloon-artery ratio comes out to be one to one, which means we're measuring the balloon size smaller as well radiographically. So I think that the points are well taken. All of us want to be more quantitative in our delivery systems, and we want to be more quantitative in assessing the vessel size. It's very difficult to institute in-lab, online, quantitative angiography that will really be useful. And I think we did achieve a good result in the study based on the dosing that was done by the investigators at the clinical sites. The balloon-artery ratios were appropriate. The outcomes were appropriate. The balloon sizings were appropriate. So I think that--I understand the point. It's actually one that we've just noted now for 10 years. But I still think that we should be using the visual assessment at the clinical site to determine what our dosage is going to be. DR. NAJARIAN: I understand in clinical practice we can all more or less do a very educated guess. But it seems like when you enter a clinical study, you know, you have to use a measurement. You know, if someone had a 40-percent stenosis and someone had a 60-percent stenosis, different people see that differently. One way you can do it is just use an internal measurement, just use the coronary artery, the normal diameter of the coronary artery as the baseline, and then compare the stenosis to that, which is done in all the carotid trials. I mean-- DR. POPMA: We've done that analysis outside, which you haven't seen the panel pack, but we published it in the American Journal of Cardiology, and that was a comparison of over 800 patients that had clinical side caliper measurements in core laboratory quantitative angiographic measurements, and that's where the 0.3-millimeter difference is coming from that I'm speaking about. So even when you do the caliper--the clinical site caliper measurements and compare then to a separate core laboratory analysis, that discrepancy is absolutely anticipated for a clinical trial. DR. NAJARIAN: I know we all struggle with measurement, and it's taking a living situation and trying to put a number on it. But then, again, that's why we're here, isn't it? MR. GREEN: We also, in designing the protocol, we wanted to make sure we designed it in such a way that they studied it in the way it would be applied when it was actually available to the user. As you said, the visual estimate of the reference vessel diameter was what we used. We were able to come out with a result that we'd hoped for, and now that can actually be applied by a user in a cath lab with the tools they use every day, which is visual estimate. DR. NAJARIAN: Okay. Thank you. ACTING CHAIRPERSON TRACY: Dr. Griem? DR. GRIEM: Mel Griem, University of Chicago. I would like to ask a couple of questions relevant to the transfer device and the response kit. The response kit has equipment to look for the source, to pick it up magnetically. There's a tweezers and so forth. But suppose the loose source is somewhere in the patient. Do you need something like a Geiger counter or something like that to find it, say either on the floor or in the patient? And what do you do when you have this kind of a problem? MR. GREEN: Actually, I think we're going to let Dr. [unintelligible] answer this question. DR. (?) : Mohawn Sinteralangum (?), University of Maryland, radiation oncologist. I was the principal investigator on the START trial. We at the university have done over a hundred cases using this system. I have been paid for my travel expenses to this meeting. In terms of handling worst-case scenarios of a source that has become dislodged from the system within the patient, certainly a Geiger counter would be utilized to identify location of source. But then the more important clinical issue is that if the source is, in fact, in the coronary vessel, then the ischemia that that source might cause is the more pertinent issue that would need to be addressed immediately. So that would be addressed through surgical techniques, again, because this is a beta isotope, in the operating room no special shielding would be required, and you would use just the source container to be able to place the source in a safe place and maintain minimal exposure to patient and personnel. DR. GRIEM: Now, on the transfer device, I'd like to ask a couple of questions. You have the single battery with the low battery voltage problem. You have a pressure monitor with LEDs. Is that voltage sensitive? MR. GREEN: Yes, it is. We actually build in the battery indicator, which is discussed in the panel pack, as a way of ensuring that the items on the transfer device, such as the LED and other items of the sensing system, have adequate power with a safety margin in order to perform a procedure. But as you're pointing out here, yes, it is battery operated, so when you start a procedure, a transfer device would go through a diagnostic. If you have enough battery power to complete a procedure, you'd be allowed to go on. If you did not have enough to complete a minimum of worst-case procedures, you would not be able to go on. So even though they do use the battery power, the system does check itself to make sure that there's adequate power for them to function properly. DR. GRIEM: Why don't you have two battery sets, like on a boat? MR. GREEN: Actually, in the device there are--the battery is composed of two cells. DR. GRIEM: But are they separate? MR. GREEN: There are two cells. I would have to actually bring up--two cells that are actually in connection with one another. DR. GRIEM: So that's one-- MR. GREEN: They're individual cells acting as one power source. DR. GRIEM: Okay. But why don't you have two power sources? MR. GREEN: The power source is simply--the source sensing system and the electronics is because of room considerations within the device, making it able to be used in the cath lab, a small device. We only have the one battery. Therefore, we built in the back-up of the battery indicator that would indicate when, say, you had about ten procedures left that you could not do any more procedures. Therefore, there isn't a need for a back-up battery because if you can begin and initiate a procedure, you can complete a procedure because you'll have enough battery power left. DR. GRIEM: We've had the Indian, Pennsylvania, episode where the source finally wandered around the countryside because the detection system failed for various reasons, and such a thing concerns me here. MR. GREEN: It should be noted also that the electronics does not control the containment of the source within the system. If the electronics were not functioning or if anything were wrong, you could remove the entire system. The sources are contained within the system. So you would not be able to--if you will remove the system, leaving the sources within the patient. DR. GRIEM: At what pressure does the catheter explode? MR. GREEN: We've done that testing, which has been submitted to the Food and Drug Administration, on the catheter. The catheter has a (?) pressure of approximately 430, 440 psi. The transfer device itself includes a pressure limiting pressure regulator that would only allow pressures of between 85 and 100 psi to be generated. DR. GRIEM: As far as radiation effects in blood vessels and connective tissue, I don't have any knowledgeable data in heart. There is some excellent in dermis, and that's published in the British Journal of Radiology Supplement No. 19, some of it with Strontium as planar sources and some with point sources in the dose ranges you're thinking about and the 2-year results of that. There's also an article in there on the rate of fibrosis and collagen formation as a function of dose with gamma radiation between 16 and 22 Gray, and how it proceeds in cyclic fashion, and that may be helpful to you. MR. GREEN: Thank you. ACTING CHAIRPERSON TRACY: Okay. Mr. Dillard? MR. DILLARD: I've got one point. Jim Dillard, FDA. Just in terms of one of the earlier questions about some of the animal data, and I believe that Dr. Waxman got up and at least talked about some information that had to do with different types of sources, and I think we've had some other people get up to perhaps talk about some things that either, number one, may not be exactly pertinent to this particular PMA and/or might be data outside the scope that isn't available in the public literature. So I just wanted to make the point that the information you should be using today is that which is in the PMA and/or available in the public literature that you know about. So just to make sure that what you're doing is factoring in that which is absolutely appropriate for this particular beta source. Thank you. DR. BAILEY: Can I ask a question in that regard? Could I ask if the clinical endpoint results and the MACE results were significant with just the 30-millimeter source data? DR. KUNTZ: We did do that analysis, and they were all positive. I think that--did you say what the TVF was? MR. GREEN: It's Table 3A, page 27, START report. DR. KUNTZ: That analysis doesn't address your question directly; that is, if you take those patients out, is there still significant differences between them? That's obviously a question of power. And so what we did so was analyze whether there was a treatment effect associated with--a main effect associated with the 40-millimeter device, which there wasn't. And there was no interaction between the 40-millimeter radiation assignment. So, indirectly, my guess is that--I don't think we actually did this analysis, but since there were only 13 cases out of the overall group, my guess is that since the treatment effect was identical, all we'll do is just reduce the pool, the sample size by 5 percent, and we'll still have significant differences among the remainder groups of 30 millimeters. DR. BAILEY: I guess the question is: If we're just applying for approval of the 30-millimeter device, shouldn't we be looking at those results? ACTING CHAIRPERSON TRACY: Can we let FDA answer that? MR. DILLARD: Jim Dillard. I don't know if you're looking at me to actually address that question. I think the sponsor's probably more appropriate to do it. But I think that one of the things that we have to look at, we certainly need to look at the overall data set. That's certainly very important in the analysis of safety. And I think in this case the sponsor has put forward an analysis for effectiveness where the predominance of the data is the actual product that they want approval on. So I think one of the things that you might be able to help us with is potentially a recommendation of an analysis. If you think it's important, that could be part of your recommendation to us. ACTING CHAIRPERSON TRACY: Okay. At this point I'd like to take a 5-minute break. We're running a little bit late, so if we could just reconvene in 5 minutes, please? [Recess.] ACTING CHAIRPERSON TRACY: At this point I'd just like to check with the panel members whether they have any burning questions that they are in dire need of asking the sponsor. DR. KRUCOFF: I do. Just relative to the physician from the University of Maryland, again, I just want to make sure I heard this correctly, talking about concerns of a coronary and the ischemia. My understanding from before is that this has not actually happened. DR. : Yes, that's actually a very important point. The question that I was being asked was what if, worst-case scenario. It's important to recognize, I think this system's been used in over 3,500 cases. They have not had one incident of catheter rupture and/or source lost in patient or--it is a closed system that actually prevents that. DR. KRUCOFF: With a blow-out valve. DR. : Right. So I was answering that as a worst-case scenario. DR. KRUCOFF: Thank you. ACTING CHAIRPERSON TRACY: Dr. Simmons? DR. SIMMONS: There's a section in the packet about a stent inside a stent and the radiation delivered to the artery maybe being reduced by as much as 50 percent. Should this be put in as a contraindication? I mean, if you're going to be delivering radioactive brachytherapy to an artery and it's actually going to be reduced by 50 percent, is it even worth doing or should it be done? Or should it be left in longer? Do you have any thoughts before we-- MR. GREEN: Are you referring to the statements made in the FDA review memos by Dr.-- DR. SIMMONS: Mm-hmm. MR. GREEN: --when he was talking about the 20 to 50 percent? When we actually did our studies with our source train, we did some studies on the bench with this. Dr. Crocker at Emory University had done some of these, Tim Fox at Emory University. We've come up with slightly different numbers, but numbers we used to predict our dosing. So for the START trial with respect to a non-stented or a non-in-stent restenosis trial. So that was built--that estimate was built into that. Dr. Crocker talked about that briefly earlier. He can go into some more detail about those numbers. But I think what we'd like to do is in our instructions for use recommend that they do the--they use the doses that were studied in the trial and that they select those doses the way they were selected in the trial, and that should allow them to get the same results that were achieved in the trial. DR. POPMA: The issue about contraindication, I think we should--again, we didn't have these stent sandwiches in our trial, so we didn't--we don't know whether radiation is effective or not effective in that subgroup. DR. SIMMONS: So at the very least, it's a warning or a-- DR. POPMA: At the very least, however ultimately it comes around. But, again, I'm taking my non--I'm not a regulator, but a contraindication for me from a clinical perspective means that we have data saying this might be harmful. And we don't have any data saying that it might be harmful, but I think it's very reasonable to say that we didn't include stents within stents in our clinical trial and that we don't have any data about the efficacy in that subset. But I think that's a well-taken point. We would want to make sure that we emphasize that these are really just single stents in arteries that we're treating, not multiple stents overlapped on one another. DR. CROCKER: And I think it's also important to add that, you know, a stent only covers a very small amount of the luminal surface, and even if you had a stent within a stent, I mean, the chances that they would be covering each other at one point would likely be very small. DR. SIMMONS: Just one other quick thing. In your training overview, I guess it's page 1, there's going to be this checklist, and it implies here that upon receipt of--that the checklist, you know, the cardiologist, the radiation oncologist, the physicist, the staff are all going to be graded on this checklist to make sure that everything is done right. And then on completion of the checklist, they'll basically be certified. I mean, what is your criteria there? Is it going to be 80 percent, 100 percent? I mean, is it going to be--and what if they don't? I mean, what if it is 80 percent? Are you not going to deliver the radiation--I mean, what--I mean, I'm sort of trying to visualize what's going on here. You're going to bring these people there. You're going to train them. You're going to do your mock trial and you're going to come back to the radiation lab. You're going to do a mock thing there. And then you're going to do three cases, and you're going to watch them and fill out this checklist. And if they don't do it right, you're not going to deliver the radiation to them anymore? MR. GREEN: Actually, there's two points here. One, of course, this is the proposed training that we're submitting to the FDA. However, what we're suggesting here is that the checklist is something that would actually be used in every phase of the training to ensure that each one of the following items is covered, in the regional training, in the on-site training, in the mock procedure and the follow-up. So it's not a grade sheet at the end, if you will, to see if they've passed, but it is--the criteria, the things they need to be trained on and show proficiency in, each one of these individual items before being released. If they don't show proficiency in these core items, then they need to be either trained more in-depth on those or there needs to be an understanding why so they can use the device in the safe and effective manner. Does that answer your question? DR. SIMMONS: I think so. I think basically what you're saying is you don't have any control over it, and, you know, you're going to try to provide them with the information. But whether they use it or not is going to be up to them, it sounds like what you're saying. DR. SPEISER: That's true. There's limited control. One is proposing credentialing for the physicians who don't have the present time have staff privileges. And what we're hoping to do is to say that you need to go through this and be checked off before the company will say that you are proficient for the procedure. The question came up what happens if something happens months later, and I don't think that there's any control that you can control an individual months down if they change their pattern. MR. GREEN: What I would say is that we are--we would be going through these things to make sure they're proficient in these. Both the proctors and other experienced users that are part of the training program for the company we would not be releasing from proctoring. We're suggesting three to five proctored cases. If after five proctored cases a team has not showed proficiency in using the device, they would not be released for, you know, solo use of the device or the system. Again, also, even after the proctored devices, we wouldn't just release the device and then assume that down the road that they were using it proficiently. There would be a need for follow-up visits with the site post-sign-off, if you would call it that, to ensure that they are continuing to use and that they are using the device correctly. And, again, this is all things that would be worked out with the FDA. This has been proposed to them, and they're reviewing it as to how best to implement training for this type of system. ACTING CHAIRPERSON TRACY: Does the sponsor have any clarifying comments they want to make before they step back? MR. GREEN: No. ACTING CHAIRPERSON TRACY: If not, okay, then you can step back, please. We'll move on then to the questions for the panel that the FDA has proposed. Okay. The first question that we were asked to consider is: Based on this information, as stated in the preliminary remarks to Question 1, please discuss your recommendations for antiplatelet therapy for patients who receive a new stent and for patients who do not receive a new stent. Any comments from the panel on this? I know it's my own impression that there does seem to be--there is in the labeling a recommendation for patients who do receive a new stent, which seems to be founded on data--perhaps not data from this particular study, but that does seem reasonable to me. However, it seems reasonable to make some statement regarding patients who have not received a new stent, even if it were a statement such that antiplatelet therapy as indicated by usual clinical practice, but some type of statement like that I believe should be added to the labeling, would be my recommendation on that. Any comments from the panel? DR. KRUCOFF: I would agree. I think the data would support patients who do not receive a new stent should be managed per the routine clinical approach of the center performing the procedure. I guess particularly just on what my own feelings would be, I think for patients who do receive a new stent, particularly with less toxic options currently available, a 60- to 90-day or a 90-day use of drug would be a reasonable inclusion precaution. ACTING CHAIRPERSON TRACY: Okay. Anybody else? DR. SIMMONS: Yes, I think, you know, the data in this packet doesn't really support a 90-day, but the data that they've brought in--I've got to trust them that it suggests that it does. I guess I'd go with 90 days for the stents and the usual clinical practice for the non-new stents. ACTING CHAIRPERSON TRACY: Okay. All right. We'll move on to Question 2, then, which pertains to device failures and malfunctions that occurred during this study. Please discuss the clinical importance of the device failure and malfunction events in the evaluation of the safety and effectiveness of the Beta-cath system. Comments? DR. AYERS: I have one. I think the implication is that when the source, other than the minor drift problems, you have problems transporting the source back out or partially into the catheter system through the valve, you have a problem in that you don't know where they're at. Unless they're in the fluoro field or in the safe--in the device, you have no idea here in the vascular the sources are located. And some of the reports we've had have had them free-floating in that basket for upwards of two minutes, and the literature tends to support that radiation doses below the therapeutic amounts can, in fact, induce stenosis. I think it would be difficult to establish that one way or another, but I think there is a clinical potential for some harm. I guess my recommendation would be that that introducer sheath is perhaps the best defense against that and would be an appropriate mechanism to use. ACTING CHAIRPERSON TRACY: Maybe I can--I don't know how this device is used other than what I've been presented here, not being an interventionalist, so I'll turn to an interventionalist to say: Would the problem of delivery be solved by not having a touie but having another type of introducer? DR. KRUCOFF: I almost wonder if there's a more generic way to approach the whole issue. It seems to me that all of the minor delivery issues have approaches either in the training of the user or in engineering or redesign or evolving design, so that it would seem to me that a decrement in the incidence of these events should be documentable. And I know personally to me the eeriest part of all this would be pulling a guide wire at a time that I'm not necessarily sure I'm ready I'd want to pull the guide wire. And ultimately I think the reduction in the percentage, in the rate of these events would be the best agenda from a patient point of view. It seems to me that a lot of these have either ideas or solutions in that a lower incidence than perhaps was seen in the START study could be something we would look for in the evolution of the device and as a result of training. DR. IBBOTT: I'm a bit disappointed that we're not able to review data relating the incidence of the malfunctions with time in the study or from one institution to another or to the training the practitioners received. But having said that, it also appears that the variations in dose received by the target lesion itself are much greater due to eccentricity of the catheter or of the lesion and placement of the sources with respect to the lesion than variations in the dwell time or in drifting of the sources. Still--so I don't--it doesn't seem that the effectiveness measures of the study would be greatly affected by the incidence of the malfunctions. I think they're more likely to be affected by those other issues of centering and so on. ACTING CHAIRPERSON TRACY: So it seems--I'm sorry. More comments down there? DR. GRIEM: The purpose of my query was to look at the transfer device and say what are the possible means of failure, and I tried to identify some of those questions just before the break. And I think that the FDA has sufficient engineering ability and talent here to look at those questions. ACTING CHAIRPERSON TRACY: Okay. So the sense of the panel seems to be that the minor drift does not seem to be a--the minor drift is not a major problem, and that the other engineering issues can be addressed ongoing with the company in monitored--see whether there is a relation to training or experience of the operator. DR. KRUCOFF: I would want to say must be addressed. I mean, I think there are a lot of pointers that they are being addressed. I'd like to see data that they are actually being addressed. I think we have to accept the fact that if this radiation therapy is capable of being effective, that part of its effectiveness is going to be the ability to deliver it accurately and to have it affect the area you're targeting and then get out of Dodge without adversely affecting other areas. So I would want to maybe voice it a little more strongly, Cindy, that it should be either--we should have augmented information over time that these fixes are--that training and that modification of the device and attention to the touie are sufficient to really reduce the incidence of these mal-deliveries relative to the START study. DR. SIMMONS: Those are the kind of things that, you know, could be addressed in a post-market kind of surveillance thing. I guess I'm disappointed--you know, Dr. Zuckerman did an analysis looking at the newest design, and it didn't seem like the lights and the locking stylet and everything else changed the rate of drift or anything else. So I think that's disappointing. So maybe it has to be training. If you're going to do training, I think maybe post-marketing is the way to go. But I liked his idea of making it mandatory that they use the arrow thing. I mean, the whole idea of screwing down on this catheter and putting a kink in it and getting it locked in the--I don't know. It just doesn't seem--it seems like it's a built-in source of error that can be eliminated. I don't know how to put that in words to get it into the protocol other than just to say it out loud, but it certainly seems like that's a very simple fix for-- ACTING CHAIRPERSON TRACY: So the answer, then, is--it will probably come up again a little bit later in the discussion, but there is concern at least over the delivery, not so much the fact that there may be 5 versus 7 seconds' worth of delivery or withdrawal. However, the issue of the adequacy, the technical adequacy of the introduction and withdrawal of the active radiation source. DR. BAILEY: I guess the other point was--I guess I'm still not sure. I thought that data were presented on the effect of the mal-delivery on one of the endpoints, anyway--MACE--which seems to conflict with your answer to your question, which was that they couldn't identify the patients who--they couldn't identify which patients had the mal-deliveries. Otherwise, I mean, if they could identify them, then they can say which order they occurred in. So I'm not sure about that issue. Maybe they can speak to that. But I would like to see the most accurate analysis possible of what the clinical effect was, and I'm not sure we saw that. And I think it's very benign, but I'd just like to see more data on that. Even though they're going to get rid of the problem, I'd like to see what, if any, the problem was that they did encounter. They have lots of power to look at the effect of drift or whatever on the change in minimum luminal diameter in the stented segment. So even post hoc I think that would be useful information for us to have. MR. DILLARD: Jim Dillard. I was just going to make the comment, you know, that the sponsor can certainly get up and address that if they'd like to at this point, too. MR. GREEN: I believe your question, Dr. Bailey, was: Could we identify the patients that had the source drift and source transit? And you are correct. We were able to identify those patients, and that's how the analysis was completed by CDAC, and we came up with that. DR. BAILEY: Does that not mean, then, that you could identify factors such as training, learning curves? MR. GREEN: We do have the details of what happened in the cases where those happened, and that's exactly the experiences we're talking about that we're implementing in our training program. DR. BAILEY: But the response to Dr. Ibbott's question was that you couldn't analyze that. MR. GREEN: If you mean after that event were there any more events like that reported at that site, so did they get better with time and experience-- DR. BAILEY: Right. He was just asking whether factors such as size of center or experience volume, training, had an impact on that. MR. GREEN: That may actually be one way of looking at that. We haven't done that analysis, but that might be something we could look at. ACTING CHAIRPERSON TRACY: Then it may be worth asking the sponsor to look at that and provide that information. Okay. We'll move on to Question 3, and I'll read the background again here in case anybody's forgotten. As demonstrated by the results included in Table 1 of the START clinical report, page 5, the incidence of the primary endpoint, target vessel failure, was significantly lower at 8 months for the treatment arm compared to the placebo. The incidence of target vessel revascularization, target lesion revascularization, and major cardiac adverse events were also significantly lower over the 8-month follow-up period for the treatment arm compared to the placebo. No incidents of stent thrombosis were detected in the treatment arm, and the frequency of total occlusions was comparable between the treatment and placebo arms. The question is: Please discuss whether you believe the probable clinical benefit of the radiation treatment outweighs the probable risks of death, myocardial infarction, total late occlusion, and late stent thrombosis posed by the device in the intended patient population. DR. SIMMONS: Yes. ACTING CHAIRPERSON TRACY: Yes. Any comments, qualifications on that? Yes? DR. WILSON: Frank Wilson. Perhaps that's a qualification. I think we've accepted that it's intuitive that the information in Table 3A doesn't draw any probable distinction between the 30-millimeter and the 40-millimeter source train. But the deal with absolutely accurate data relative to what's in the application for the 30-millimeter, I would probably like to see this table reworked to confirm that that data is what we think it is. ACTING CHAIRPERSON TRACY: I think that's a fair enough addendum to our answer yes, but we'd like to see the information presented specifically with the 30-millimeter. DR. KRUCOFF: I think the other potential addendum to the yes, since the question is posed, Does beta radiation therapy achieve this? where the answer clearly is yes, is that this set of data still leaves the question as to whether the delivery system itself either rubs or irritates the interventional site independent of the delivery of radiation. And I think that awareness of that in this or future device designs, that the device has to be placed into the lumen of a territory that has been designated optimally dilated, and then in 21 percent of cases, by the time the radiation therapy has been administered acutely, another step has to be added, is a potential proviso that would be a reasonable thing to survey over time and with wider experience. So this question as posed for the radiation, which I think the data clearly support, does get a yes, but the delivery device as an intracoronary device I think we still have to recognize was not randomized in this study, and 21 percent of patients, after that device was removed, had additional manipulation of the site. DR. SIMMONS: Well, the other thing was something that you brought up: What happened to the patients who had reintervention after the stenosis? I mean, we don't have any data on those either. I was surprised. Those patients appear not to have been followed up, so the patients who had re-restenosis in the radiation group, did they do as well as the patient population that has restenosis without the radiation? Was there excess bleeding, you know, problems with the surgeons finding adequate sites to reimplant the grafts? We have no data on that. ACTING CHAIRPERSON TRACY: So the answer becomes yes, but-- DR. SIMMONS: I was looking at the acute-- ACTING CHAIRPERSON TRACY: Right. We need the information on the shorter device, and we also recognize that it is not just the radiation but there's the delivery device system. So follow-up on that and understanding of what the mechanics of the delivery system does to the vessel and the late effects of radiation, if there are any late interventions, would be useful information to follow over time. Okay. Question 4 is a multi-part question. One aspect of the pre-market evaluation of new product is the review of its labeling. The labeling must indicate which patients are appropriate for treatment, identify the product's potential adverse events, and explain how the product should be used to maximize benefits and minimize adverse effects. Please address the following questions: 4A, Please comment on the indications for use section page 12 as to whether it identifies the appropriate patient population for treatment with the device. And probably we should turn to that. DR. KRUCOFF: Page 12? DR. SIMMONS: Page 12, Section 4. DR. KRUCOFF: I think one of the things that needs to be emphasized in the indications is that the patients tested in this study had successful angioplasty results or successful percutaneous intervention results. And I think the way it's currently worded, patients who have undergone PTCA for discrete lesions, is far less specific to a host of operators than the patients who were, in fact, included in the START study. In fact, as was stated earlier, we don't know how many patients came into the labs with in-stent restenosis who were not included in the study. Roughly, if that's 10 percent or so by the operators' estimates, I think that that's an important element to the indications for use, that these were patients who had a successful angioplasty result, who then underwent beta radiation. DR. SIMMONS: Let me put my foot in there. Successful PTCA with a lesion less than 30 percent? Thirty percent or less? Residual lesion 30 percent or less, something like that. ACTING CHAIRPERSON TRACY: I think that would be what was included in the study, so that seems to be reasonable to include it in the indications here. MR. DILLARD: Jim Dillard. Just an additional question. One thing that we struggle with frequently is how much actually had to do with entry criteria or inclusion criteria in the study, how much of that needs to go into specifically defining the intended use or indication for use versus how much of that should go in the clinical section of the labeling that actually gives us the descriptive characteristics of the clinical study and the outcomes. And so one of the things if you're thinking about where to put it or where your recommendation might be, should it actually be in the indication for use or should it just be a very prominent section contained within the clinical section of the labeling, if you can make some differentiation between the two of those, because what I don't like to see in the direction we're trying to go at the agency is try not to put so much detail in the indication for use statement that you as the clinician stop reading them because they become three pages long and then the important clinical section becomes very tiny. So I know it's a trade-off, but, you know, any recommendation you might have on that would be appreciated. ACTING CHAIRPERSON TRACY: It may be worth being a little bit less specific, then, and perhaps just saying successful rather than specifying a degree of stenosis since different labs have different abilities or technologies for measuring the lumen and the residual stenosis. But this study really did pertain to those patients who had had a successful in-stent restenosis. So I would say maybe just the word successful. DR. BAILEY: Should atherectomy be included in this, or is that implied? ACTING CHAIRPERSON TRACY: I would think it would--there were a variety of different techniques by which the vessel was opened. Whether it was atherectomy or dilatation or--I don't think we should specify--I would think we shouldn't specify that. DR. KRUCOFF: Actually, this is a small detail, but I guess small can be important. Rather than PTCA, it might be worth just substituting PCI as an indication, that the patients who were enrolled in this study essentially had a reduction of their lesion based on operator discretion, selection of tools, not just balloons. DR. GRIEM: But excluding new stents. DR. KRUCOFF: But excluding stents. That's true. ACTING CHAIRPERSON TRACY: Okay. So we're going to change the word from PTCA to PCI and add the word successful. DR. KRUCOFF: You might want to add excluding stents. That's a good point. ACTING CHAIRPERSON TRACY: Okay. Well, I'm not--maybe we can move on to 4B. Somewhere we've got to deal with the stents, but 4B is: Please comment on the contraindications section, page 12, as to whether it identifies all conditions under which the device should not be used because the risk of use clearly outweighs any possible benefit. And the contraindications, as stated on page 12, say unprotected left main disease and patients in whom antiplatelet and/or anticoagulant therapy are contraindicated. Any comments from the panel? DR. AYERS: Bob Ayers. I noticed one thing that crops up here is that nowhere indicated--and it could bridge contraindications and warnings and precautions. There's nothing there about use that hasn't been established in the trial and that a priori would probably be considered a contraindication or potentially harmful, such as treating across a bifurcation or trying to step a source like this. In other words, trying to use a 30-millimeter source train to treat 60 millimeters of artery, which the system's not designed really to do. MR. DILLARD: I'll make a comment. Jim Dillard. Without going into a lot of detail about where we currently view precautions, warnings, and contraindications, let me just give you a little bit of a sense. Contraindication I think in our mind would be that where we have got some negative data, some data that we would know would be adverse to a patient or a patient group. I think that would be very important and that should drive a contraindication. I think generally--and perhaps it's implied, although many times it ends up in a warning about any other patient population or any other subgroup that was not studied in the trial. Of course, there's no safety and effectiveness information known on those patients. Many times statements like that show up. They don't show up generally in contraindications, but a lot of times they would show up in a precaution or a warning type of statement. So I think we're trying to clarify that and not make contraindications when a contraindication is not warranted, but nonetheless, if it's something that might have either an effect on the device or an effect on a patient population where we don't know very much, then that should drive either a precaution or a warning, respectively. DR. SIMMONS: What about a special consideration? I don't think I've heard that term used before. Where did that one come from? MR. DILLARD: Jim Dillard again. I think there can be special considerations, there can be notes, there can be a lot of other sections to a manual, and I think the important thing, at least--you know, maybe we can put this to bed a little bit--is that if your recommendation is that there should be some consideration in the labeling about a certain type of issue or a certain type of item, I think we can work very closely with the sponsor about where to appropriately put it. I think if you can give us a recommendation about how strongly you think it's needed, that will be, you know, very beneficial to us. ACTING CHAIRPERSON TRACY: I think as regards that comment, the issue of the stent perhaps comes up somewhere where you can decide with the sponsor where it should be mentioned, that this study was not specifically meant to deal with radiation therapy for stents placed within stents for restenosis. And the other section that I think that we should think about are the whole group of exclusion criteria for the investigational protocol that did not make it up either as contraindications or warnings and precautions. And I wouldn't think they should be contraindications, but I think it was Dr. Simmons who had pointed it out earlier that certain patient characteristics or ejection fractions were not included in this patient population. That should, I would think, be stated somewhere, called something. DR. SIMMONS: Yeah. How about, you know, the MI less than or equal to 72 hours prior to the procedure? There may not be specific data on that, but it would--you know, certainly that artery is probably not the same as an artery that's, you know, stable, anginal-type symptoms and maybe that should go under precaution just--there may not be negative data, but there's certainly no data and there's certainly implied data, that maybe it isn't a normal artery. ACTING CHAIRPERSON TRACY: Yeah, I think that there is a gradation, it seems, within the exclusion criteria, some of which would probably rise to a precaution level. And I think the infarct is a good example of that. DR. KRUCOFF: Yeah, Mitch Krucoff. I'd sort of extend the list not, again, so much as a contraindication as to helping interventionalists understand where we just don't have data to probably include thrombotic lesions, vein graft segments, stent sandwiches, and at least to the best of my appreciation of the START data, to diffuse proliferative responses that extend beyond the margins of previously implanted stents. ACTING CHAIRPERSON TRACY: Do you think that--are you proposing those as precautions or warnings, or do you think those are just indications of what was not studied? DR. KRUCOFF: Well, I'm actually not sure what the right answer is. I think being cautious in an arena where we don't have data and where we have potentially different physiologic substrates about the use of radiation technology is one of the reasons that I would impute this START study was designed to look at a fairly pristine patient population, and we have data from a fairly pristine patient population, and we have a community of knuckle-dragging interventionalists who will probably run all over that outside of that pristine patient population. So anything we can do to assist them to stay within the boundaries of what we know and have learned from these data I would encourage be in the package labeling. DR. SIMMONS: So maybe under precautions? ACTING CHAIRPERSON TRACY: I'm sorry? DR. SIMMONS: Maybe under precautions rather than warnings kind of thing. ACTING CHAIRPERSON TRACY: Maybe we'll leave the semantics to the FDA. DR. SIMMONS: How about these? Could we look at these? As long as we're looking at--we sort of lump these all together. ACTING CHAIRPERSON TRACY: Which one is this? DR. SIMMONS: What about these special considerations? Do you want to read through those and see what you want to do? I have some problems with those. I don't know what to do with them. Like these vessels or lesions that would preclude revascularization or placement of the Beta-cath delivery system, it's kind of obtuse to me. I don't understand what it's trying to say or what it's a special consideration for. ACTING CHAIRPERSON TRACY: That is a little--I would--I took that as meaning an unsuccessful dilation or an unsuccessful atherectomy. DR. SIMMONS: I don't know... ACTING CHAIRPERSON TRACY: I think we've moved that into the indication--I think, if I'm reading that correctly, that they're talking there about an unsuccessful intervention. DR. SIMMONS: Or a tortuous vessel or a vessel that's too small to get the catheter into or something-- ACTING CHAIRPERSON TRACY: It could be, but they have specifically indicated that the study involved 2.7- to 4-millimeter vessels, so we shouldn't even be talking about those other types of vessels I would think. DR. SIMMONS: So I think that's vague and-- ACTING CHAIRPERSON TRACY: So maybe it's too vague, just leave it out. DR. SIMMONS: How about patients having undergone prior chest radiotherapy? What does that have to do with anything? I mean, maybe one of the radiation oncologists could help us here. If you've had previous radiation therapy, does that preclude you from getting a successful result with this radiation therapy? Or are they worried about cumulative radiation dosages? I don't understand that one. DR. WILSON: I suppose that was the concern. I read that several times and decided it was acceptable where it is. It's a special consideration, should be given all due weight in the thinking of the user before proceeding. But I don't have a strong feeling about it. I doubt this is ever going to be a clinical problem in individual cases or in groups of patients. But it's something to weigh. DR. SIMMONS: And how about the other one, the women of child-bearing potential? Is that just a special consideration? Is that a warning, a precaution? Is it really important if they're having coronary anatomy that needs to be opened up whether they're of child-bearing potential or not? ACTING CHAIRPERSON TRACY: I think it is a special consideration in that you're exposing that woman and that child to radiation of one form or another, whether it's through fluoroscopy and whatever limited radiation at a distance you get from the beta source. DR. SIMMONS: So just leave it in this sort of vague special considerations category rather than a warning or a precaution. ACTING CHAIRPERSON TRACY: I would think so. DR. WILSON: I looked at that also, and I wondered about that. It seemed to me that it's all right where it is. Obviously, if you have a woman who is in the first trimester, that is, during the period of organogenesis, you're going to be very concerned about the dose from fluoroscopy in that patient. But it seems to me it's still something you're going to take under advisement rather-- DR. SIMMONS: Yeah. DR. WILSON: --than say it's an absolute contraindication towards. DR. AYERS: Bob Ayers. It kind of supports, goes along with our regulations. We have no prohibition whatsoever as to the practice of medicine of giving radiation to a fetus, but we do have a prohibition against unintentionally giving radiation to a fetus. So if the physician doesn't determine the patient is pregnant and then goes ahead and gives an excessive dose of radiation to the fetus, that's a problem. But if he knows the patient's pregnant and decides it's still in the best interest of the patient and goes ahead, that's okay. So it's unintentional radiation exposures to a fetus that are a problem for us. ACTING CHAIRPERSON TRACY: Okay. So maybe that--under their label section special considerations we would suggest removing the second one, which seems to us to be another statement of unsuccessful or inappropriate, so we'd take that out. And the other comments that we made regarding the inclusion, specific inclusion, and the exclusion criteria from the study should be reflected somewhere, probably for the most part under either an additional statement of inclusion/exclusion criteria of the study or precaution for the particular of acute infarct vessel. DR. KRUCOFF: Cindy, Mitch Krucoff. I guess I should speak up since to me this second one actually does make sense. I should at least voice it. Lesions that preclude revascularization to me is a way of saying to the interventional community this gadget is not magic; you know, that basically it is indicated for use once you've achieved a successful angioplasty result in a vessel that can be revascularized. And maybe more importantly, the second part that precludes placement is a reminder that this is not a supple balloon catheter design. This is a hydraulically protected closed system of tubing that is going to be a little more stiff and potentially ramming it into a coronary anatomic location that's obviously tortuous or heavily calcified would be something you'd be worth taking a special consideration about delivering this device compared to other devices. I would take that as at least the spirit of why it's there and why it's sort of isolated in this section as a way of giving interventionalists an alert that this is not a sleek, supple balloon catheter, this is a larger plastic system that you better think about, and you might not want to slide in through a tiny guide catheter, you might want to use a larger guide catheter, et cetera, et cetera. ACTING CHAIRPERSON TRACY: Is that, then, a contraindication or special consideration? DR. KRUCOFF: Well, for me, I would leave it as a special--it might, for instance, influence me in the selection of the guide catheter that I would use to make sure I had more support or that I might take more time if I had a more moderate turn in a coronary to make with this thing, that I might be a little more aggressive in placing a guide catheter or other sort of technical elements like that. ACTING CHAIRPERSON TRACY: Okay. DR. SIMMONS: Make sense to me. It's great. ACTING CHAIRPERSON TRACY: It's the anti-- DR. SIMMONS: It just left me cold. I read it and I went blank. ACTING CHAIRPERSON TRACY: Okay. DR. NAJARIAN: Just as far as leaving it in, I agree with that. But you could add another word like morphology or anatomy, you know, vessel or lesion morphology and/or anatomy, and that kind--I think that's what you're trying to say. Just by saying vessels or lesion, I mean, it's kind of vague. ACTING CHAIRPERSON TRACY: Okay. So we'll ask that they add morphology or anatomy in there. DR. NAJARIAN: How's that? ACTING CHAIRPERSON TRACY: Sounds good. I think we've already talked about 4C, which was: Please comment on the warning and precautions section as to whether it identifies all potential hazards regarding device use. Any additional comments regarding that? My only comment, just from the standpoint of finding them, it's just a little bit difficult. It might be nice if they--it's fine that they're interspersed throughout the text, but it might be nice to have it put together in one particular--one concise table so that it would be easy to look at it rather than having to kind of go through the whole section. DR. KRUCOFF: Mitch Krucoff. I don't know if this really belongs in the warning section, but to me one of the biggest print items I'd like to see in this package is a message to the users on the importance of your whole team and having the team together and sustaining that team. And we'll come back to this, I think, when we talk about training. But one of my biggest concerns is that as this device becomes more routine in a cath lab, the tendency for us to move cases along and to orchestrate the team work that really is implied in the safe and effective use of this is going to be lost to the larger agenda of moving faster and in general use. So I don't know if warnings and precautions is the place exactly to say it, but to me a warning and precaution that you really need every key member of the team to safely and effectively apply this device in patients is implied in the data and the experience to date. ACTING CHAIRPERSON TRACY: Any other comments? [No response.] ACTING CHAIRPERSON TRACY: Okay. Question 4D: Please discuss whether any improvements could be made to the labeling to help minimize the occurrence of device failures and malfunctions as discussed under Question 2. Maybe that's where some statement being made about the importance of maintaining a team approach to this could be made. DR. AYERS: And this is probably the place to talk about that sheath introducer, too. ACTING CHAIRPERSON TRACY: Okay. Any additional comments on 4D? [No response.] ACTING CHAIRPERSON TRACY: If not, 4E: Please comment on the remainder of the device labeling as to whether it adequately describes how the device should be used to maximize benefits and minimize adverse events. DR. KRUCOFF: Mitch Krucoff. I guess my one concern here is still whether the emergency bail-out procedure is optimally done pulling the guide wire or whether the actual time and exposure involved to--even if you're not sure where the source train is within the system, to backing the system out over the wire, which is something that would take less than a second to do versus something that would probably in skilled hands take maybe 10 seconds to do, but in less than a second you lose the lumen of the vessel by pulling out the guide wire, where in 5 to 10 seconds you take the train out and put it in the box but leave the guide wire. I'd like to hear from--actually, if it's appropriate, from some of the people who have used this what they really think. Is the optimal bail-out procedure so urgent that it's worth pulling the wire? DR. : Again, I'm Mohawn (?) from University of Maryland. It's been our experience--again, I think if you look in the data, I think there were six circumstances that the manual removal procedure was initiated. In our clinical experience at the university, I would agree with you that time difference in skilled hands is really not significant, and I think leaving the wire down clinically, at least in our cardiologists' hands, that's much more important of an issue than the actual 10 seconds that it might take to keep it down. So we keep the wire down, but, again, I think that has to be a clinical determination. I don't think that the dosing is so significant, that difference, that we need to make a recommendation that it always has to be pulled. ACTING CHAIRPERSON TRACY: Dr. Popma? DR. POPMA: Jeff Popma. Just in the lab, these are in-stent restenosis lesions that you're dilating, and recrossing the lesion is not usually a big deal. And I think as we're talking about the importance of getting this out and getting it in a contained space as quickly as possible, the fact that there may be some delay in doing an over-the-wire exchange technique and at least given the perception that all this period of time we have no idea where the seeds are is the reason that we said to go ahead and pull everything out. Basically, this is an in-stent restenosis trial, and the patient's had a successful treatment, and to rewire the lesion isn't that big a deal. But I understand if any individual operator has really had difficulty wiring the lesion and it's been very difficult to do, that's a clinical decision they're going to have to make because they may spend another 30 minutes getting it back across the lesion. But the majority of cases, it's more important, I think, to get the device into a box than it is to worry about recrossing the lesion. ACTING CHAIRPERSON TRACY: Maybe I could just summarize to say that there seems to be a divergence of opinion whether it's better to leave the wire or pull the wire, but that that, as an electrophysiologist's point of view, would be an individual experienced operator's choice probably should not be mandated in the labeling for bail-out. I wouldn't think that that would be something you would want to mandate. DR. KRUCOFF: Well, it's already pretty explicitly mandated in what's here. ACTING CHAIRPERSON TRACY: Can you point--I'm having a hard time finding what you're looking at. [Pause.] DR. KRUCOFF: Page 42 of the IFU. It's basically lines 12 through about, whatever, 24. ACTING CHAIRPERSON TRACY: If the source does not return to the active transfer device, is that where-- DR. KRUCOFF: Correct. Loosen the hemostatic use for more saline, remove the Beta-cath and guide wire from the patient. ACTING CHAIRPERSON TRACY: All right. Is that--this looks like what they're discussing here, the active source is considered to be lodged in the delivery catheter. DR. KRUCOFF: That's what bail-out's basically for. DR. KRUCOFF: I can take Dr. Popma's point, though. I think that, again, in this--as a well-defined patient population who have in-stent restenosis that's been dilated with a good result before you engage in a radiation direction, that the recrossing with a wire is probably less of an issue than an whole general population, while the indeterm