Document Yjo6JMno1aVbKYyDnwYb7a5D

101 1 buying a pig in a poke, you have to believe in the basic 2 integrity of the thing and in the observations and all kinds 3 of things, which you never do get to check, which is sort of 4 why the robustness of the findings and anything that can 5 make it seem less dependent on day-to-day differences adds 6 to the persuasiveness for me, and the things you cited, the 7 evidence of control group, the knowledge that there is 8 quality control in the system that generated the data, all 9 those things certainly help us. 10 CHAIRMAN BUNN: For the consumers on the 11 Committee, one of the things when the FDA approves an agent, 12 it allbws a manufacturer to market or advertise, you know, 13 their product, whereas, they are not allowed to do that 14 without the FDA approval. 15 Most of the time when a company comes in with an 16 extra indication, it is because they feel it is important to 17 advertise or market, and that has been the case in the past. 18 Now we have these new things where companies are just 19 denying drugs, saying, well, the FDA hasn't approved it, 20 therefore, we are not going to pay for it, and that is an 21 unfortunate thing that I think consumers and physicians, 22 even the FDA, as we have just heard, is really not in favor 23 of that type of approach to medical care. 24 DR. DELAP: I would just reiterate that there is 25 no way that we can fully keep up with the medical practice MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 .CZ02)-546-6666 ` 102 1 and that it is really unfortunate when insurance companies 2 make decisions based on FDA approvals, which are always 3 going to be -- despite our best efforts -- are always going 4 to be a little bit behind the cutting edge of clinical 5 science. 6 So I think we should represent ourselves as trying 7 to keep the labeling up to date for the use of the insurance 8 companies because I don't really see that as an appropriate 9 use of labeling to make a decision about payment. 10 CHAIRMAN BUNN: But for those in the audience, I 11 think we all agree that this was an appropriate use of . 12 literature-based and important for the FDA, important for 13 physicians, and important for consumers, and this is an 14 appropriate way to do things. Hopefully, it won't be 15 extended to how insurers should pay for drugs. 16 That being the case, we will take a break for 17 lunch until 12:15. 18 [Whereupon, at 11:15 a.m., the proceedings were 19 recessed, to be resumed at 12:15 p.m.] 20 MILLER REPORTING COMPANY, INC 507 C Street, N . E . Washington, D.C. 20002 (202) 546-6666 103 1 A F TE R N O O N P R O C E E D I N G S 2 [12:30 p.m.] 3 DR. DEVOUS: Let's start to introductions to 4 begin. Steve Pollitt, actually, needs to talk about a 5 voting conflict. 6 MR. POLLITT: Dr. James Ingle has recused himself 7 from voting on this issue after finding out that his 8 institution is doing a study on NR-LU-10. He will be 9 allowed to participate in discussion but will not vote on 10 the issues at hand. 11 Dr. Devous? 12 DR. DEVOUS: On my right, I want to introduce 13 Barbara Sortino who is here as a Patient Representative and 14 is a non-voting member in that regard. Perhaps the best 15 thing for us to do to geo on firm footing first is to just 16 do some personal introductions. 17 George, Could you start and we will just go around 18 the table. 19 DR. MILLS: I am Dr. George Mills from the Center 20 for Biologies. 21 MR. BERKOWER: Ira Berkower from the Center for 22 Biologies. 23 DR. BOTSTEIN: Paula Botstein, the Center for Drug 24 Evaluation and the Office of Drug Evaluation III. One of 25 our Divisions is the Medical Imaging Division. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 at 1 DR. JAY SIEGEL: Jay Siegel, Office of 104 2 Therapeutics in the Center for Biologies. 3 DR. KEEGAN: Patricia Keegan, Center for 4 Biologies. 5 DR. GELBER: I'm Rich Gelber a biostatistian on 6 the ODAC Committee from Boston. 7 DR. FORASTIERE: Arlene Forastiere, medical 8 oncologist from Johns Hopkins. 9 DR. INGLE: James Ingle, Professor of Oncology, 10 Mayo Clinic. 11 DR. OCHS: Judy Ochs, pediatric oncology, Arkansas 12 OD A C . 13 MR. POLLITT: Steven Pollitt. I am the Executive 14 Secretary for the Committee. 15 DR. SWAIN: Sandra Swain, medical oncology, 16 Bethesda. 17 DR. OZOLS: Bob Ozols, Fox Chase, medical 18 oncology, ODAC. 19 DR. KROOK: James Krook, Duluth, medical oncology, 20 OD A C . 21 MS. BEAMAN: Carolyn Beaman, science educator, 22 Consumer Representative, ODAC. 23 DR. DEVOUS: We have introduced Barbara. I am 24 Mike Devous, University of Texas, Southwestern Medical 25 School, Dallas. I am the Chairman of MIDAC and, for five or MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6656 105 1 ten minutes or so, the Acting Chairman here until Dr. 2 Parkinson arrives. 3 DR. MARILYN SIECEL: Marilyn Siegel, pediatric 4 radiologist, Mallinckrodt Institute of Radiology, St. Louis. 5 DR. ABBITT: Pat Abbitt, radiologist, University 6 of Florida in Gainesville. 7 DR. MALCOLM: Arnold Malcolm, member of MIDAC, a 8 radiation oncologist. 9 DR. BARRY SIEGEL: Barry Siegel, nuclear medicine, 10 from Washington University. I am a consultant to MIDAC. 11 DR. DEVOUS: Thank you. I have seen the agenda! 12 for about 30 seconds, so I will struggle through with this. 13 If I make any gross mistakes, somebody please raise your 14 hand and jump in there. 15 This is the post-lunch session. My understanding 16 is the first order of business is George's presentation; is 17 that correct? The sponsor presentation? Oh; I'm sorry. It 18 does say that. 19 MR. POLLITT: As people come to the podium, would 20 you please state your name and your affiliation. 21 PLA 94-1308, NR-LU-10 22 KARL THOMAE PRESENTATION 23 DR. DAVID BRILL 24 DR. DAVID BRILL: My name is David Brill. I am 25 the Director Drug Regulatory Affairs at the sponsor, MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 f9n?v cac-fissp 106 1 Boehringer Ingelheim. I would like to say,- to start, good 2 afternoon and an early "Happy holidays" to the members of 3 ODA.C, the consulting members from MIDAC, the representatives 4 from the Food and Drug Administration and all the attending 5 industry colleagues that are sitting in the audience today. 6 As I said, my name is David Brill and I am 7 representing the sponsor, Boehringer Ingelheim. That is 8 hard to say. It shouldn't be. It is my pleasure and true 9 privilege to be introducing and opening today's Advisory 10 Committee meeting discussion on Verluma. 11 [Slide.] : 12 Verluma is a kit for the preparation of Technetium 13 99 nofetumomab merpentan. I have to tell you that is not 14 easy for me to say, as well. Verluma, which you may have 15 seen or heard about under a different name such as Oncotrac 16 or NR-LU-10 or, as we finally refer it to as "narlu-10," is 17 a product opportunity which has come to be based on a very 18 extensive collaboration involving many organizations. 19 [Slide.] 20 Verluma was developed by the NeoRx Corporation in 21 Seattle, Washington. The manufacturer of Verluma is Dr. 22 Karl Thomae GmbH which is an affiliated company of 23 Boehringer Ingelheim International GmbH. I should say that 24 it is Dr. Karl Thomae who is the official sponsor of the 25 pending PLA and will be the holder of the license following 1 MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546 - 6 6 6 6 107 1 approval of that application. 2 In North America, Verluma will be distributed by 3 the Radiopharmaceutical Division of the DuPont Merck 4 Pharmaceutical Company in Billerica, Massachusetts. 5 [Slide.] 6 The development of Verluma, as you can see, was a 7 collaborative effort. So, too, is today's Advisory 8 Committee presentation. I would like to just very briefly 9 introduce some of our participants. 10 Vie have already met. Dr. Darrell Salk is 11 currently with the consulting firm Biotech Med/Reg. 12 However, during the majority of the development of Verluma, 13 he served as a Vice-President of Medical and Regulatory 14 Affairs an NeoRx Corporation. 15 Dr. Wil Nelp of the University of Washington 16 Medical Center is a nuclear-medicine physician who served on 17 the Verluma development team since its inception. Dr. Nelp 18 participated in the Phase I and II clinical trials. He 19 assisted in the development of the Phase-III protocol design 20 and also served as a blinded reviewer during that Phase-III 21 study. 22 Dr. Andrew Turrisi of the Medical University of 23 South Carolina is an oncologist and brings to the Verluma 24 Program that critical perspective of how, where and under 25 what conditions fits in that clinical community. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 108 1 We have additional team members here that will be 2 available to answer some questions as deemed necessary. 3 They are Dr. Paul Abrams who is the President of the NeoRx 4 Corporation; Dr. Hermann Allgaier, who will serve as the 5 responsible Head in Dr. Karl Thomae, GmbH; Dr. Michael Bjorn 6 who is the Director of Clinical Research at the NeoRx 7 Corporation; Dr. Brent Blumenstein of the Fred Hutchinson 8 Cancer Center who is our consulting biostatistician; and Dr. 9 Alan Fritzberg of the NeoRx Corporation who designed the 10 Technetium labeling technology. 11 Let me move, then, very quickly into the agenda, 12 and get us started. 13 [Slide.] 14 Dr. Darrell Salk will give a clinical introduction 15 and profile of Verluma and give a brief background on small16 cell lung cancer. Dr. Wil Nelp will present some images 17 which might typically be seen using the product and, also, 18 go through the image-review procedure and other product 19 performance characteristics. 20 Dr. Salk will present the pivotal-study results 21 and will focus in on what are the key results and 22 conclusions from that Phase III study. Dr. Andrew Turrisi, 23 as I mentioned before, will present the critical 24 oncologist's perspective and will really focus in on issues 25 related to clinical benefit and relevance of Verluma. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (2.02) 546-6656 109 1 Dr. Darrell Salk will then return and serve as 2 moderator for the remaining section of this discussion and 3 will take measures to insure that all your questions are 4 adequately answered. 5 So, having adequately, I hope, set the stage for 6 the rest of the presentation, it is my pleasure to introduce 7 Dr. Darrell Salk, and I would like to turn over the 8 microphone and the presentation to him. 9 Dr. Salk? 10 DR. DARRELL SALK 11 DR. SALK: Good afternoon. Thank you, Dr. Brill. 12 My name is Darrell Salk. I was, as Dr. Brill said, formerly 13 Vice-President of Medical and Regulatory Affairs at NeoRx 14 Corporation and, from 1987 to 1994, my primary 15 responsibility was in the Phase-III clinical development of 16 Verluma. Since 1994, I have continued my involvement in the 17 Verluma project as a consultant to NeoRx. I currently 18 consult through Biotech Med/Reg in Seattle and I am an 19 Affiliate Associate Professor of Pediatrics and Pathology at 20 the University of Washington. 21 Today, we want to describe for you a product that 22 we have developed for staging patients with small-cell lung 23 cancer in a non-invasive way. We are very confident of the 24 safety and efficacy of this product. 25 We have focussed initially on small-cell lung MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 . 110 1 cancer in the clinical development of Verluma because 2 accurate staging is so important in the treatment of this 3 disease. The antibody that we chose to radiolabel and to 4 use for imaging small-cell lung cancer was actually derived 5 originally using small-cell lung-cancer tumor cells. It is 6 an FAB fragment of an IgG 2B murine monoclonal antibody that 7 is specific for a 40 kiloDalton glycoprotein expressed on 8 the surface of small-cell lung-cancer cells as well as the 9 surface of many other epithelial cancers. 10 [Slide.] 11 This is a molecular model of an FAB fragment. ; 12 This is where the antibody attaches to the antigen. We 13 developed a labeling method that tightly binds Technetium 14 99m in a defined and stable way that does not affect the 15 immunoreactivity of the molecule. /. , 16 As you know, Technetium 99m is a radionuclide that 17 is routinely used in clinical nuclear medicine. We show 18 here, at the same scale, a model of the Technetium chelate 19 at an exposed lysine amine of the protein. The structure of 20 the Technetium N2S2 chelate is shown here with the side 21 chain for formation of an amide linkage with the FDA 22 fragment. 23 This chemistry results in an attachment of a 24 stable- metal Technetium chelate using a defined covalent 25 linkage to the protein. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 111 1 Today, we want to summarize for you the key points 2 from our work that demonstrate the usefulness of this 3 diagnostic agent and support the following indication. 4 [Slide.] 5 Verluma imaging is indicated for initial staging 6 of patients with small-cell lung cancer. It establishes a 7 diagnosis of extensive disease in 77 percent of patients 8 with extensive disease -- in other words, it has a 9 sensitivity of 77 percent -- with a positive predictive 10 value of 94 percent which is very high. For these patients, 11 no further staging tests-are needed before moving on to : 12 therapy. 13 [Slide.] 14 If there is no evidence of extensive data by 15 Verluma imaging, then additional tests can be done to 16 accurately establish the diagnosis of limited disease which 17 is a diagnosis of exclusion. Of course, for a diagnosis of 18 exclusion, the more tests that are done that are 19 complementary to each other and remain negative, the greater 20 the accuracy of that diagnosis. 21 [Slide.] 22 Verluma imaging should not be used for 23 differential diagnosis of small-cell lung cancer because it 24 also localizes to other carcinomas. This precaution is 25 important because Verluma images must be interpreted with MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6665 112 1 this cross-reactivity in mind just as radiologic images and 2 those from other diagnostic imaging procedures must be 3 interpreted, recognizing that the histology of any observed 4 abnormality is presumed rather than defined. 5 [Slide.] 6 Like all Technetium-labeled nuclearmedicine 7 imaging agents, Verluma is provided as a cold kit with all 8 of the components necessary for the addition of Technetium 9 99m in the radiopharmacy at the clinical site. 10 By the way, the name NR-LU-10 stands for NeoRx 11 Lung Antibody No. 10. Despite rumors to the contrary, there 12 is no mythological or astrological significance to that name 13 and Elvis was not involved in creating it. 14 [Slide.] 15 Verluma imaging is performed as an outpatient 16 procedure with administration to the patient in the 17 afternoon and then routine gamma-camera imaging 14 to 17 18 hours later the following morning. Our safety experience 19 with Verluma is actually quite extensive. It has been used 20 in many Phase.1-11 studies with a variety of different tumor 21 types. 22 Among 515 patients, Verluma has been shown to be 23 very safe with only a very small percentage of patients 24 showing laboratory changes without any clinical 25 significance. Other clinical events have been very rare and MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 113 1 insignificant. 2 [Slide.] 3 53 patients in the pivotal study had blood 4 collected at baseline, 6 weeks and 9 weeks after imaging 5 and, thus, were evaluable for the development of HAMA - 6 that is, human anti-mouse antibodies which may occur in 7 response to administration of a murine protein. 8 As it turns out, 50 of 53 small-cell lung-cancer 9 patients had no evidence of HAMA response when tested with 10 an antibody-specific and very sensitive ELISA. Only three 11 patients had low-level responses that were transient. 12 [Slide.] 13 A small-cell lung cancer represents 20 to 14 25 percent of all lung cancers. There are about 40,000 new 15 cases in the United States every year. It is an aggressive 16 disease and, if not treated, survival is only a few months. 17 Chemotherapy is the mainstay of treatment in this disease 18 with a very high response rate. However, staging, accurate 19 staging, is critical for determining prognosis and the 20 choice of therapy. 21 [Slide.] 22 Small-cell lung cancer is staged as limited 23 disease or extensive disease based on the distribution of 24 the lesions and diseased organs. A limited disease is that 25 in which the disease is within a single radiation port, one MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 54 6 - 6 6 6 6 r 114 1 hemithorax, the mediastinum and ipsilateral supraclavicular 2 lymph nodes. 3 Extensive disease, on the other hand, is disease 4 when there is any metastasis outside of this limited region. 5 [Slide.] 6 For patients with limited disease, the intent of 7 treatment is curative using chest radiation therapy within 8 that radiology port in addition to combination chemotherapy. 9 Patients with extensive disease, on the other hand, the 10 intent of treatment is palliative using chemotherapy alone. 11 ' The reason for this difference in approach to 12 treatment is because patients with limited disease who are 13 treated with chest radiation therapy in addition to 14 chemotherapy demonstrate an increased long-term survival 15 compared to patients who are treated with chemotherapy 16 alone. 17 On the other hand, patients with extensive disease 18 gained no benefit from the additional chest radiation 19 therapy and its attendant cost, acute toxicity and potential 20 morbidity. 21 [Slide.] 22 Because of the importance of accurate staging in 23 small-cell lung cancer, current staging procedures that are 24 necessary include four diagnostic-imaging tests and an 25 invasive procedure evaluation of the bone marrow. MILLER REPORTING COMPANY, INC. 507 C Street, N . E . Washington, D.C. 20002 )1707 546-SS66 115 1 [Slide.] .. 2 The objective of our pivotal study was to evaluate 3 the ability of Verluma imaging to stage patients with small- 4 cell lung cancer in comparison with current staging 5 modalities. 6 Before discussing the pivotal trial and its 7 results, Dr. Nelp will describe the characteristics of the 8 images obtained with Verluma. As Dr. Brill mentioned, Dr. 9 Nelp was involved in the Phase I-II studies of Verluma, 10 defined the pharmacokinetics, biodistribution and dosimetry 11 and he was also one of two nuclearmedicine consultants who 12 read Blindly all of the images in the pivotal study. 13 Dr. Nelp. 14 DR. WIL B NELP 15 DR. NELP: Thank you, Dr. Salk. I am Wil B. Nelp, 16 Professor of Medicine and Professor of Radiology at the 17 University of Washington School of Medicine in Seattle. I 18 have been the Director of the Nuclearmedicine Programs there 19 for a number of yea r s . 20 What I will do is show you the basic 21 biodistribution and kinetics and excretion of the NR-LU-10 22 Technetium-labeled FAB product and then I will show you some 23 examples of tumor labeling as it has been applied to small- 24 cell lung cancer. 25 [Slide.] MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 :202) 546-6666 116 1 This is done immediately after injection. This is 2 an anterior gamma-camera imaging planar view of the chest. 3 A.s you would expect with this protein, you see predominantly 4 vascular distribution, cardiac blood pool, major vessels of 5 the neck, lungs, liver. 6 [Slide.] 7 If we look lower in the abdomen, we, again, see a 8 blood-pool image, faint visualization of the kidneys, the 9 renal blood pool, other vascularity. Notice in the left 10 lobe of the liver sort of a defect. This is a large 11 metastatic lesion from carcinoma of the colon. This was a 12 Phase-II study where we did a lot of the biodistribution 13 studies, a Phase-II patient. 14 This image was probably made 20 minutes after 15 injection. You can see very early peripheral adherence of 16 antibody to the tumor. 17 [Slide.] 18 If we come back in three hours and look at the 19 head and neck region, now we see this image. We see the 20 thyroid. This is not free pertechnetate. This is a cross 21 reaction between the antibody and a naturally occurring 22 antigen that is in normal thyroid tissue so you always see 23 the normal thyroid as part of the image picture. 24 Up here, the pituitary. This is also another 25 cross-reactive antigen in the anterior pituitary and, very MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 '207 546-6666 117 1 often, one can define that. 2 [Slide.] 3 If we go down into the abdomen, now, at three 4 hours, we notice now we have fairly intense activity in the 5 kidneys because this is the major route of excretion of the 6 product. Some slight clearance in vascularity and, note, 7 there has been increased labeling of that tumor mass now at 8 the three-hour period. 9 [Slide.] 10 Eight hours; some further clearance, continued 11 activity in the kidneys. This is relatively more intens 12 compared to surrounding tissue. And now we see 13 radioactivity in the bowel. This is because, as the product 14 is degraded, the lysine portion, which is linked to the 15 chelate which traps the Technetium, comes off and is . 16 excreted. Part of that excretion goes through the liver 17 into the biliary system and is excreted into the bowel. 18 [Slide.] 19 If we look further down at 8 hours, this is 20 bladder, lower abdomen, we see activity in the testes. This 21 is a non-specific uptake in the testes, not specific for 22 this antibody but it is seen with a number of antibodies. 23 We can't identify cross-reactive antigens, but it will be 24 seen in the images. 25 [Slide.] MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 118 1 If we go out to 24 hours, which was what our 2 biodistribution studies did, now we can see much clearance 3 of general activity, still prominence of the kidneys, 4 residual activity in the gut and then this tumor now is 5 sitting pretty much in very high contrast to the surrounding 6 tissues. 7 If we put all this information together, we get 8 biodistribution data that looks something like this. 9 [Slide.] 10 This is serum clearance over time, 24 hours, very 11 rapid'disappearance of activity from the serum due to 12 immediate distribution of the product and early excretion. 13 This dominates the curve, has a T1 half of about one-and-a14 half hours, and then there is a minor longer-phase component 15 here. 16 [Slide.] 17 If we do total body counting on subjects, we get 18 this kind of a pattern. This is the disappearance of 19 activity from the total body over a period of 24 hours at 20 which time, on average, about 38 percent remains in the body 21 and the reciprocal of that, of course, is the urinary 22 excretion curve, the major route of excretion. 23 What comes out in the urine initially? The 24 preformed FAB intact, and that continues to come out. And 25 then degradation products, again, predominantly the lysine MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-SS66 119 1 adduct with the chelate and Technetium still bound together. 2 There is really no free Technetium, per se, that comes out 3 in the urine. 4 Now, I am going to show you some images of small5 cell lung cancer. These images were done at the 14-hour 6 time period which was the way the study was carried out. I 7 will show you the following. 8 [Slide.] 9 This is a patient with a pleural-based lung 10 lesion, primary tumor here, soft-tissue mass here, 11 supraclavicular lymph nodes, small-cell cancer of the lung. 12 ' [Slide.] 13 This image was done at 14 hours and it is in the 14 posterior projection. So this is the head, the thorax, the 15 tumor. The lymph nodes are on the right side now, which is 16 over here. Notice the intense labeling in the diseased 17 lymph nodes here. This is the pleural-based lesion which 18 looks somewhat more extensive that was appreciated on the X19 ray and on CT. 20 Notice these multiple discrete areas of activity 21 down the spine. This is distinctly abnormal. This 22 patient's bone scan was completely normal. This is activity 23 in marrow, but to confirm this, an MRI was done and showed 24 diffuse and punctate areas of marrow invasion of small-cell 25 lung tumor corresponding to these areas and several areas MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 . 120 1 where bony invasion had also take place. But the bone scan 2 remained negative. 3 In this patient, bone-marrow aspiration did show 4 small-cell lung-cancer cells. 5 [Slide.] 6 The liver. The chest is up here, not much 7 activity seen. It is cleared. This is the region of the 8 liver. This patient has multiple hepatic metastases 9 demonstrated on CT imaging. You can see these discrete 10 areas of tumor labeling within the liver that correspond to 11 those abnormal findings, and you note the good contrast ' 12 between the lesion and the surrounding normal tissue. 13 Here is the bowel, again, you see. This activity 14 could be further reduced by a very successful catharsis. 15 However, it is variably present and invariably present to 16 some degree. 17 [Slide.] 18 If we look at this patient's anterior view of the 19 chest, we see this type of image. Here is the primary lung 20 tumor. Here are some mediastinal nodes. Here is, actually, 21 a small lymph node next to the thyroid. 22 We can do SPECT, which is a tomographic image 23 where we take slice through the chest and we can, then, go through the area of interest and remove activity above and remove activity behind the lesion and enhance the contrast. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (7.07) 546-6S56 . 121 1 [Slide.] 2 SPECT on this patient shows these two lung lesions 3 nore dramatically and separates them out a little bit 4 better. This image is rather blurry, but that is the nature 5 of the instrument that was used by this investigator. 6 [Slide.] 7 Another anterior view of the chest. This is the 8 primary tumor, pleural-based and extending toward the 9 mediastinum. But notice thyroid. Notice all of the 10 activity in the ribs, sternum, et cetera. This, again, is 11 distinctly abnormal. ' 12 [Slide.] 13 If we look at the posterior view of this patient, 14 it looks like a normal bone scan. This is distinctly 15 abnormal. Notice that there is activity down here in the SI 16 joints and in the pelvis, as well. The kidneys, of course, 17 are seen. 18 This is marrow tumor. This patient did have an 19 abnormal bone scan but, as it turns out, only one vertebral 20 body was abnormal and several ribs. The bone-marrow 21 aspiration on this patient, however, showed heavy invasion 22 with small-cell lung tumor. 23 So this is part of the learning process to learn 24 these patterns. What else did we see? We can see some 25 artifacts. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 122 1 [Slide.] 2 This is an anterior lesion. These are the 3 axillary folds. These are not abnormal lymph nodes. This 4 is an artifact where skin folds come together and you get 5 forward-angle scattering commonly recognized in Technetium 6 images procedures. This can be eliminated by simply 7 elevating the arms at the time of imaging. 8 [Slide.] 9 This is an interesting case here. Here the 10 investigator put the intensity very high on this film to 11 illustrate the point. So this is a posterior view of the.' 12 chest that had -- this is upscatter from the kidneys because 13 the intensity is so high. Here is the spin. Now, this 14 little triangular area, which is sort of in the posterior 15 basilar segment of the lung, clearly has abnormal deposition 16 of activity. It was noted so by the blinded reviewer. But 17 this is a post-operative site. This patient had surgery in 18 this tissue three days prior to the image. 19 The investigative reviewer, who was not blinded, 20 realized this and properly identified it as an operative 21 artifact. But what happens is the tissue is injured, and 22 there is some edema. There is a protein leak into that 23 space. So you will see collections of antibody under those 24 circumstances. 25 It can also happen in sites of intramuscular MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 123 1 injection and other infectious areas but, if you know the 2 clinical history, these things are usually relatively easy 3 to sort o u t . 4 How is the learning process for interpreting these 5 images in the nuclearmedicine community? I had the 6 opportunity to visit many of the participating sites, talk 7 with the investigators and, in particular, the 8 nuclearmedicine physicians. Nuclearmedicine physicians 9 spend about 80 percent of their time, at least, interpreting 10 images of Technetium. So they have -- this is a familiar 11 isotope energy. Their equipment is very well designed to: do 12 Technetium imaging, and I would go through biodistribution 13 artifacts, tumor labeling and so forth with them. 14 I would characterize the learning curve as fairly 15 straightforward and relatively simple. 16 Thank you. I am going to turn the podium back to 17 Dr. Salk. 18 DR. PARKINSON: Can we just ask some questions 19 while you are there? 20 DR. INGLE: As in a bone scan, a very sensitive 21 but not specific as a Technetium bone scan, say, in breast 22 cancer, in lung cancer, you have described a number of, I 23 guess, false positives -- that is, areas of increased uptake 24 that are not tumor. Do you have any other thoughts on this? 25 Trauma, you said, will do it. How about arthritis, other MILLER REPORTING COMPANY, INC. ' 507 C Street, N.E. Washington, D.C. 20002 :202) 546-6666 at 124 1 benign conditions. Will they give increased uptake such 2 that if you have an abnormal scan, you need to corroborate 3 the abnormal findings with some other images technique. 4 Just any comments on that. 5 DR. NELP: The false-positive rate, actually, was 6 quite small in this pivotal study, but we do see deposition 7 of protein in spaces where injury occurs and the best answer 8 to your question is, in the practice of medicine, you 9 usually evaluate the patient, you usually examine the 10 patient, you get the correlative information, and if you 11 have an inflammatory joint or if you have a traumatic 12 arthritis of the joint or if you have a post-operative 13 region, these things are usually easily identified. 14 But false positives in terms of what looks like 15 tumor localization, like in the marrow and this sort of 16 thing, those are not false positives. But I think it is 17 easy to sort out at the practical-application level. 18 DR. PARKINSON: I am going to suggest that we have 19 further discussion after the full presentation and we have a 20 chance to look at the data. 21 DR. SALK: That question will be answered, in 22 part, by some of the data that you are going to see now. 23 Basically, in the pivotal study, there were only three 24 instances of inadvertent false positives. There were two 25 axillary skin folds and one, that patient you saw with the MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 125 1 post-cervical thing. 2 The other kinds of things that Dr. Nelp referred 3 to are either theoretical or have been observed with other \ 4 antibodies in other situations as non-specific hyperemia 5 inflammation. But it was not a problem in this study. 6 DR. DARRELL SALK 7 [Slide.] 8 DR. SALK: The objective of the pivotal study, as 9 I said before, was to evaluate the ability of Verluma 10 imaging as shown to you by Dr. Nelp, to stage people with 11 small-cell lung cancer in comparison with current imaging- 12 methodologies. 13 [Slide.] 14 All patients had a new diagnosis of biopsy- 15 confirmed small-cell lung cancer and they underwent a series 16 of non-invasive tests including physical examination and 17 four diagnostic imaging procedures, traditional ones, and 18 including a Verluma imaging. 19 The results of the physician exam and traditional 20 imaging tests define the traditional stage. The results of 21 the Verluma imaging, by itself, defined the Verluma stage. 22 These traditional tests were evaluated and read as routinely 23 practiced in clinical medicine; that is, in an unblinded 24 manner with access to all clinical and laboratory 25 information as well as the results of other tests. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 126 1 In what we report here today, the Verluma images 2 were interpreted in a single blinded review performed 3 centrally by one of two nuclearmedicine consultants. No 4 information was available about the imaging procedure 5 including information about the patient, other tests or, as 6 we mentioned, any technical information related to the. 7 imaging procedure, itself. 8 Thus, the data that we present to you here today 9 represent the most stringent challenge for Verluma, the 10 comparison of a completely blinded review of a single test 11 with an unblinded assessment of four tests, a physical ' 12 examination, and access to complete clinical and laboratory 13 information. 14 [Slide.] 15 96 patients were enrolled in the study and 89 were 16 evaluable. Among the 89 evaluable patients, there were 610 17 individual lesions and 333 affected organs; in other words, 18 organs that had at least one lesion. It was these 333 19 affected organs that formed the basis of the staging 20 analysis in this study. 21 These demographic characteristics are similar to 22 the general population of patients with newly diagnosed 23 small-cell lung cancer. There were no apparent sampling 24 anomalies. 25 [Slide.] MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 fono\ 127 1 This slide summarizes the statistical results for 2 staging patients with small-cell lung cancer using Verluma 3 imaging as a single test read unblinded in comparison with 4 the unblinded reading of four tests plus a physical exam. 5 The sensitivity for detection of extensive disease 6 was 77 percent for Verluma imaging by itself with a very 7 high positive predictive value compared with 88 percent for 8 the full battery of standard tests. The specificity for 9 detection of extensive disease is also quite high with a 10 negative predictive value of 69 percent. 11 ' Thus, a fully blinded reading of Verluma images- 12 alone compares very favorably with an unblinded assessment 13 of four imaging tests in a physical examination. 14 [Slide.] ' 15 For purposes of illustration, here is a clinical 16 model that uses the study data to predict what would be 17 expected to be the outcome of staging 100 patients in 18 clinical practice using either the full battery of non 19 invasive tests or Verluma imaging read in a blinded manner. 20 With the full battery of traditional tests, 56 21 percent of the patients are accurately identified as having 22 extensive disease and 34 percent of the patients are 23 accurately staged as having limited disease, for an overall 24 accuracy of staging of 90 percent. 25 With a blinded reading of Verluma images, the MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 128 1 overall accuracy of staging is 82 percent, 49 for extensive 2 disease and 33 for limited disease. A very small percentage 3 of patients is overstaged; that is, they are said to have 4 extensive disease when, in fact, they have limited disease. 5 Such patients might not receive potentially beneficial chest 6 radiation therapy. 7 This false-positive rate is the same for 8 traditional methods and for a fully blinded reading of 9 Verluma images. Thus, a low rate of false positives is a 10 problem for all staging methods of small-cell lung cancer 11 and it points out the need for good clinical judgment in 12 making critical staging decisions that are based on a single 13 positive finding. 14 As Dr. Nelp pointed out, and we discussed a moment 15 ago, some of the false-positive findings experienced in an 16 artificially blinded reading of Verluma images were due to 17 technical artifact that would be recognized and properly 18 interpreted in an actual clinical setting. 19 These 8 percent of patients would be understaged 20 by the traditional battery of non-invasive tests. Thus, 21 they would be said to have limited disease although they 22 actually have extensive disease. Such patients might 23 undergo costly, potentially toxic, chest radiation therapy 24 that would be of no benefit. 25 The patients that were understaged by this full MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 54 6 - 6 6 6 6 129 1 battery of traditional tests, of course, would not have been 2 identified as false negatives if Verluma imaging had not 3 been available. Because Verluma imaging is complementary to 4 other staging tests, it identifies some disease that may not 5 be seen initially by other tests. 6 In our study, when a positive Verluma finding 7 suggested the possibility of a false negative by the 8 standard battery, we did additional evaluations to confirm 9 whether the tumor site was positive or not and only if those 10 evaluations corroborated the Verluma finding was the lesion 11 defined as positive. 12 Although this model shows that Verluma alone 13 understages 15 percent of patients, in practice, all 14 patients who have no evidence of extensive disease would 15 continue to have additional standard tests until either, by 16 exclusion, all tests are negative and a diagnosis of limited 17 disease is made or these 15 would be recognized to have 18 extensive disease. 19 Thus, in practice, using Verluma imaging first and 20 then following with additional tests as needed, both these 21 sets of false negatives would be eliminated, reducing, 22 therefore, the number of patients exposed to non-beneficial 23 non-radiation therapy. 24 Having evaluation Verluma alone against an entire 25 battery of standard tests, we compared Verluma images MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 / r\ > \ t: a . r fT t? r ~ - 130 1 against each individual test in order to assess which test 2 provides the most information. 3 [Slide.] 4 It is not surprising, since it is a whole-body 5 imaging agent that is not limited to a single anatomic 6 region or organ system, Verluma imaging identifies the most 7 number of lesions, the most diseased organs and, therefore, 8 the largest number of patients with extensive disease, more 9 than any other single test. 10 This fact may appear to be self-evident, but the 11 analysis here illustrates why Verluma imaging is, therefore, 12 the best test to use first in staging patients with small 13 cell lung cancer. More patients can move rapidly on to 14 therapy after only a single test. 15 In addition, this analysis suggests that CT of the 16 abdomen is the next-best test to use which does, indeed, 1. 17 turn out to provide the most additional information when 18 added to Verluma imaging. 19 [Slide.] 20 In summary, Verluma is safe. Verluma is the most 21 sensitive single test for establishing the diagnosis of 22 extensive disease with a high positive predictive value. 23 For staging patients with small-cell lung cancer, a single 24 blinded reading of Verluma images is comparable to an 25 unblinded evaluation of all four traditional imaging tests MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (2021 546-6666 131 1 plus a physical examination. 2 [Slide.] 3 These conclusions support the indication that 4 Verluma can be used for the initial staging of patients with 5 small-cell lung cancer. 6 I will now ask Dr. Turrisi to take the microphone 7 and he will comment on the usefulness of Verluma as seen 8 from the perspective of an oncologist. 9 DR. ANDREW TURRISI 10 DR. TURRISI: Good afternoon. My name is Andrew 11 Turrisi. I am Professor and Chairman of the Department of 12 Radiation Oncology at the Medical University of South 13 Carolina. I am trained as a medical oncologist and I began 14 my radiation training here in Bethesda and, subsequently, 15 have spent my last 13 or so years involving myself in 16 clinical trials. 17 I have had an interest in small-cell lung cancer 18 since medical school days when a patient that I cared for 19 with extensive disease responded to chemotherapy and had a 20 Lazarus-like response. He got out of the bed and walked. 21 [Slide.] 22 Some of this may be gratuitous to most of you that 23 are here, but I just want to focus on the fact that we deal 24 with limited and extensive and not this traditional AJC 25 systems of staging with the TN&M system. This is used in MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 . 132 1 lung cancers and, sometimes, on small-cell for patients 2 going to surgery, but is not routinely useful. 3 Why do we do it this way? Because the limited and 4 extensive categories provide prognosis. In limited disease, 5 we believe that goal is the cure and that, in extensive 6 disease, the goal is palliation, a worthy goal. But 7 distinguishing these two different sets of patients is the 8 first step in how we go on to treat these patients. 9 Dr. Salk kindly provided you some what I will 10 charitably call "dated information" about outcomes in small11 cell lung cancer published in the DeVita textbook. But the 12 cure rates now for limited disease in group studies in 1995 13 show a 40 percent 2 -year survival and, perhaps, a 20 to 14 30 percent 5-year survival in patients with limited disease. 15 This year, at the ASCO Meeting, in the Program 16 Book, there is going to be an abstract that also shows that 17 a reanalysis of extensive disease shows that we haven't made 18 much progress in the last 20 years. So, separating out that 19 group of who has got extensive and who has got limited, is 20 very crucial. 21 Why is it crucial? Because it is going to guide 22 treatment and, as I have just mentioned, it influences 23 prognosis. 24 [Slide.] 25 I would like to categorize for the next couple of MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 133 1 minutes what this distinction does for patients and then 2 what it does for physicians. For patients, they are 3 generally in a hurry to get going on treatment. They are 4 scared and they want to move because they think that the 5 therapy is going to provide them with their best measure and 6 hope for cure. 7 So staging, now, will consume less of that time if 8 we have Verluma. The process becomes a little bit more 9 efficient. We get accurate staging with that study 10 82 percent of the time by the data Darrell Salk showed you a 11 few minutes ago. And it is the most sensitive test to use 12 first. We can identify that extensive group of patients 13 77 percent of the time. 14 Once Verluma is positive, you don't have to bother 15 to do all those other -- and I have phrased it "meaningless" 16 -- studies at that time. That eliminates, or certainly 17 should reduce, the cost of unhelpful tests in three-quarters 18 of the patients. 19 Another important feature is about the feature of 20 false positives. At least in the study that we have shown 21 you today, that of the three false positives with clinical 22 input, those false positives are eliminated. But I think it 23 is safe to say that the frequency of false positives is very 24 small. 25 [Slide.] MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 134 1 What are physician benefits, and, I guess, health 2 care benefits, for that matter. We now can use tests much 3 more efficiently. We have got a test that shows us that 4 with a positive Verluma image, there is no need for tests in 5 extensive disease save for if you want to get that test to 6 further confirm and follow that lesion to see if it has 7 improved with therapy. 8 Again, in this study that we showed today, three 9 false positives were eliminated once you knew that the 10 clinical input said the patient had surgery or their arms 11 were down and it was a technical artifact. So we had an '. 12 accurate 'staging of 82 percent of the patients screened. 13 Again, the sensitivity for extensive disease had picked up 14 77 percent of the patients. 15 If you have a negative Verluma image, that means, 16 in many cases, that the patients have limited disease and 17 only these patients require further testing. More extensive 18 staging procedures in this group can pick up some of the 19 patients that are "understaged, and we can accurately 20 diagnose those patients as was shown by Dr. Salk's 21 presentation. It shows you that the traditional stages 22 don't always pick these things up, so the tests are 23 complementary. This one merely is better to use first. 24 [Slide.] 25 So, to summarize this, and what I have just told MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 5 4 5 - 6 6 6 6 135 1 u, is that Verluma decreases the overutilization of 2 staging procedures, reduces the time spent defining stage 3 and that should lead to a reduction in cost. Only the 4 aegative Verluma gets the full battery of tests. This leads 5 bo an increased comfort for patients and eliminates, or at 6 least decreases, the frequency of bone marrows which, in and 7 of itself, takes three days to do, to decalcify, to read and 8 then interpret. 9 And all of this you get with no compromise in 10 accuracy. 11 ' I think I can say, as a clinician and as an . 12 investigator, that I will find having Verluma around, 13 particularly in these days when we are asked to be efficient 14 and use resources more carefully, a great boon and something 15 that I look forward to having. 16 I thank you for your attention. 17 DR. SALK: We would like to take questions now. 18 DR. PARKINSON: Sure, if you are ready. At this 19 point, I think it is appropriate for me to, first of all, 20 apologize for being what I thought was half an hour early 21 for this meeting -- the last schedule that I had had. I 22 haven't been half an hour early for anything in years and I 23 guess the record still continues. 24 At this point, I think it is appropriate to open 25 up for questions to the sponsor. Dr. Siegel, you had a MILLER REPORTING COMPANY, INC. 507 C Street, N . E . Washington, D.C. 20002 (202) 546-6666 - at 1 question earlier. 136 2 DR. BARRY SIEGEL: Actually, I had a comment 3 earlier, and I will use it to start, namely in response to 4 Dr. Ingle. I agree with what Dr. Nelp said completely. In 5 routine clinical practice in nuclearmedicine, we do chat 6 kind of image problem-solving all the time every day where 7 we see things that look like they are conforming to the line 8 of a glenohumeral joint, where we go out and talk to the 9 people, look at the available chest radiograph, figure out 10 that it is osteoarthritis and go on from there. 11 So I really am not troubled by the concept, that 12 the kinds of things like recent surgical incisions that Dr. 13 Nelp pointed out, are going to cause problems, significant 14 problems, in clinical practice. So that is my first 15 comment. 16 Question for Dr. Turrisi, actually. The way you 17 all have built your model, assuming that the prior 18 probability of limited disease in patients who would be 19 subjected to imaging is going to be somewhere around 20 30 percent. What fraction of people who walk in the door 21 with small-cell lung cancer have clinically evident 22 extensive disease right at the time they are first seen so 23 that they really wouldn't be candidates for imaging at all, 24 because they come in with a superior venacaval syndrome, and 25 the initial chest film shows bilateral mediastinal and MILLER REPORTING COMPANY, INC. 507 C Street, N.E. ' Washington, D.C. 20002 C202) 546-6666 " 137 1 hyaler adenopathy, or they have a seizure and the CT scan 2 shows a brain metastasis. 3 I am actually wondering if this will apply to a 4 smaller fraction of small-cell lung cancer than you 5 intimated in your broader modeling approach. 6 DR. TURRISI: I think that is a good question and 7 I am going to refer back to Dr. Salk because I think he has 8 some specific data on this, but it is not overwhelmingly 9 common that the simple things, like chest X-ray and physical 10 examination, make the diagnosis as extensive. But, you are 11 absolutely right. If you have that information in today's 12 world, unless there are specific requirements for someone to 13 go around and study. 14 But, in the real world, if you know that they have 15 extensive disease, you probably don't need those imaging 16 tests. 17 DR. BARRY SIEGEL: But you wouldn't have found it 18 in this trial. 19 DR. TURRISI: I agree with that. 20 DR. BARRY SIEGEL: Because those patients probably 21 wouldn't have ever gotten into the study because they would 22 have been undergoing radiation therapy within two hours of 23 getting to the hospital. 24 DR. SALK: That is a possibility. Those patients 25 were not excluded from the study. They could participate, MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 ' 138 1 but there may have been some selection beforehand. It turns 2 out, in this study, if you actually look at what is added to 3 Verluma by physical examination and chest X-ray or, 4 conversely, what Verluma adds to a physician exam and chest 5 X-ray, it is not a great deal either way. 6 Do you have that slide available? I will tell you 7 which one it is. It is M-l, Slot 53. 8 [Slide.] 9 As I said, this shows the information that is 10 added to a physical examination and chest X-ray by Verluma 11 or, conversely, to Verluma by physical examination and chest 12 X-ray. The vertical axis shows the number of diseased 13 organs. It may be hard for you to read, so I will read 14 across the bottom here. This is lung, mediastinal lymph 15 nodes, pleura, bone, liver, abdominal masses and lymph 16 nodes, other abdominal lesions, brain, other head, neck, 17 axillary lymph nodes, other categories and bone marrow. 18 You can see here the purple is that information 19 that was collected by a physical examination and chest X20 ray, in other words, in the patient's first walk in the 21 door. The green is what is added by Verluma imaging. 22 Conversely, this is the Verluma imaging, the red. 23 And you can see the blue, the additional information from a 24 physical examination and chest X-ray. It is interesting 25 that in such areas as in the neck and axillary lymph nodes MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington', D.C. 20002 ion-51 sAC-fifiKS 139 1 where the physical examination picks up a lot, this non 2 invasive procedure also is very good at picking up those 3 things which would normally be found only on physical 4 examination. They are not very well defined by other 5 imaging modalities. 6 Similarly, the bone marrow you see here, Verluma 7 alone and, again, here Verluma alone, is picked up and it is 8 obviously, other than a full-body MRI, the only non-invasive 9 test that will give you an assessment of bone marrow. 10 Does that answer your question? 11 DR. BARRY SIEGEL: Yes; I think so. ' 12 " DR. FORASTIERE: I am wondering if you have that 13 same data for physical examination and CT because most 14 people who have an abnormal chest X-ray, nobody does a 15 bronch on them or a percutaneous needle aspiration until 16 they have gotten a chest CT. That includes, usually, the 17 upper abdomen. 18 So, since that really is the baseline evaluation, 19 a physical examination and at least a minimum of a chest CT 20 and, most of us, if you suspect lung cancer, do the full 21 abdomen anyway. But let's say it is just the chest CT and 22 upper abdomen, how much would Verluma add in that situation 23 to defining someone who has extensive disease? 24 DR. SALK: Verluma imaging adds, in that 25 situation, as well because it reads something beyond what MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 14 0 1 you see just in the chest. Dr. Abrams, do you have a 2 comment on that? 3 DR. PAUL ABRAMS: My name is Paul Abrams. I am a 4 medical oncologist as well as President of NeoRx. The 5 comment I would make on that is that today I certainly 6 accept what you say as what the procedures a r e . I think 7 that what we are trying to show here is that the data 8 suggest that if you have a chest X-ray from which you 9 suspect small-cell lung cancer, if you have a needle biopsy 10 that shows small-cell lung cancer, that Verluma, in fact, is 11 a more efficient way of staging patients because, not only12 will -you find what is in the abdomen, but you will" also find 13 what is in other tissues, as has been defined by Dr. Salk. 14 So the question is which one to do. Our 15 suggestion, based on the data, is that Verluma, in fact, is 16 a more powerful test and that fewer patients, following the 17 CT of the chest and following the CT of the abdomen, would 18 require any further evaluation. 19 DR. FORASTIERE: But how many further? That is 20 what I was asking, how much of a difference does it make 21 because, the fact is that you wouldn't do the scan until you 22 have the diagnosis and you wouldn't make the diagnosis until 23 you had the chest C T . 24 DR. ABRAMS: That really depends on how you want 25 to practice. If you have a chest X-ray and you do a needle MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 54S-6SSS - . 141 1 biopsy of the lung, or a fiberoptic bronchoscopy or .a..lymph 2 node tnat is needed, and you have the diagnosis, you then 3 have a choice of what test to do. It can be a CT of the 4 chest and abdomen. It could be a CT of the head. It could 5 be a bone scan or it could be Verluma. 6 Again, as the data shows, Verluma is the most 7 powerful test to use in order to find out whether the 8 patient has extensive disease or n o t . Specifically, to 9 answer her question - 10 DR. SALK: Let me ask you to flip it over a little 11 bit because in the study that we did here, we did look at. 12 what happens if you do contrast of the abdomen in" addition 13 to Verluma, the two tests together versus Verluma alone, et 14 cetera. It turns out that when we do this fully blinded 15 analysis, a single reading under the settings that I 16 describe, the CT of the abdomen and Verluma together have 17 essentially the same negative predictive value as the entire 18 battery of standard tests. 19 So, under these stringent conditions, your 20 negative predictive value or evaluation of limited disease 21 is equivalent. When we do the analysis according to the 22 design we had in the protocol which we felt provides a 23 better estimate of what is liable to happen in practice - 24 in other words, it takes in to consideration the technical 25 artifacts and so forth -- the antibody imaging alone was MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 . .. (202) 546-6665 142 1 equivalent in its negative predictive value to the full 2 battery of standard tests. 3 And those numbers are 80 percent, 80 and 81 4 percent, or 74 and 77 percent, under the two different forms 5 of analysis. 6 DR. FORASTIERE: I understand what you are saying. 7 I am just making the point that current practice for a 8 pulmonologist to do bronchoscopy or someone to do a needle 9 aspiration, they have the information from the a chest CT 10 before they proceed to make the diagnosis. You wouldn't do 11 the Verluma scan until you have the diagnosis. 12 - DR. SALK: Absolutely. We certainly can't dictate 13 clinical practice here. If somebody already has a diagnosis 14 of extensive disease, they shouldn't get Verluma any more 15 than the other way around. 16 DR. OZOLS: I think Dr. Forastiere makes a good 17 point. I think that is really the crux of the matter about 18 the clinical utility. You are talking about the 19 sensitivities tests; we are certainly going to do a physical 20 examination. This test, Verluma, doesn't appear to be very 21 good for head problems, so you are going to have to do a CT 22 scan of the head. 23 You are likely, as you point out, going to want to 24 do that CT scan of the chest before you do anything. If you 25 do those tests as part of the routine practice right now, it MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 143 1 seems that you are really coming down to a very small number 2 of indications or instances where Verluma would help you. 3 The standard of care when somebody walks in with a 4 diagnosis of suspected lesion, what are you going to do? In 5 your study, in your pivotal study, what was done before you 6 got the tissue biopsy. Any one of these tests; right -- for 7 your new diagnosis of small-cell. 8 DR. SALK: We didn't control for that. 9 DR. OZOLS: Right. So a lot of tests were done 10 before that. 11 DR. SALK: No, not necessarily. Typically, I : 12 think the histologic diagnosis was made very early on, that 13 patients may have come to a tertiary center already with 14 that diagnosis made. But the point about, for instance, 15 sensitivity in the head, you are right. Based on the 16 numbers that we show here, the very small number of brain 17 lesions -- there were only 12 in the entire study -- the 18 detection rates are low. 19 However, when you look at staging the entire 20 patient, there is very little impact from that because there 21 are very few times that a brain lesion alone defines 22 extensive disease. 23 DR. OZOLS: That's correct. 24 DR. SALK: Therefore, you only have to go on and 25 do your CT of the head if you have still have not found MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 ' (202)" 546-6666 . 144 1 extensive disease by other tests. One thing I want to 2 emphasize again is that with Verluma imaging, anywhere in 3 the process, if you have done all of these things and you 4 still have no extensive disease, then your diagnosis of 5 limited disease is more accurate than the one that is 6 available just with the traditional battery of tests 7 currently. 8 D r . Turrisi, do you have a comment? 9 DR. TURRISI: The point I would make is that that 10 is a good observation that maybe if you did things the way 11 we have always done them, it doesn't add that much, if you 12 have already expended all four imaging tests. But if you 13 look at it the other way, it accurately picks up three14 quarters of them and that means you only need to do that 15 battery of four extra tests or two extra tests if I give you 16 that they have a chest CT ahead of time and maybe even some 17 of the abdomen. 18 But, it reduces the number of total tests that you 19 have to do and you only get the head CT scan in that small 20 proportion rather than all 100 percent of the patients. 21 DR. OZOLS: Then what would you do first before 22 you did the biopsies, right now? 23 DR. TURRISI: If I were to see these patients now 24 and they a lesion that looks like a lung cancer, you get, 25 probably a chest CT scan, today. I don't know what you get MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 . ' (202) 545-6666 145 1 in California, whether you have to go to bronch first or you 2 get a needle on that before you are allowed to do the 3 expensive image. 4 But if we know that patient has small-cell by 5 either getting the CT or getting the bronch next, the next 6 test for me would be to get the Verluma because after the 7 Verluma, you get two nicely divided groups of patients. The 8 three-quarters that have extensive don't need any more tests 9 and then the one-quarter to one-third that does have limited 10 disease apparently then gets the rest of the battery of the 11 test to find out whether, in fact, they really are limited, 12 about-a quarter of the patients, or that understaged group 13 of patients that you then need to find out and prove that 14 they have got extensive disease. 15 DR. OZOLS: If the Verluma showed extensive 16 disease, you are not interested in the CT scan of the head? 17 DR. TURRISI: As a radiation oncologist, if that 18 person is clinically asymptomatic, there is actually data 19 now suggesting that you don't need to intervene with 20 radiotherapy off the top. But managing those with systemic 21 chemotherapy is a reasonable first go around for those 22 patients. 23 Clearly, if they develop a symptom, then you would 24 come to the beam and you would, in that case -- I am going 25 to tell you, half of the patients, ultimately, are going to MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 146 1 go on to have brain metastasis so those patients are going 2 to have an MRI or a C T . But it can be down the road because 3 we don't know today, even, if it is negative or positive 4 what that influence is. 5 DR. PARKINSON: Let me ask you a related question. 6 What about subsequent therapy. How important do you think 7 it might be to a medical oncologist or radiation therapist 8 to have baseline scans? We have no information on the 9 utility of Verluma in following patients, presumably. So 10 would you do subsequent scans without a baseline scan? 11 DR. TURRISI: Second-line therapy in small-cell; 12 lung -cancer does not work very effectively. If you want to 13 find out how that person is going to progress, I think 14 getting a battery of tests, the time that someone loses 15 performance status, has a new symptom, in order to guide 16 prognosis or further therapy to symptomatic lesions, that is 17 a sensible approach. 18 But, as I was trained, we did -- in the era when 19 cost was not considered, you would do all of these screening 20 tests. I think, today, that is going to be relooked at. In 21 cooperative groups, we are asked, regularly, now to go back 22 and see do we really need to repeat the test with the 23 frequencies that we have had or, in fact, should we start to 24 back off and really think about, good and hard, whether we 25 need those extra tests. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 . 147 1 So, to directly answer your question, I don't know 2 if I need the baseline anymore. I used to, modeling things 3 after other diseases where it was important. I am really 4 not sure I need it today. 5 DR. MALCOLM: Actually, it is the same question, 6 Dr. Turrisi. But I am not sure if it was clear to me what 7 you were saying. If you have the patient "staged," with the 8 Verluma study, that question is with that patient's response 9 to therapy, what do you use to evaluate the response. I am 10 not saying for returns. But, to evaluate that patient, I am 11 not clear if you answered that. : 12 DR. TURRISI: Let me answer it because that is 13 slightly different. But if I knew that I had a Verluma- 14 positive patient and, let's say, it was only the bone 15 marrow. I will use one of Dr. Nelp's examples where the 16 vertebral bodies of the spine were positive. If I wanted to 17 have a followable test, I would order an MRI on that patient 18 to see the way the marrow looked. And then you would have 19 to MRI to follow that patient rather than Verluma. 20 The other side of the coin,* let's say it was a 21 liver lesion that glowed. You get a liver scan, or an 22 abdominal CT scan, and you can then follow that study from 23 that point forward. But you don't need a full battery of 24 every other test known to man, at least the way I look at 25 this. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-S66S 148 1 DR. SALK: Also, what lesions do you follow 2 clinically most of the time? Is it, indeed, one of these 3 distant metastases or what you have seen in the chest? 4 DR. TURRISI: The most common thing you would 5 follow and, actually, even in extensive disease, the most 6 common site of failure is the chest. I guess the point has 7 been made but I will make it again; the brain is the other 8 thing that shows up with increasing frequency. This is not 9 a time to academe in whether PCI should be used or not. I 10 won't get into that. 11 DR. SWAIN: I wanted to get back to the questions 12 we asked-at the beginning. Since it is established that you 13 do get a chest and abdominal CT in these patients before 14 diagnosis, how many patients with small-cell lung cancer 15 actually have a negative abdominal CT or negative liver and 16 positive bone scan? Do you know that? 17 DR. TURRISI: We don't get 58 percent. 18 DR. SWAIN: But in the literature. I don't know 19 the data, myself. I don't take care of lung cancer 20 patients. Negative liver and positive bone. 21 DR. TURRISI: How common is that? 22 DR. SWAIN: How common is that, because that would be the group DR. TURRISI: Negative liver scan and a positive bone scan? MILLER REPORTING COMPANY, INC 507 C Street, N.E. Washington, D.C. 20002 149 1 DR. SWAIN: Right. 2 DR. TURRISI: A positive bone scan, as an isolated 3 finding, is not a very common feature. On the study, if you 4 looked at the bone scan, it was, perhaps, the least 5 sensitive test to establish distant disease. But the other 6 question is how common is the abdominal CT negative in 7 extensive-disease patients, Dr. Swain? 8 DR. SWAIN: No. I am trying to get at the group 9 that this is going to help. If we have an abdominal CT that 10 is negative, then you would normally go on to do the other 11 scans like the Verluma -- 12 DR. TURRISI: The premise is not entirely correct. 13 I don't think we would have to get-an abdominal CT scan. I 14 will just correct the illusion that the simple CT from chest 15 down through adrenals is a good enough evaluation of the 16 liver. If you really want to look at the liver, you have 17 got to ask for it and you have to get a separate contrast 18 injection as the scan goes through that spot. 19 But your point about abdominal CT scans; there are 20 a fair number that are negative. Certainly, all of the 21 limiteds are negative. But I don't have a good handle on 22 which ones are negative of the rest of the group. It would 23 be the sensitivity of the CT on the - 24 DR. ABRAMS: I think it is worth pointing out that 25 if you just look at the abdominal CT, it only detected MILLER REPORTING COMPANY, INC. 507 C Street, N.E. . Washington, D.C. 20002 150 1 58 percent of all those patients with extensive disease. 2 So, without picking it out by whether it is bone or bone 3 marrow or whatever, it is only detecting 58 percent whereas 4 the Verluma test is detecting 77 percent. 5 So you have almost a 20 percent improvement in the 6 pickup of patients.with extensive disease through one or 7 more of those organs. 8 DR. OZOLS: I think that takes it out of context 9 of what you do when you see the patient. You are still 10 going to have the physical examination which is about 11 18 percent. You are going to do a chest X-ray and that is 12 going to put some of these -- so just to say that you have 13 got the Verluma scan at 77 percent may overstate, in light 14 of what you are going to do anyway. You are going to get 15 some of these tests before you make that diagnosis, and you 16 are certainly going to do a physical examination and a chest 17 X-ray and, likely, a CAT scan as well. 18 I just don't see how you are going to avoid doing 19 those tests. 20 DR. SALK: I can make a couple of points about 21 that. The first is that in the study, we did evaluate what 22 other information was provided first and how much was added 23 by each of the tests when mixed together and combined. The 24 data that was showed you indicate that the Verluma gives you 25 the most information of any test with a single test. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 151 1 And, in fact, it does so more than any combination 2 of the other tests. Clearly, people are going to practice 3 as they see and we can't advocate that everybody change 4 their practice and you do what we suggest. 5 We, nevertheless, feel that the data that we have 6 here indicate that Verluma provides the most information of 7 any test and, therefore, done first, is beneficial. 8 DR. OZO L S : I am not advocating to do the test 9 before you have the diagnosis. We are talking about what 10 the minimum that you would do before you did the diagnosis. 11 And before you did the diagnostic tests, you are not going 12 to do the Verluma. 13 DR. SALK: Before you do the histologic diagnosis. 14 DR. OZOLS: Right. That is your indication. It 15 says after the diagnosis is made. So what are you going to 16 do before you make that diagnosis? You are certainly going 17 to do a physical exam and you are certainly going to do a 18 chest X-ray and a CT scan. 19 DR. SALK: Right. And based on the data that we 20 have here, the physical examination, the CT, still allows 21 Verluma to provide a great deal of information. I am going 22 to ask Dr. Turrisi and Dr. Kyle Bryan to comment because 23 they are both oncologists. 24 DR. TURRISI: I just want to bring this one up 25 because I know it is not that long ago, but the CT scan MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 r.Oh'i1 152 1 wasn't around forever. We used to diagnosis small cell lung 2 cancer on the basis of a bronchoscopy and a chest X-ray. If 3 you ask me to define what the role is of CT scans and how 4 they have aided our management, they are like something that 5 we know we want them, but we don't know exactly why we have 6 to use them. 7 I can't argue you out of the idea that today our 8 clinical practice is to get CT scans, both the chest and the 9 abdominal. But I ask you to take that look into the future. 10 Will we be allowed to do those in all people if there is 11 another test and a better test, a more sensitive test, that 12 says you don't need to do those. 13 If you could make the diagnosis without doing a 14 chest CT scan, which is a little bit larger leap, you 15 certainly can make the diagnosis without doing an abdominal 16 CT scan. That, by itself, is a good enough reason as far as 17 I am concerned, not to mention not needing to do a head CT 18 or a bone marrow or a bone scan. 19 DR. SALK: Dr. Kyle Bryan is a physician at NeoRx 20 Corporation. 21 DR. KYLE BRYAN: My name is Kyle Bryan. I am a 22 physician. I work at the NeoRx Corporation. In addition, I 23 have a clinical oncology practice within a managed-care 24 setting. I think I can address a couple of the issues that 25 were brought up. .. MILLER REPORTING COMPANY, INC. 50^ C Street, N.E. Washington, D.C. 20002 /'**rvo\ c:a r r r?c C ' 153 1 First, the issue of the negative liver CT scan 2 with the positive bone images. To my recollection, I 3 believe the last addition of DeVita mentions 15 to 4 20 percent where a positive bone scan is the only evidence 5 of extensive disease. 6 With regard to following and the issue of the CT 7 scan versus chest X-ray and the timing within the diagnosis, 8 most of the patients, at the point when I see them, have 9 often had a chest X-ray followed by bronchoscopy. But many, 10 many of them have had chest CT scanning done at times 11 including the liver and adrenals. . 12 - But, as Dr. Turrisi pointed out, to get an 13 accurate assessment of the liver for full evaluation, you do 14 need to have a special contrast injection and you need to 15 tell the scanner that you are looking for liver lesions. 16 Often, at this point, the diagnosis is not made and so you 17 have to go back and scan again to get a better look at the 18 abdomen. 19 With regard to how to follow this, these patients 20 have some evidence of disease. Either they have had a 21 peripheral supraclavicular node that has been aspirated. 22 There was a chest X-ray lesion that detected, or a chest CT 23 lesion. nA We rarely, in the absence of any clinical evidence of bone pain or a neurologic finding, need to get a CT scan MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 r . 154 1 of the head in order simply to have a lesion to follow to 2 assess response to therapy. There is always some type of 3 lesion that you can follow. 4 DR. PARKINSON: Dr. Forastiere, I think that this 5 may be the last question before we go on. 6 DR. FORASTIERE: I was wondering if you have any 7 information on the development of HAMA with multiple 8 injections because I could envision a scenario, if this were 9 approved, where a patient received their first injection; it 10 was negative and they were shown to have limited-stage 11 disease, and they go on and receive treatment. Then they, 12 might be restaged down the road. It might be very- useful to 13 do this one test to get an indication of whether or not they 14 were still in a complete response on not because of the 15 sensitivity. 16 So if you were to do repeated restagings and use 17 this, would there be a problem with HAMA? 18 DR. SALK: First, I want to point out that in this 19 application, we specifically limit it to a single use of the 20 product. 21 DR. FORASTIERE: Okay. 22 DR. SALK: The very limited information we have is 23 difficult to come by because HAMA happens so infrequently 24 that any repeat imaging very rarely is done in a patient who 25 has had it before. We did happen to have one patient who MILLER REPORTING COMPANY, INC. 507 c Street, N.E. Washington, D.C. 20002 155 1 had -- one of the three patients there, that I described, 2 that had a transient HAMA that then went down -- who went on 3 to reimaging and the HAMA level, once again, went up and 4 came back down. 5 This is essentially one anecdote but, again, this 6 indication is for a single use of the product. 7 DR. FORASTIERE: Thank you. 8 DR. PARKINSON: Thank you very much. 9 DR. JAY SIEGEL: Just a brief comment about the 10 issue under discussion. In the FDA briefing package, it is 11 Page 14 of the draft I have. I don't know if everybody else 12 has the same draft and table. There is some data addressing 13 the impact of SSU. If the 57 patients in this study who had 14 a clear clinical extensive stage, there were only two where 15 that was identifiable by chest X-ray, suggesting, perhaps, 16 that there was some tendency not to include such patients in 17 the trial. 18 There were an additional 9, or only 8 correct, 19 that were identified on physical examination. I shouldn't 20 actually say additional. We don't know that those 8 might 21 not also have included a tumor with chest X-ray. 22 There are 5 identified by CT scan of the chest 23 which probably includes those who are positive on chest X24 ray so there may be anywhere from 8 to 13 patients there. 25 There are, certainly, far more where the CT of the abdomen, MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 . (202) 546-6556 156 1 had it been done, would have give evidence of disease. 2 DR. SALK: Very few patients are identified by 3 physical exam or chest X-ray to have extensive disease. The 4 reason the chest CT and chest X-ray do not frequently 5 identify extensive disease is not because of some bias in 6 entering the study but because it is relatively uncommon to 7 have bilateral lung disease which is the only instance of 8 extensive disease in small-cell lung cancer detectable in 9 the chest. 10 DR. PARKINSON: Thank you. 11 The first FDA presentation is by Dr. Berkower. . 12 FDA PRESENTATION 13 MR. IRA BERKOWER 14 MR. BERKOWER: My name is Ira Berkower. I am in 15 the Center for Biologies of FDA. I will be presenting the 16 first part of the FDA perspective on this product. 17 I would like to begin this afternoon by telling 18 you the Committee members which will be on the first slide. 19 [Slide.] 20 I will be presenting the chronology and the 21 findings of the early Phase I and Phase II studies and then 22 George Mills, who is one of the Cochairmen of the Committee, will be presenting the Phase III pivotal trial. I will just say that the PLA Committee consisted of the two Cochairmen, George Mills and Tom Bull and also MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 157 1 included the other members, Satish Mishra, Florence. 2 Kaltovich, Chris Joneckis, Dennis Klinman and myself, Ira 3 Berkower. 4 [Slide.] 5 The slide that was before this said NR-LU-10 is a 6 FAB fragment of a monoclonal antibody that is to be used for 7 imaging small-cell lung cancer. This slide shows a brief 8 chronology of the application. The Phase-I and Phase-II 9 studies were initiated with the filing of an IND application 10 in May of 1987. I will be discussing the 47 patients in 11 this study. : 12 This led to a Phase-III trial in August of 1988 13 involving 89 evaluable patients. That will be presented by 14 Dr. Mills. In December of 1989, these studies led to the 15 filing of a PLA. In January of 1992, for business reasons, 16 manufacturing of the monoclonal antibody was suspended for a 17 period of time. In September of 1993, a new manufacturer 18 took up this product and that sponsor filed the current PLA 19 in March of 1994. 20 So I will just be describing the Phase-I and 21 Phase-II studies. 22 [Slide.] 23 This is a list of the questions addressed during 24 these studies. The first question was the dose of the 25 protein which was found to be 10 mg, the dose of the MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 54b -6666 158 1 radiolabel, which was 15 to 30 mCi, and the optimum time of 2 imaging, which was found to be 14 to 20 hours. I will go 3 into this a bit more on the next slide in terms of 4 pharmacokinetics leading to these parameters being chosen. 5 The safety was evaluated in these studies both in 6 terms of clinical safety, which there were very few adverse 7 reactions, human antibodies to mouse which, in the Phase I 8 and II was detected in 5 out of 48 patients and, as well, 9 the radiation dose absorbed by these patients, indicated as 10 RADs; the radioactivity was that 4 RADs were absorbed in the 11 kidneys, 1 RAD in the thyroid, and all the other organs were 12 1 or less, which was considered a very good profile for this 13 type of imaging agent. . 14 The other thing that had to be evaluated at this 15 phase was which tumor type to go after since 16 immunohistochemistry studies revealed that this antigen is 17 found on a number of different tumors. Of course, these 18 studies led to small-cell lung cancer. Other lung cancers 19 were evaluated as we l l . 20 [Slide.] 21 So, on the next few slides, I would just like to 22 go into a little more detail on some of these issues. 23 First, the half life of the Technetium-labeled FAB fragment, 24 both in terms of its biological half life, which the 25 excretory phase was 10.5 hours, and the physical half life MILIiER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 /OHO) 159 1 of the radiolabel, which was 6 hours. 2 The good match between these two times, in part, 3 accounts for the equality of the images obtained because the 4 material was cleared from the blood pool relative quickly 5 for an antibody fragment, faster than in a whole antibody, 6 and yet the 6-hour half life limited the total amount of 7 time available for imaging, and most imaging is done at 8 about 3 half-lifes. 9 [Slide.] 10 This slide shows some of the artifacts that were 11 observed right from the first Phase-I and -II trials, the12 imaging artifacts. Some of these could be attributed to 13 specific binding, recognizing the antigen on tissues. These 14 included, as was shown by Dr. Nelp, the thyroid gland, which 15 seems to have an antigen-specific binding. Salivary glands 16 are sometimes picked up and the anterior pituitary. 17 As well, non-specific localization has been 18 observed on organs such as the testes and nasopharynx. 19 Finally, an imaging artifact occurs due to the excretion 20 pathway which is roughly two-thirds renal so there is a 21 strong renal image, kidney image, and the hepatobiliary tree 22 leading to some artifacts in the GI tract at various times 23 after imaging. 24 It is a very important point that has persisted 25 with this product through all the trials, that there is an MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 160 1 education process so that the reader will be able to take 2 these artifactual images into account in looking for tumor 3 images. 4 [Slide.] 5 Finally, in the Phase-I and -II trials, they 6 address the question of which tumors were the most likely to 7 succeed in terms of imaging. What I am showing here are a 8 variety of tumors that were tested. There were roughly 10 9 patients for this one, this one, this one and this one. I 10 am just showing the detection rate of involved sites. 11 This is not really an indicator of specificity Dr 12 efficiency of detecting them, just simply getting an idea 13 which were the good tumors to go for. As you can see, 14 small-cell lung cancer, 1 2 out of 21 sites, were detected 15 even in these early studies. 16 Non-small-cell lung cancer is also detected, as is 17 colon, breast, kidney, prostate, liver and cancers of 18 undetermined origins. Some of them were detected less well. 19 So the results of the Phase-I and -II study were interpreted 20 as showing that small-cell lung cancer would be the target 21 for further studies and this was pursued in a Phase-III 22 trial which will now be discussed by Dr. Mills. 23 DR. GEORGE MILLS 24 DR. MILLS: I am George Mills with the Center for 25 Biologies. We are going to discuss the pivotal Phase-III MILLER REPORTING COMPANY, INC. 507 C Street, N.E. . Washington, D.C. 20002 - (202V 546-6666 161 1 trial. 2 [Slide.] 3 This is PLA 94-0308. It was submitted into the 4 Center for Biologies on March 28, 1994. Under IND 2633, all 5 the clinical trials we will be discussing were performed. 6 Our sponsor is Dr. Karl Thomae. 7 [Slide.] 8 The pivotal Phase-Ill trial was titled a trial of 9 Technetium-labeled monoclonal antibodies for staging a 10 small-cell cancer. The dates of this trial were July 6, 11 1988 through July 6, 1989. All of the patients that are ; 12 accrued under the pivotal Phase-III trial that are submitted 13 in this PLA were performed at that time. 14 There were 24 sites in the United States. This is 15 a single-arm, open label, multi-center trial. 16 [Slide.] 17 The sponsor's proposed indication has been 18 reviewed by the sponsor, but let me take you through it 19 again, that the proposal for NR-LU-10 is for initial staging 20 of patients with biopsy-confirmed small-cell lung cancer. 21 For those patients with no evidence of extensive disease - 22 that is, staged as limited disease by NR-LU-10 -- the 23 sponsor is recommending that they have the standard diagnostic tests in order to continue staging because the standard diagnostic tests have demonstrated that they will MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 5 4 5 - 6 6 6 6 162 1 find additional sites of disease which are not apparent by 2 NR-LU-10. 3 [Slide.] 4 NR-LU-10 also, as noted by the sponsor, is not 5 indicated for differential diagnosis of suspected lung 6 cancer or suspected mtastass. Clinical trials have shown, 7 in their Phase I-II studies, that they can image other 8 tumors with NR-LU-10; non-small-cell lung cancer, primary 9 and metastatic tumors, also of breast, ovary, colorectum, 10 prostate, kidney and liver. 11 [Slide.] 12 - The sponsor had three objectives in the pivotal 13 Phase-III trial. The first was the comparison of NR-LU-10 14 FAB imaging with the current standard diagnostic imaging 15 tests to assess the utility of NR-LU-10 for staging of 16 patients with small-cell lung cancer. 17 The second objective was estimation of sensitivity 18 and positive predictive value of NR-LU-10. The third was 19 the evaluation of safety for NR-LU-10. 20 [Slide.] 21 The population in the pivotal Phase-III trial were 22 adults, newly diagnosed, histologically confirmed, small- 23 cell lung cancer. They had to have a least one known lesion 24 detected by physical exam or standard diagnostic evaluation. 25 They were to have no prior history of chemotherapy, MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 163 1 radiation therapy, or other investigational agents.2 [Slide.] 3 The treatment plan began with the standard staging 4 evaluation, physical examination, CT of the chest, head and 5 abdomen, whole-body bone scan and bone-marrow aspiration 6 and/or biopsy. 7 [Slide.] 8 The second step in the treatment plan was a single 9 intravenous administration of NR-LU-10 .FAB label with 10 Technetium. There was a maximum of 10 mg of the NR-LU-10 11 FAB fragment and a minimum of 5 mg; a minimum of 15 mCi and 12 a maximum of 30 mCi of Technetium-99, pertechnetate as the 13 label; total volume of administration is 15 to 20 ml, and it 14 was chelated with normal saline to get that volume; and it 15 was administered over 3 to 5 minutes. 16 [Slide.] 17 The imaging protocol in the treatment plan was 18 ideally to image these patients at 14 to 17 hours. Imaging 19 was regional, planar imaging; whole-body planar imaging, and 20 SPECT to the chest was optional and performed in many 21 patients. 22 [Slide.] 23 The interpretation of the scintiscans; the 24 original treatment protocol identified that all of these 25 studies would be evaluated at the clinical site unblinded. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 ron?) 164 1 A second step was all of the scintiscans were to be 2 evaluated in a completely blinded evaluation. Comparison 3 between the first and second steps were then to make sure 4 that they agreed. 5 If there was a disagreement, then they were 6 submitted to a third interpreter and that third interpreter 7 would evaluate those sites that were in question and then 8 render an unblinded third opinion. 9 The Center for Biologies, noting that this is 10 being proposed as a first-best test for evaluating patients il with small-cell lung cancer, have chosen to go with the 12 fully blinded interpretation for the evaluation of this 13 trial. 14 That evaluation; all scintiscans from all 15 investigational sites were moved to an independent 16 interpretation. There was no consultation between the 17 interpreters. They had no knowledge or access to the 18 clinical history, the findings of the standard diagnostic 19 imaging, nor did they know the technical imaging findings or 20 artifacts, which is significant when we talk about false 21 positives, especially. 22 [Slide.] 23 The treatment protocol for safety evaluation; 24 patient monitoring was first with observation and vital 25 signs. Next was serial laboratory evaluation with MILLER REPORTING COMPANY, INC. 507 C Street, N . E . Washington, D.C. 20002 ^ a c - dace. 165 1 hematology, serum chemistry and HAMA evaluations, and 2 urinalysis. 3 [Slide.] 4 The results of the pivotal study Phase-III trial; 5 first of all, the demographics and patient characteristics. 6 We started with 96 patients being entered on this trial. 7 89 patients were evaluable for efficacy. The mean age of 8 the population is 60.9 years ranging from 34 to 88 years. 9 77 percent of the patients being admitted on the trial were 10 m a l e . 11 [Slide.] ". 12 Seven patients who did not qualify for efficacy 13 analysis. One did not undergo gamma-camera imaging. One 14 did not complete the standard diagnostic studies. One 15 patient received chemotherapy prior to the antibody 16 injection. One patient had a second malignancy, a colon 17 cancer. Three patients had primary tumors, not small-cell 18 lung cancers. Two were non-small-cell lung cancers of the 19 lung. One was a small-cell carcinoma of the parotid. 20 [Slide.] 21 This summarizes the evaluation of the Phase-III 22 trial. We start with 89 patients in this trial and we 23 notice, on the left side, final stage being determined. The 24 final.stage is different from those patients in terms of the 25 standard diagnostic modalities. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 2000^. ,, ion-?], iic-cccc , 166 1 Seven patients plus 2, a total of 9, were 2 reevaluated, sometimes based upon the imaging from NR-LU-10, 3 other times in terms of re-review of bone marrow as well as 4 CT evaluations. Therefore, when we look across for the 5 final stage of extensive disease, we have 57 patients who 6 were staged as extensive disease, 32 patients who were 7 staged as limited disease, totalling the 89. 8 Looking on the bottom, NR-LU-10, extensive 9 disease, identified 47 patients as extensive disease. Of 10 these, they were correct on 44 patients as extensive 11 disease. They were incorrect on 3 patients as extensive ' 12 disease. These were overstaged and we will review those 13 initially for you in just a moment. 14 Looking at NR-LU-10 for limited stage, they 15 identified 42 patients as limited stage. They were correct 16 on 29 patients. They were incorrectiand-unde-rsJtaged ,,on 17 13 patients and we will review those. 18 [Slide.] ' 19 Sensitivity for NR-LU-10 imaging. Of 57 patients 20 with extensive disease, they identified 44 correctly, 21 77 percent sensitivity. All standard diagnostic tests; 22 those 57 patients. All standard diagnostic tests identified 23 50 for a sensitivity of 88 percent. Specificity; 32 24 patients identified with limited stage, NR-LU-10 imaging 25 correctly identified 29 for a specificity of 91 percent. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 . 167 1 All standard diagnostic tests; of the 32, they correctly 2 identified 30 for 94 percent. 3 [Slide.] 4 Comparative performance for extensive disease; 5 accuracy, PPV and NPV may vary with the prevalence of the 6 disease and the population. In our review, in the briefing 7 document, you will notice that CBER went forward and looked 8 at different population prevalences, at 60 percent, 9 69 percent and 75 percent to compare and see if we could 10 feel that the accuracy, PPV and NPV reported in this patient 11 population for the clinical trial could be reliable. ' 12 The population prevalence in this trial was 13 64 percent and we determined that the accuracy PPV and NPV 14 had minimal variations and, therefore, we could rely on the 15 values for this clinical trial as being representative. 1:6, . . _The_ accuracy for,NR-LU-10 imaging, 82 percent. 17 All standard diagnostic modalities, 90 percent. PPV for NR- 18 LU-10 imaging, 94 percent. All standard diagnostic 19 modalities, 96 percent. NPV, 69 percent for NR-LU-10 20 imaging, 81 percent for all standardized diagnostic tests. 21 [Slide.] 22 This is a comparative example of the accuracy of 23 single diagnostic modalities. You have seen this before. 24 It lists, on the left column, each of the individual 25 diagnostic modalities that were in this clinical trial. NR- MILLER REPORTING COMPANY, INC. 507 C Street, N.E. . Washington, D.C. 20002 168 1 LU-1G is at the'top. CTV abdomen is No. 2. And the others 2 are listed below. 3 Limited stage, the number correct as compared to 4 the total, is in the second column, extensive stage in the 5 third column, and accuracy in the fourth. As we see, 6 accuracy at 82 percent for NR-LU-10 is the leading accuracy 7 of all of the single standard diagnostic tests as compared 8 to NR-LU-10.. . 9 This is somewhat artificial. You are not going to 10 use the individual tests alone. But, from that standpoint, 11 it is the most accurate individual single test in this 12 trial. Extensive disease; it identified 44 of 57. The next 13 closest is CT of the abdomen, identifying 33 of 57. As a 14 comparison to that 82 percent accuracy for NR-LU-10, CT of 15 the abdomen demonstrated an accuracy of 71 percent. -16 .......- -- The- other diagnostic modalities -beiow. all decline 17 in accuracy and all decline in the number of extensive-stage 18 disease identified. Of note, in the limited stage; actually 19 NR-LU-10 was the lowest in terms of identifying limited- 20 stage correctly. 21 [Slide.] 22 This is a sensitivity, a secondary analysis, if 23 you will, of NR-LU-10 by organ evaluation. In the top 24 column, you will see the white listing. All of this area 25 above, all of these areas, .,,were identified with a detection MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002^. >*.**>\ I-AS- ' . 169 1 rate or sensitivity greater than 75 percent, in the lung, 2 neck, axilla, mediastinum, bone and bone marrow. 3 Liver and pleura demonstrated a detection rate or 4 sensitivity at 63 percent and 53 percent, respectively. 5 Noted is brain which has -a relatively insensitive 6 demonstration of 3 of 12 at 25 percent. And there were 7 9 metastatically involved adrenal glands in the trial. None 8 of them were detected by NR-LU-10. 9 [Slide.] 10 False positives. There were 3 patients who were 11 overstaged in a fully blinded evaluation. The interpreters 12 had no knowledge of the technical artifacts and these of 3 13 patients that were overstaged, one had a recent surgery 14 site, and Dr. Nelp demonstrated that scan to you. There is 15 no need to show it to you again. Indeed, it is positive 16 and-r- if you were -on the clinical site, you-would have-- - 17 identified that and easily detected it. 18 There were two skin folds in the axilla. This 19 product does demonstrate positive localization of the skin 20 folds. It should be recommended for this agent that always 21 the arm should be elevated. But they weren't in the trial, 22 as you saw in the picture. 23 In the opinion, therefore, in my review, all three 24 Qf these false-positive patients would have been properly 25 staged in a clinically competent or routine clinical ' MILLER REPORTING COMPANY, INC. ' 507 C Street, N . E . . Washington, D.C.. 20002 -im] cac-Cc.cc . 170 1 setting, so the false positives here were three. I feel all 2 of them are represented because of the fully blinded 3 evaluation. 4 [Slide.] 5- False negatives;'13 patients were understaged. Of 6 these 13 patients, 5 were because they did not identify 7 brain metastatic disease primarily. Remember that 3 of 12 8 that I showed you before. It will detect brain in some 9 patients but it is relatively insensitive. It is a whole- 10 body imaging technique and, as a result, it is not and it 11 cannot be expected that it is going to identify brain 12 metastasis with a high efficiency. ~ 13 4 patients were missed because of liver metastatic 14 disease which was not identified. 4 patients, 4 bone 15 metastasis, total of 13. All of these patients were 16 correctly.staged- and identified for the metastatic disease-- 17 at the same time by the standard diagnostic modalities as 18 tiiey were missed by NR-LU-iO in this evaluation. 19 [Slide.] 20 Review of the imaging. Dr. Nelp has given you an 21 excellent set of images. . *. I will go through and step through 22 what I consider to be the significant findings for, number 23 one, imaging and confirmed tumor sites; yes, NR-LU-10 will 24 identify tumor sites as demonstrated. ,, 25 Adequate target to nqn-target imaging resolution MILLER REPORTING COMPANY, INC. 507 C Street, N.E. , Washington, D.C. 20002 _ 't.a c . c c c c 171 1 is apparent in the studies. The earliest image tumor that I 2 could identify in the scans that were submitted to me was at 3 3 hours and the latest were 24 hours. 4 [Slide.] 5 Non-tumor localization is in three areas in these 6 studies, and he has reviewed those for you, also. First of 7 all, the clearance routes and we will step through those in 8 just a moment. Antigen-specific localization; remember, we 9 are working with an antigen-antibody response and non 10 antigen-specific localization. 11 [Slide.] 12 Clearance routes were two. The first and the 13 predominant is through the kidneys and the urinary bladder. 14 It is seen as early as 3 hours in the images and it is 15 intense through the 24-hour set of images. Gall bladder, 16 smalt bowel -and- large- bowel and liver are a second- route of 17 clearance. It is significant with approximately one-third 18 of the clearance going this route. It is seen as early as 19 3 hours through 24 hours. 20 [Slide.] 21 Antigen-specific localization; salivary glands, 22 thyroid and adenohypophysis have all been demonstrated on 23 imaging. A fourth is also known in terms of 24 immunohistochemistry and that is the pancreas. We were not 25 able to identify pancreas localization on the scans, but MILLER REPORTING COMPANY, INC. 507 C Street, N.E. . Washington, D.C. 20002 (202V 546-66S5 . 172 1 this may well be because of the overlying bowel in the area 2 that we were not able to define the structure. 3 [Slide.] 4 Non-specific localization; typical for 5 nuclearmedicine-type imaging. Normal structures in the 6 heart, vascular structures, nasopharynx, spleen, testis and 7 skin folds. Inflammatory responses; we have noted the 8 recent surgical site that you have seen. I would anticipate 9 that you are going to see this agent in injection sites as 10 common with any type of Technetium-labeled material. 11 I would also anticipate you are going to see this 12 in inflammatory responses and you have to be concerned that,, 13 with any type of agent that sees inflammatory responses, the 14 potential for abscess localization is there, too. Neither 15 of those last two, though, were identified in the clinical 16 trial for us. ,, ......- .-- . 17 [Slide.] 18 * * Non-tumor localization. NR-LU-10's ' interpretation 19 will be of limited value for detection of potential 20 metastatic disease in regions of non-tumor localization. 21 The extrahepatic abdomen presents one concern in terms of 22 the amount of activity clearing through the bowel as well as 23 in the region of the kidney and bladder, remembering those 24 adrenal glands; they were 0 for 9 in terms of identifying 25 metastatic disease. ' MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 173 1 [Slide.] 2 Integrated safety summary; 515 single patient 3 intravenous doses reviewed. These are in multiple trials in 4 multiple different types of tumor. Their experience from 5 those 515, 30 patients, 6 percent showing minor adverse 6 events. 7 [Slide.] 8 The safety summary in terms of those events; 9 clinical events, mild and transient, fevers in 2 patients. 10 Allergic reactions in three patients. Laboratory value 11 increases, low-level and transient. Amylase in 10 patients 12 and lipase in 13 patients, remembering this agent is 13 targeting salivary glands and pancreas. And whether one or 14 both of these organs are demonstrating slight elevations in 15 amylase and lipase would be determined. 16 -- - Thyroid-function studies were elevated-and changed . 17 in 2 patients transiently, remembering also that this 18 antilbody targets specifically: the" thyroid and there is some 19 change. No adverse events were considered of clinical 20 significance. 21 [Slide.] 22 HAMA response evaluated in 53 of 89 patients in 23 the pivotal, Phase-III trial. Of those 53 patients that 24 were evaluated, 3 patients developed elevated HAMA tests. 25 Of those levels, one was a two-fold increase over baseline. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. ' Washington, D,C. 20.003- - 174 1 One was a six-fold increase over baseline, and one was 40 2 fold increase over baseline. 3 The 2 patients with two-fold and six-fold both 4 returned to the baseline within 4 months. The third patient 5 was declining but did not return to the baseline at 4 months 6 and was lost to follow up. 7 [Slide.] . 8 Dosimetry; total body radiation exposure is 0.47 9 RAD/30 mCi. The kidney exposure, our first route of 10 clearance, is 3.9 RAD/30 m C i . Our gall bladder, second 11 route of clearance, 5.6 RAD/30 mCi. And the thyroid, 0.91 12 RAD/30 m C i . From this standpoint, all of these are within 13 acceptable limits with Technetium products and represent no 14 concern in terms of a medical or clinical evaluation. 15 [Slide.] 16 : -- . Summary points. Tech NR-LU-10 FAB whole-body - 17 imaging identified extensive disease in 44 of 57 patients 18 for a sensitivity of 77 percent. Those patients had a PPV 19 of 94 percent. There was a low rate of false positive 20 interpretation in 3 of 47, and I have reviewed those false 21 positives for you, in terms of being probably representative 22 from the fully blinded evaluation. There were 13 of 57 23 patients understaged. 24 - [Slide.] 25 , NR-LU-10 images small-cell lung cancer, other MILLiER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 . 175 1 malignancies, inflammatory sites at recent operative site, 2 we know from this trial; benign neoplasms, as we know from 3 immunohistochemistry; and normal tissues. Readministration 4 has not been evaluated for safety or efficacy. And the 5 performance characteristics of NR-LU-10 who have received 6 anti-neoplastic therapy, chemotherapy and/or radiation 7 therapy, are unknown. 8 [Slide.] 9 NR-LU-10 FAB imaging is safe for clinical use with 10 a 6 percent adverse-events reported in 515 single-patient 11 injections and no adverse events were clinically ' 12 significant. 13 That concludes the pivotal study Phase-III trial. 14 DR. PARKINSON: Thank you, Dr. Mills. I think it 15 is appropriate at this point to ask if there are any 16 questions either for Dr. Berkower or for Dr. Mi3rls. 17 , DR. GELBER: In the pivotal study, could you . - * 18 comment on the determination of the final stage? How was 19 that actually achieved? 20 DR. MILLS: In terms of the final stage whereby 21 standard diagnostic modalities -- and, again, there were 22 9 patients who were actually rereviewed on bone marrow as 23 well as CT, sometimes being led back by the NR-LU-10, sometimes by other clinical findings, that, were final stage changed from the initial on-study staging. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 t'702) 5 4 6 - 6 6 6 6 176 1 DR. GELBER: I see -- because your comment, about 2 using the test study to go back over caught my attention. 3 DR. MILLS: It caught my attention, too, in terms 4 of when we initially started reviewing this that you will, 5 from time to time, find sites on NR-LU-10 during that 6 clinical trial which would refocus their attention and go 7 back looking at CT or looking at bone marrow. So, from that 8 standpoint, while we consider the standard diagnostic 9 modalities to be a gold standard, it may have a slight tinge 10 of brass to it in terms of interpretation. 11 DR. GELBER: Just to push that a little bit : 12 further; if you didn't use the evidence from the agent we 13 are evaluating to restage, what would the parameters of the 14 test performance look like? 15 DR. MILLS: In terms of if we had not done that? 16 Actually:, when we compared, when we were looking at the 17 sensitivity, we were looking the standard diagnostic 18 modalities on stage evaluation for that comparison of 19 sensitivity and specificity. 20 DR. GELBER: So it wouldn't change it. 21 DR. MILLS: No. We were comparing that because we 22 wanted to identify final stage in terms of the outcome for 23 it. But when we compared sensitivity and specificity, we 24 used the on-trial staging as initially being evaluated. 25 DR. GELBER: So, did you recalculate positive MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 ' (202) 546-6666 177 1 predictive value, then, using sensitivity and specificity 2 and putting in a prevalence, because the 94 percent, I 3 assume, comes out of the data in the table. 4 DR. MILLS: That is correct. 5 DR. GELBER: Rather than a recalculation. 6 DR. MILLS: That's correct. We estimated that 7 prevalence rate only to make sure with this trial, because 8 we had 89 patients, that the prevalence would not move those 9 accuracy PPV and NPV values. 10 DR. GELBER: And your lowest prevalence for that 11 sensitivity was 60 percent? 12 DR. MILLS: We attempted 60 percent, 69 percent 13 and 75 percent. 14 DR. GELBER: Was the 60 percent a reasonable lower 15 limit for the portion of patients with extensive disease? 16 , -L. DR. MILLS: We felt it was. If someone, in terms 17 of difference, in terms of that -- but we felt that those 18 would extend to any type of prevalence rate that we would 19 expect to see clinically. 20 DR. GELBER: And a last question about the study. 21 The 7 patients who were excluded. Did you go back and take 22 a look at what their small-cell lung-cancer stage would have 23 been? 24 DR. MILLS: No; I did not. From that standpoint, 25 they were excluded from i t . I reasonably went back and MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002,^ (202) 545-6666 178 1 evaluated each one of them in terms of the reason for 2 exclusion to make sure there was a reasonable reason to move 3 those patients off. 4 DR. INGLE: For the indication, the positive 5 predictive-value indications sought -- the positive 6 predictive value is very important. What you are telling us 7 is that if you have the clinical material available, and the 8 way it would be used was approved, that, in fact, the 9 positive predictive value is 100 percent; is that right? 10 That is what I am assuming from the data; that is, 11 there were 3 false positives that, in fact, if you had the 12 clinical data, you would have called them negatives. 13 DR. MILLS: I would have. 14 DR. INGLE: So your positive predictive value, 15 with all the clinical information, is 100 percent. That 16 almost sounds too good to be true. ~ '- - 17 DR. MILLS: As a nuclearmedicine imager, it is too 18 good to be true. But, from that standpoint, in terms of a 19 fully blinded interpretation, the three false positives were 20 correctable in terms of standard clinical evaluation. It is 21 a remarkable one in terms of being able to do extensive 22 disease evaluation. It is, obviously, not, in terms of 23 limited disease. 24 DR. INGLE: That is correct. But they are not seeking that as an indication. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) .546-6666 - 179 1 DR. MILLS: That's correct. 2 DR. INGLE: I know that the studies were blinded. 3 Did you review the studies that were called negative? 4 DR. MILLS: Yes. 5 DR. INGLE: You have got uptake from the neck and 6 the head and down into the abdomen and the bowel, et cetera 7 -- well, thank you. I feel much better. 8 DR. MILLS: From that standpoint, number one, is 9 we did review all 89 scans in terms of being able to 10 determine if we could evaluate that. Number one is that if 11 you are looking for extensive disease, when you are looking 12 for multiple organ involvement, this study does quite well. 13 When you see that much activity, limited-stage disease 14 presents a problem for you in terms of determining it. 15 Looking at the abdomen -- and one of the concerns 16 -I have about this study is it clears through into the small 17 bowel, large bowel, and shows quite a bit of activity over 18 the abdomen. What that did, in terms of the evaluation, was 19 that the interpreters became insensitive in that area and 20 actually under-read adrenal areas. 21 So, from that, standpoint, the experienced 22 reviewer, if you will, became very cognizant of those areas 23 that were of concern and down-read them and, so, tended to 24 understage patients. . 25 DR. INGLE: Just one last question, if I could. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 - (2 0 2 ) 180 1 Whenever you have a test that has 100 percent positive 2 predictive value, it opens the door for basically 3 overstaging somebody; that is, you gave the example that you 4 would expect it to be taken up in benign lesions such as 5 abscess. 6 DR. MILLS: Absolutely. 7 DR. INGLE: Any thoughts about suggestions if it 8 was used in positive as a part of clinical practice that one 9 uses, focuses in one lesion to prove metastatic disease with 10 a more conventional diagnostic test, anything to really 11 protect patients so that you don't overstage somebody? \ 12 DR. MILLS: I think what will happen with this 13 study - 14 DR. INGLE: That's what will happen. 15 DR. MILLS: Absolutely. They will understage 1-6 -these -patients,- characteristically, and that---is what you saw 17 in this trial evaluation. There is so much activity in the 18 field and you know where the positive areas are going to be 19 which are going to confuse you, that abdomen, that activity, 20 both kidneys. 21 So what we see is that the interpreters were 22 actually understaging patients. So what is going on is that 23 they have a degree of confidence in the extensive. They were able to call that. DR. INGLE: One point regarding that that I am MILLER REPORTING COMPANY, INC. 507 C Street, N .E. Washington, D.C. 20002 (202) 546-6666 . 181 1 sure will be reflected in the labeling is that although one 2 might argue that no incidences of false positives were 3 observed, when one looks at the confidence interval around 4 either specificity or PPV, given that you are talking 30 or 5 40 people in the group, even with no observation, the 6 confidence interval for specificity drops down as low as 7 about 90 percent, meaning that 10 percent of patients with 8 limited disease may be positive and it just may not have 9 shown up on this. 10 Similarly, if you look at the PPV, the confidence 11 interval could be as low as 92 percept. So I don't think, 12 the data yet are strong enough where, even based on these 13 readings, one would write a labeling saying they are not 14 going to occur if this is approved. 15 DR. MILLS: To offer to you, the same way. My 16 consern of having, dealt with nuclearmedicine- images fo-r 17 many, many years is that nothing has 100 percent. But, from 18 the standpoint of looking at this, the natural tendency of 19 the reviewer who is experienced and who has been trained 20 will be to under-read these images, not over-read these 21 images. 22 DR. BARRY SIEGEL: I certainly agree that nothing 23 in nuclearmedicine has 100 percent with confidence. Dr. 24 Gelber, I think, was trying to get at an issue of 25 verification bias as it entered into the assessment of the MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 ' (202) 546-6666 182 1 final extensive versus limited disease classification. I 2 think that is an important problem. But there is another 3 twist to it and that is, since, the data were available at 4 the clinical site, since the scans were read and it was an 5 open-label study at the clinical site, do you all have a 6 feel, or has the sponsor presented data, as to what extent 7 verification bias operated in the actual clinical 8 interpretation of the standard diagnostic tests and prompted 9 additional workup? 10 How often did the scan results lead to something 11 that wouldn't have been done in routine clinical practice'or 12 change the interpretation of the CT scans? 13 DR. MILLS: We know that there were 7 patients who 14 had CT re-reviewed in terms of that interpretation, if you 15 wanted to draw th a t . We did not come back in terms of 16 additional ---I'm seeing -- Dr. Salk is at "the ready to give 17 me some input. As soon as we can get him seen, he is going 18 to tell u s . 19 DR. BARRY SIEGEL: Part of the point of the 20 question is that, in a way, the design of this trial is 21 stacked against NR-LU-10 and stacked in fayor of the 22 conventional diagnostic test. 23 DR. JAY SIEGEL: It is my understanding, though, 24 that if a CT was re-read and the CT or biopsy are ultimately 25 confirmed disease, the reading for the standard diagnostic MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 183 1 test, however, was the initial reading not the re-read - 2 DR. BARRY SIEGEL: But even the initial reading - 3 DR. JAY SIEGEL: Might have been informed - 4 DR. BARRY SIEGEL: If the NR-LU-10 showed a lesion 5 in the left lobe of the liver and the initial CT scan showed 6 only a subtle hypodense lesion of the left lobe of the 7 liver, the CT interpreter, knowing that NR-LU-10 was 8 positive, might well have called that a metastasis that he 9 would have skipped otherwise. 10 DR. JAY SIEGEL: I think they were done 11 sequentially, though, where the standard tests were all done 12 before the NR-LU-10, at least that was a protocol. 13 DR. SALK: Yes. Almost invariably, all the 14 standard tests were done before the Verluma imaging. There 15 were a very few exceptions that were made because of 16 scheduling reasons. - ~ 17 ,, 18 Can you get me Slot No. 42. * * DR. MILLS: By the way, just as you are getting 19 your Slot 42, one of the things to understand is that we, 20 indeed, stacked this against NR-LU-10 as stringently as we 21 could in terms of making sure there was no bias introduced 22 in terms of knowledge from the site. That is why I 23 specifically wanted a fully blinded interpretation with this 24 study. 25 We have data that we were originally presented in MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 184 1 the i-rial in terms of the on-site. We did not want to have 2 that influencing because of the performance characteristics. 3 DR. SALK: I want to show you two slides that I 4 think will help to see what is going on here. One of them - 5 - 22 will come after this one. 6 [Slide.] 7 This shows you the previously unsuspected true 8 positives. In other words, these were the false-negative 9 lesions, organs or patients, those three different levels of 10 analysis, by the traditional imaging test, by the standard 11 evaluation. These are all of them in the study. 12 Among 610 lesions, there were 39 lesions that were 13 seen in NR-LU-10 images that had not been previously seen, 14 and raise the question of whether they were real or not. In 15 contradistinction to what you said, Dr. Mills, in fact, all 16 of these were first seen by the Verluma imaging". 17 Because they were seen, we went back and either 18 reread the baseline CT, obtained another test, went and did 19 an MRI, and, in one instance where we found a positive bone 20 marrow in the Verluma images, went back and looked at the 21 original bone marrow and found abnormal cells that had been 22 missed previously. 23 Those were the only instances in which, when it 24 was corroborated by a standard modality, most of the time 25 that was because it just didn't happen to be seen in the MILLER REPORTING COMPANY, INC. 507 C Street, N . E . Washington, D.C. 20002 (7 0 7 ) 185 1 first reading, there were only 6 percent of those in the 2 entire study in the lesions. At the organ level, among 333, 3 only 5 percent and, at the patient staging level, the 7 4 patients that Dr. Mills has been referring to were the 5 unsuspected positives. 6 [Slide.] 7 This shows the detection rates for positive 8 disease, the percentage detection of lesions, affected 9 organs or patients with extensive disease at the three 10 different review levels. The patterns are similar at all of 11 these. Let me just stick with this one. : 12 The orange one is what we referred to as R-l. 13 That is the primary investigator at the clinical site. This 14 is the person for whom we did not control blinding. He had 15 access to everything. One or two investigators chose to 16 read everything blinded, but we had no control or any 17 information about that. 18 * All of the images were then sent, submitted, to 19 NeoRx and were read, fully blinded, as I said before, by 20 either Dr. Help or Dr. Grief. They alternated odd and even 21 cases. That reading is referred to as the R2 reading. That 22 is the one we report to you here today. As you see, those 23 blinded reviewers tend to be more conservative in the 24 reading of the images. 25 As Dr. Mills mentioned, in order to try the MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002* (202) 5 4 6 - 6 6 6 6 186 1 protocol design for this review process, it was fairly 2 complicated. In order for us to try and define something 3 unlike an artificial blinded situation, it would give us a 4 more accurate prediction of what the experience would be in 5 clinical practice. 6 It, therefore, went to when there was difference 7 between the first and second reviewers, it went to a third 8 reviewer which actually was fully blinded, initially, in 9 order to make a final determination of the stage. We did 10 add an extra step related to unblinding in order to identify 11 those three cases in which technical artifact played an 12 impact, put that back in. 13 So this, what we referred to as our final blinded 14 analysis, is the result of that blinded review process. 15 This was designed this way in the original protocol in an 16 attempt to make a better prediction of what would happen-in 17 practice. As you can see, all of these are very similar to 18 each other and the pattern is that a reviewer on site with 19 all information tends to identify more disease. A fully 20 blinded reviewer is more conservative. 21 When you put in technical information and other 22 things, and you increase uniformity -- because this, again, 23 was done by two reviewers centrally as opposed to the 24 variety of reviewers with a variety of different equipment 25 throughout the many sites. When you increase uniformity, MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 5 4 6 - 6 6 6 6 187 1 you get this experience. They are all very similar to each 2 other. 3 DR. GELBER: I don't know how to summarize this. 4 I am just a little confused about the seven cases. I don't 5 know how it would turn out. You mentioned that this . 6 analysis you did was conservative, if you will. You stacked 7 the deck against i t . 8 On the other hand, part of the information we are 9 presented is the hierarchy of what the individual studies 10 can produce with respect to the different sensitivities and 11 accuracies. It is against those that people are going to12 carry away the message, "This is much better." I haven't 13 done all the math and so on, but, if out of the 44 patients 14 that are in the cell and table that are extensive disease, 15 final stage, plus extensive disease by the test, seven of 16 -those got -into that cell because the test stimulated a 17 rvaluation of previous studies based on CAT scan. 18 Therefore, when you look at the results of 19 sensitivity, specificity here and you compare it to the 20 results if you look at other types of studies that were done 21 you might suspect, because you are using that information to 22 define your final stage, you could, again, get a better test 23 performance for the test agent that you are using to 24 stimulate. 25 That" is just my first feeling. On the other hand, MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 54-66G6 188 1 you might say, "Well, there really was a gold standard of 2 disease present and this agent was better able to pick it 3 up." So my question, really, is, besides looking at the 4 images, was there any other further evaluation of the status 5 and what summary numbers can we take away with respect to 6 sensitivity, specificity for the different tests given that 7 you are using your test to classify the final stage. 8 How do we get around that issue to report the 9 characteristics? 10 DR. MILLS: Remember, the sensitivity and 11 specificity are based upon the initial evaluation by the 12 standard diagnostic modalities that we presented to you, not 13 the final stage numbers. 14 DR. SALK: Let me also comment. You have 15 identified something which is a problem for the evaluation 3:6 of every-diagnostic test we have available' for which -- - 17 DR. GELBER: I don't mean to pick on you. You 18 just happen to be up there. 19 DR. SALK: -- for which a true gold standard 20 doesn't exist. Particularly in this disease, with small 23 cell lung cancer, you can't do biopsies of the entire body 22 and slice it up and know exactly what all of your -- it is 23 difficult to get that past a review committee. 24 So this is, indeed, a problem that one has. It is 25 uniform for all studies like this. Our feeling is that what MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (yen) 189 1 we did was we tried to define the best truth that we could 2 and to use that in order to come up with a reasonable 3 prediction of what the experience is liable to be. 4 DR. GELBER: I guess the bottom line comment is it 5 would be useful, because people will be using the numbers 6 that are reported here in a comparison way, to perform some 7 type of a sensitivity analysis. For example, you did one on 8 the prevalence. That was very nicely presented and an 9 important question to ask. 10 But, also, with respect to some of the other 11 statistics that come out of the study, with respect to how 12 you classify -- in addition to the 95 percent confidence 13 intervals, take a look at what those numbers would be if you 14 used other kinds of evaluations. 15 DR. GELBER: I think it is very interesting 16 suggestion. I- am going to suggest that you-- get together 17 with Dr. Blumenstein afterwards and discuss it because we 18 have spent many, many hours discussing this kind of thing. 19 I terms of making comparisons with other agents, remember, 20 the other agents are also tested within the same kind of 21 setting where the gold standard doesn't exist and may be 22 influenced. 23 DR. JAY SIEGEL: I think part of that question, 24 which I didn't hear answered, is, on those 7 patients who 25 were, at least by initial standard diagnostic modalities MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 190 1 negative but wound up positive, do you know how many of 2 those were histologically confirmed as positive as opposed 3 to confirmed by rereading the standard methodology? 4 DR. MILLS: We had two of them. Bone marrow. 5 DR. SALK: I think, perhaps, two. It was a small 6 number and those were, like, bone marrows for which you can 7 do histology in this disease or biopsy. 8 DR. MILLS: The others were reinterpretation of 9 scans which were quite apparent. 10 DR. PARKINSON: I think we may get additional 11 clarification. Could you please use the microphone and 12 identify yourself, please. 13 DR. SATISH MISHRA: Satish Mishra from CBER. 14 Regarding the sensitivity analysis, I looked at the numbers 15 very carefully. There were 8 persons who were identified as 16 extensive by physical examination, alone.. Two patients were 17 identified as extensive by chest X-ray, so making up 18 lCf percent identified as having extensive disease by 19 physical examination and chest X-ray alone. And they were 20 identified as extensive by other tests like this - 21 DR. OZOLS: That is an important point that we 22 keep coming back to. If you take the test of a single test, 23 the Verluma test, and you get 44 patients who were 24 identified out of the 57 as extensive stage, that is very 25 impressive. But, if you take into account that everybody is MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 f?n9' KAC-CCCC 191 1 going to have a physical examination, everybody with small- 2 cell is going to have a chest X-ray, and then you compare 3 those 8 patients with physical examination, 2 patients with 4 chest X-ray, to CT of the abdomen, you are left with 5 43 patients. 6 So, in the real world, is this test any better 7 than a chest X-ray, a CT scan of the abdomen and a physical 8 examination? 9 DR. MISHRA: This analysis I did not do. But if 10 we exclude those persons, maybe we could get some number 11 indicative. I think; that the numbers will not change that 12 much. That is my fueling. But, again, sample size, too, is 13 small. That is one comment. 14 Another comment I wanted to make was the on-site 15 reviewer had identified two different patients than the 16 blinded reviewer as having false false positives-. So there 17 was some reader-to-reader variability seen also in this. , 18 ^& DR. SALK: Dr. Ozols, I have a series of slides 19 which may help to answer that question. 20 DR. OZOLS: I would actually like the FDA's 21 opinion on that. 22 DR. SALK: Oh; I'm sorry. If you want to know the 23 truth, it is the slides that were prepared by Dr. Mills. He 24 asked us to make them. 25 DR. MILLS: Why don't you go ahead and bring them MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washi/n~gr\too\n,era D.C. c cc c 20002 c 1 up. I anticipated what we were going to have. 192 2 DR. SALK: L-l through L-6; that's 46 through 51. 3 DR. PARKINSON: Those are the contract slides. 4 DR. SALK: We went to the trouble to make them for 5 him and he didn't show them, so I wanted to make sure they 6 were here. 7, DR. MILLS: It is good grist for the response, 8 now. Go for it. 9 [Slide.] 10 This is, again, a similar diagram to what I showed 11 you before and that is the number of diseased organs here 12 and the different organ systems down here. The top of the 13 bar represents the total number of organs, effected organs, 14 in the study. And the orange indicates the number of those 15 organs that were detected by, in this case, chest X-ray or 16 CT of the chest. ' 17 ; 4 18 [Slide.] - ** The next one shows what happens if you move on to 19 CT of the abdomen. 20 [Slide.] 21 Bone scan. 22 [Slide.] 23 CT of the head. 24 [Slide.] 25 And when you add all of those together, this is MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 1707) RAfi-KfiKK ' 193 1 the pattern that you get. You will notice that the bone 2 marrow and neck and axillary lymph nodes are not marked very 3 much. So, after stepping through that whole series, you end 4 up with this information. 5 By comparison with the Verluma imaging, with a 6 single test, this is the information which you get including 7 evaluation of the bone marrow in a non-invasive manner and 8 either confirmation or new identification of neck and 9 axillary lymph nodes which are otherwise not well 10 identified. 11 DR. MILLS: One of the more interesting things of 12 these tables is to identify that NR-LU-10 imaging is the 13 second best test in each organ system. Its constellation of 14 findings is where it is able to perform for extensive 15 disease. Its limitation, therefore, is why it is limited in 16 limited staging. ' 17 DR. PARKINSON: Are,there additional questions *' ' r - 18 either for the FDA or Dr. Salk? * 19 DR. MALCOLM: We were just discussing the fact 20 that, in clinical practice, many times, individuals get 21 chest CT scans which include the abdomen. The question we 22 wanted to get clarified here; were these actually separate 23 tests? What I am saying was a CT scan of the chest and then 24 a separate CT scan of the abdomen - 25 DR. PARKINSON: Or was it a CT to include MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002-" ' 194 1 adrenals. . 2 DR. MALCOLM: Right. 3 DR. MILLS: I can't confirm if every one of them 4 was divided separately. We had reports but whether they 5 were done at the same setting or as a continuous, that I 6 cannot tell you for sure. Depending on the clinical 7 practice site as to whether or not they are going to 8 continue right on down. .... 9 DR. MALCOLM: I would beg to venture that the 10 majority of these patients had a CT scan of their chest 11 which included the upper abdomen which is the standard way 12 to do it and the patient did not undergo a separate CT scan 13 of the abdomen. 14 DR. SALK: Actually, remember this study was 15 performed back in 1988 and that was not the routine practice 16 every place. So, in many, many, many -instances,''it""was an 17 independent CT of the abdomen that had to be performed. 18 DR. BOTSTEIN: I wanted to ask the sponsor 19 something. If this were a drug, we would be asking you to 20 have two adequate and well-controlled clinical trials, the 21 second one replicating, scientifically, the findings of the 22 first. So really the only exception to that would be if a 23 product were of such major medical importance that you 24 couldn't go and do the second trial once you had the results 25 of the first trial. - MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 195 1 So what is different here in the requirements for 2 a biological, or in the scientific setup, that makes you 3 think one study is enough? This isn't a situation where you 4 could look at the findings from separate study sites and say 5 they replicated each other because there aren't enough 6 numbers for you to get sensitivity and specificity and all 7 that sort of stuff. 8 DR. SALK: Actually, Dr. Mills, would you like to 9 comment on that? 10 DR. MILLS: I was going to say Dr. Siegel - - a t 11 that level, will discuss that. ' 12 DR. JAY SIEGEL: I think one of the issues here is 13 that this isn't an issue of hypothesis testing, of looking 14 for a p value to reject the null hypothesis. This is a 15 issue of estimation -- two things; estimating past 16 performance in terms of sensitivity, specificity, predictive 17 values, accuracy and so forth and, also, estimating utility 18 in a clinical setting. 19 I would argue, in biologies in general where we 20 obviously are very interested in replication or in how 21 convincing the data are--- we don't have quite as, perhaps, 22 consistent a view as to the multiplicity of trials, but I 23 think in a case where you are trying to get estimates of 24 statistical values, that one can certainly make, as opposed 25 to hypothesis testing, the case that one can get more MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 " . (^ r\n\ v.a c _c c c.c 196 1 precise data from one larger trial than from a larger number 2 of smaller trials each yielding rather inaccurate data. 3 It is, of course, important, I think, to look at 4 the data that insure that there aren't local either biases 5 or techniques that may have caused some sites to perform 6 differently from others. Although that hasn't been 7 addressed, that is a routine part of our review. 8 DR. SALK: Our experience at NeoRx has primarily 9 been with biologies rather than drugs. This obviously was a 10 question that we discussed pre-Phase-III, early on, about 11 the design of the study. The resolution was to do it the- 12 way we did. Our experience in looking at other biologic 13 products is that many of them are also handled that way. 14 So your general question of whether drugs are 15 different from biologies and why, I am not going to attempt 16 "to answer.- " "' ~ " 17 DR. PARKINSON: Is there data on performance at 18 different sites? I guess it was unremarkable enough that it 19 20 DR. MILLS: There were 24 sites. We did not see a 21 .difference in terms of that remarkable presentation. But, 22 remember, you have very small experiences at any one site 23 with 89 patients. 24 DR. SALK: The other important thing is that all 25 of these images were ready in one place by one pair of MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 iom'i r,-cc.cc 197 1 blinded reviewers so that the variability from site to site, 2 et cetera, is eliminated by the review process. Unlike some 3 other studies where either none or a sampling of images are 4 read by a blinded way, we read 100 percent of the images in 5 this blinded process. '~ 6 DR. BARRY SIEGEL: That doesn't actually address 7 the clinical problem in the real world because unless you 8 have Dr. Nelp read all the scans, you can't be sure -- and 9 that would keep him busy for a long time to come -- but you 10 can't be sure that you can replicate this level of observer 11 performance at multiple sites. 12 You have the opportunity with at least several 13 sites to look at the data because you have several sites 14 that have large enough n values to try to get some estimate 15 of lesion sensitivity and specificity. You can't do it on a 16 ptirent basis. -~ ~ " 17 DR. SALK: It, certainly, potentially, could be 18 done. The experience we actually had in training physicians 19 for the trial was a session, not unlike what Dr. Nelp did 20 for you here today. They were either done personally or, 21 sometimes, with a book. We sent a book of ijnages with some 22 descriptions. A nuclearmedicine physician read through 23 that, looked at the normal biodistribution, saw a few 24 examples of the types of artifact, and they were very 25 comfortable. MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 198 1 There was no obvious -- nothing jumped out 2 obviously as people having particular problems at one site 3 compared to another. So, practically speaking, it was not 4 apparent. 5 DR. PARKINSON: Can I now consider myself 6 certified to be able to read these scans? 7 DR. SALK: Not necessarily, but you can be 8 comfortable. .... 9 DR. BARRY SIEGEL: You can't be licensed to 10 possess the radioactive material, but you can be certified 11 to read the studies. That is an NRC -- 12 DR. PARKINSON: There are people who could be 13 certified immediately. 14 DR. TOM BULL: I am Tom Bull. I am with CBER. IN 15 addition to the reading of these images that they have 16 mentioned today, we also had three radiologists at CDER, the 17 Center for Drugs, read these images, read 22 of them. They 18 were Dr. Lionel Lieberman, Dr. David Woodbury, and Dr. Eric 19 Jones. George has looked at 100 percent of them, and your 20 assistant has looked at 100 percent of them. 21 So we have had five radiologists within the Food 22 and Drug Administration look at them and they are all 23 getting the same answers. So I think that is a fairly good 24 certification that the reading results can be replicated. 25 DR. GELBER: Did you also review all the CAT scans MILLER REPORTING COMPANY, INC. 507 C Street, N.E. Washington, D.C. 20002 (202) 546-6666 199 1 that were done? ' 2 DR. MILLS: From the FDA's standpoint, no, we did 3 n o t . We have all the source documents for all of the 4 interpretations from them, but we did not actually review 5 all of the CT scans. From that standpoint, we would have to 6 have a record room fairly extensive to do that in any of 7 these trials. 8 DR. GELBER: It, again, goes to the issue of the 9 comparability - 10 DR. MILLS: Absolutely. From our standpoint, we 11 have to rely on the on-site interpretation of the comparable 12 gold-standard studies. 13 DR. SALK: And, as you said, we stacked the deck 14 against Verluma by using the routine evaluations on site 15 rather than attempting to do a very, very difficult thing 16 which would be to make uniform all of the other "tests as 17 well and to review those centrally. * 18 -* DR. GELBER: I guess that last comment is the one 19 that I haven't -- I am going to need more time to convince 20 myself that that is the way the deck has been stacked 21 because if one can conjecture, with 24 sites, maybe three 22 patients from each site, studies done in '88, '89, in terms 23 of the uniform reading of the other diagnostic tests against 24 a test that has been reviewed and reviewed and reviewed by 25 experts in the field who train themselves and are really MILLER REPORTING COMPANY, INC. 507 C Street, N .E. . Washington, D.C. 20002 . >'no. 1 sAfi-ccs*; 200 1 clued in to evaluating that scan, and then.to look at the 2 performance characteristics of that procedure against the 3 performance characteristics of detecting extensive disease 4 at local institutions. 5 That is the standard that we are comparing against 6 and I am just a little uncomfortable. 7 DR. SALK: Remember, once again, when I showed you 8 the Verluma experience at the sites compared to the central 9 blinded reading, at the sites, they were higher. I would 10 expect that, in general, an unblinded reading, for instance 11 of the traditional imaging tests, is going to be higher than 12 a blinded reading. 13 So this is why I feel that we stacked the deck 14 against Verluma by that method of analysis. 15 DR. DEVOUS: I would reinforce that. In 16 diagnostic imaging, in general, blinded readings always do 17 worse, often substantially worse, then informed ones. So, I 18 understand that question you are asking, but, 19 experientially, the notion of giving a lot of clinical 20 information to somebody frequently makes them more accurate. 21 So, in this particular case, while that is an 22 appropriate challenge in a diagnostic imaging setting, I 23 think I would agree with his perspective that this is an 24 anti-Verluma design. 25 DR. GELBER: One last question related to that, MILLER REPORTING COMPANY, INC. 507 C Street, N . E . Washington., D.C. 20002 (20:2) 54 6-6666