8:00 a.m.

Tuesday, September 11, 2001














Versailles Ballroom

Holiday Inn - Bethesda

8120 Wisconsin Avenue

Bethesda, Maryland




Associate Clinical Professor

Oncology Associates, P.C.

Helen & Harry Gray Cancer Center

Hartford Hospital

85 Retreat Avenue

Hartford, Connecticut 06106


Advisors and Consultants Staff, HFD-21

Food and Drug Administration

5600 Fishers Lane

Rockville, Maryland 20857


Professor of Medicine

Division of Hematology/Oncology

Loyola University Medical Center

Cancer Center, Room 109

2160 South First Avenue

Maywood, Illinois 60153


Wilshire Oncology Medical Group, Inc.

50 Bellefontaine Street, Suite 304

Pasadena, California 91105


Professor of Medicine

Division of Hematology and Oncology

University of Alabama at Birmingham

1530 3rd Avenue South

Birmingham, Alabama 35294-3280


Professor of Biostatistics

Department of Biostatistics and Bioinformatics

Box 3958

Hanes House, Room 219

Trent Drive at Erwin Road

Duke University Medical Center

Durham, North Carolina 27710

ATTENDEES (Continued)



Chief, Gastrointestinal Oncology Service

Memorial Sloan-Kettering Cancer Center

1275 York Avenue

New York, New York 10021


Professor of Medicine and Cancer Prevention

The University of Texas M.D. Anderson Cancer Center

Department of Clinical Cancer Prevention

1515 Holcombe Boulevard, HMB 11.192c, Box 236

Houston, Texas 77030

JODY L. PELUSI, F.N.P., PH.D., Consumer Representative

Phoenix Indian Medical Center

4212 North 16th Street

Phoenix, Arizona 85016


Associate Director

Stem Cell Transplant Program

Center for Cell and Gene Therapy

Baylor College of Medicine

6565 Fannin Street, M964

Houston, Texas 77030


Associate Professor of Internal Medicine

Division of Hematology/Oncology

University of Michigan Comprehensive Cancer Center

7216 Cancer Center

1500 East Medical Center Drive

Ann Arbor, Michigan 48109-0948



Departments of Medicine and Pathology

Indiana University School of Medicine

Indiana Cancer Pavilion

535 Barnhill Drive, Room 473

Indianapolis, Indiana 46202

ATTENDEES (Continued)



Professor of Medicine

Medical Director, Palliative Care Services

Division of Clinical Oncology

University of Kansas Medical Center

3901 Rainbow Boulevard

Kansas City, Kansas 66160-7353




Maryland Regional Cancer Care LLC

40 West Gude Drive

Rockville, Maryland 20850


Professor of Medicine

University of Southern California/Norris Cancer Center

Division of Hematology, MS 34

1441 Eastlake Avenue

Los Angeles, California 90033


Associate Director

Developmental Therapeutics Program

Division of Cancer Treatment and Diagnosis

National Cancer Institute

Executive Plaza North, Suite 8018

6130 Executive Boulevard, MSC-7458

Rockville, Maryland 20852




Austin, Texas




2 Heritage Hills Court

Skillman, New Jersey 08558-2340

ATTENDEES (Continued)



Professor, Radiation Oncology

Radiation Oncology Department

University of Alabama, Birmingham Medical School

619 South 19th Street

Birmingham, Alabama 35233-6832






















BLA 125019, ZEVALIN (ibritumomab tiuxetan)





By Dr. Templeton-Somers 9


By Eleanor Metz 11


By Dr. Marjorie Shapiro 14



By Dr. Leslie Shelly 17

Scientific & Medical Summary of Zevalin

By Dr. Christine White 21

Questions from the Committee 50



By Dr. Ruby Meredith 82


By Dr. Philippe Bishop 101

Questions from the Committee 119



(8:00 a.m.)

DR. NERENSTONE: Good morning. If the committee will take their seats, we'll get started.

This morning we're going to be talking about Zevalin's application, BLA 125019. I'd like to thank everyone for coming.

I'd like to start by introducing the committee, and we'll start with Mr. Ohye.

MR. OHYE: I'm George Ohye, nominee for industry representative.

DR. SLEDGE: George Sledge, medical oncologist, Indiana University.

DR. BRIDGES: Don Bridges, radiation oncologist, Maryland Regional Cancer Care, Rockville, Maryland.

DR. REDMAN: Bruce Redman, University of Michigan Cancer Center.

DR. TAYLOR: Sarah Taylor, medical oncologist, University of Kansas.

DR. PELUSI: Jody Pelusi, oncology nurse practitioner, Phoenix Indian Medical Center, and consumer representative.

MS. KRIVACIC: Susan Krivacic, patient rep, Austin, Texas.

DR. GEORGE: Stephen George, biostatistics, Duke University.

DR. BLAYNEY: Douglas Blayney, medical oncologist, Wilshire Oncology Medical Group, Pasadena, California.

DR. NERENSTONE: Stacy Nerenstone, medical oncology, Hartford, Connecticut.

DR. ALBAIN: Kathy Albain, medical oncology, Loyola University, Chicago.

DR. LIPPMAN: Scott Lippman, M.D. Anderson Cancer Center, Houston.

DR. LEVINE: Alexandra Levine, hematologic oncologist, University of Southern California, L.A.

DR. PRZEPIORKA: Donna Przepiorka, cell and gene therapy, Baylor, Houston.

DR. KELSEN: David Kelsen, medical oncologist, Sloan-Kettering.

DR. CARPENTER: John Carpenter, medical oncologist, University of Alabama in Birmingham.

DR. MILLS: George Mills, acting branch chief for the oncology section in CBER.

DR. BISHOP: Philippe Bishop, CBER, medical oncologist.

DR. KEEGAN: Patricia Keegan, Deputy Director, Clinical Trials, CBER.

DR. TEMPLETON-SOMERS: Thank you all for coming. I realize that space is a little tight. If people from the public can find some seats. I think there are a couple in the front row or in the last two rows over on the side. You're welcome to go ahead.

The following announcement addresses the issue of conflict of interest with respect to this meeting and is made a part of the record to preclude even the appearance of such at this meeting.

Based on the submitted agenda and information provided by the participants, the agency has determined that all reported interests in firms regulated by the Center for Drug Evaluation and Research present no potential for a conflict of interest at this meeting with the following exceptions.

In accordance with 18 U.S.C., section 208(b), full waivers have been granted to David Kelsen, M.D.; Douglas Blayney, M.D.; Susan Krivacic; George Sledge, M.D.; Scott Lippman, M.D.; and Alexandra Levine, M.D.

Further, in accordance with 21 U.S.C. 355(n)(4), Dr. Douglas Blayney, Susan Krivacic, and Dr. Sarah Taylor have been granted waivers that permit them to vote on matters concerning Zevalin.

A copy of these waiver statements may be obtained by submitting a written request to the agency's Freedom of Information Office, room 12A-30 of the Parklawn Building.

In addition, Dr. Kathy Albain and Dr. Sarah Taylor have interests that do not constitute financial interests in the particular matter within the meaning of 18 U.S.C., section 208, but which could create the appearance of a conflict. The agency has determined, notwithstanding these interests, that the interest of the government in their participation outweighs the concern that the integrity of the agency's programs and operations may be questioned. Therefore, Dr. Albain and Dr. Taylor may participate in the committee's discussions and vote concerning Zevalin.

With respect to FDA's invited guest speaker, Dr. Ruby Meredith has reported interests that we believe should be made public to allow the participants to objectively evaluate her comments. Dr. Meredith is a co-investigator of clinical studies and a projected principal investigator at her institution for proposed clinical studies.

We would like to note for the record that George Ohye is participating in this meeting as an industry representative, acting on behalf of regulated industry. As such, he has not been screened for any conflicts of interest.

In the event that the discussions involve any other products or firms not already on the agenda for which FDA participants have a financial interest, the participants are aware of the need to exclude themselves from such involvement and their exclusion will be noted for the record.

With respect to all other participants, we ask in the interest of fairness that they address any current or previous financial involvement with any firm whose product they may wish to comment upon.

Thank you.

DR. NERENSTONE: We're now going to turn to the open public hearing part of the meeting. Eleanor Metz.

MS. METZ: Good morning. I'm sorry for the delay.

Distinguished panel, I am here representing myself and I'm impatiently and anxiously waiting for Zevalin to be approved by the FDA. I've been told by my wonderfully supportive oncologist at John Hopkins, Dr. Richard Ambinder, that this drug would again be one of the best choices of therapy for me as my follicular mixed non-Hodgkin's lymphoma is relapsing again. I implore you to expeditiously approve this drug for people like me.

So, how do I know about this drug? I was fortunate to be part of the phase III trial of, at that time known as, Y2B8, Zevalin, on April 26, 1999, as a patient at Johns Hopkins. And for the records, I was patient 23. Trust me, I had a great deal of trepidation, but under the excellent care provided by Dr. Ian Flinn, who I believe is in the audience today, research nurse Amy Goodrich, and a whole bunch of other people at Hopkins, Zevalin turned out to be a very good choice for me.

I've been a patient at Hopkins for 11 and a half years and have gone through many types of chemotherapy as my cancer has reoccurred. And it's true that I needed some transfusions shortly after Zevalin was administered, but through it all, I was able to work, lead a normal life. It's been two-plus years, and I work. Until very recently there's been little growth internally.

As I said, two-plus quality years. During this time, I've become a grandmother, and since my children were 9 and 17 when I was originally diagnosed, I didn't think I'd get to see that day. This drug Zevalin, along with the care at Hopkins, my own determination, the will of God, et cetera has gotten me this extra time.

My drug therapies are very limited. As a matter of fact, before I went on Zevalin, I was going to have to go on a very hard chemotherapy called ESHAP. Well, let me share a short story. I promise you it will be short.

I was visiting my son one day, my son the lawyer I say with a smile.


MS. METZ: When his friend, the doctor, -- I was telling him about my upcoming chemotherapy. He saw my son a few days later and he said I should ask about Y2B8. Initially I wasn't considered a candidate, but then I heard that there was the phase III trial and Dr. Ambinder felt I should give it a try and work with Dr. Flinn. And it just felt right. It turned out to be right for me. This drug has been wonderful for me.

In my professional life, I work with gifted and talented students in a local school system, and I hope the skills that I'm giving these students will help them go into the field of developing these incredible drugs.

We cancer patients make very difficult choices. We're dealing with this horrible disease, and I know I'm not telling you anything you don't know. But I hope I'm giving you the personal touch with the human face. I, like so many others, including those that don't live near centers that have these wonderful trials, need the opportunity to do these researched options.

I implore you to approve this so that I and others, with the guidance of our oncologists, get a chance. My chance with Zevalin was successful. Two-plus years is very good for me. My daughter is soon to graduate from College Park. I want to see my grandson grow up, and I want to continue to serve and enjoy my family, my education community, and my religious community. Basically I want to live. So, I hope the results from this drug give you the courage and the wisdom to approve this so that I and others can benefit from it.

Thank you.

DR. NERENSTONE: Thank you very much.

DR. TEMPLETON-SOMERS: We received a number of e-mails from the public also in support of Zevalin and other drugs, and those letters are available for the public to view out at the desk in the books. And all of the committee members have been supplied copies of those letters ahead of time. But in the interest of time, I think we're going to leave them to being read privately as opposed to out loud. Thank you.

DR. NERENSTONE: We're going to now turn to the drug application. Dr. Shapiro of the FDA will do the introduction.

DR. SHAPIRO: Good morning. I'm Dr. Marjorie Shapiro, one of the product reviewers and the chair of the Zevalin BLA committee.

This morning the agency and IDEC Pharmaceuticals will present an analysis of the data obtained from the clinical trials involving Zevalin. First I will introduce the BLA Review committee and then briefly describe the product. After the presentations by IDEC and our guest speaker, Dr. Philippe Bishop will finish the FDA presentation.

Besides myself, the other members of the review committee are Philippe Bishop, George Mills, Satish Misra, Dave Green, Mary Andrich, Leon Epps, Deborah Trout, Kevin O'Brien, and Michael Noska.

The 2B8 antibody, or ibritumomab, is a murine IgG1 kappa anti-CD20 monoclonal antibody. It is the murine parent of the chimeric molecule, rituximab. Ibritumomab is manufactured by standard procedures and subsequently conjugated with a chelator/linker, MX-DTPA, or tiuxetan. Ibritumomab tiuxetan conjugate is packaged into kits for radiolabeling with indium, and another kit is used for radiolabeling with yttrium 7 to 9 days later.

This is a schematic diagram of the antibody conjugate prior to radiolabeling.

The kit components include the ibritumomab tiuxetan, the 50 millimolar sodium acetate solution, formulation buffer, and a reaction vial. Even though the kit components for radiolabeling with indium or yttrium are identical, there will be separate kits for radiolabeling with the different isotopes.

If the committee recommends approval of this product, you should be aware that there are some outstanding manufacturing issues that need to be resolved. To prepare radiolabeled Zevalin, the nuclear pharmacy will place one order with the distributor for IDEC. This distributor will coordinate the shipping of the indium kit, the yttrium kit, and the yttrium. The nuclear pharmacy will place a separate order for the indium directly from an approved source. Both radiolabeled forms of Zevalin will be prepared on site at the nuclear pharmacy.

The agency considers all components, including rituximab, indium Zevalin, and the yttrium Zevalin, to be part of the total Zevalin therapy. The step 1 components, rituximab and the indium Zevalin, and the step 2 components, rituximab and yttrium Zevalin, are administered 7 to 9 days apart.

The proposed indication for Zevalin is for the treatment of patients with relapsed or refractory low-grade, follicular, or CD20 positive transformed B-cell non-Hodgkin's lymphoma and rituximab-refractory follicular non-Hodgkin's lymphoma.

Thank you.

DR. NERENSTONE: Obviously the FDA will continue their presentation. Now we'll turn to the sponsor for IDEC Pharmaceuticals' presentation.

DR. SHELLY: Good morning, Dr. Nerenstone, members of the Oncologic Drugs Advisory Committee, and FDA staff. I am Leslie Shelly, Associate Director of Regulatory Affairs at IDEC Pharmaceuticals.

Today I would like to provide an overview of a novel agent, ibritumomab tiuxetan, or Zevalin, which represents a new class of therapies, radioimmunotherapeutics.

Specifically IDEC is requesting approval of Zevalin for the treatment of patients with relapsed or refractory low-grade, follicular, or CD20 positive transformed B-cell non-Hodgkin's lymphoma and treatment of patients with rituximab-refractory follicular non-Hodgkin's lymphoma.

Non-Hodgkin's lymphoma is predominantly of B-cell origin and expresses the CD20 antigen. Collectively non-Hodgkin's lymphoma ranks fifth in cancer incidence and mortality, with an incidence in the United States of approximately 55,000 per year and a prevalence of approximately 300,000 cases. Of these, 65 percent are low-grade or follicular. With progressive disease, there's an increased incidence of transformation from indolent non-Hodgkin's lymphoma to a more aggressive histology.

The median age at diagnosis is 60 years.

The median survival for low-grade or follicular is 6.2 years. For transformed patients, the estimated median survival is 7 to 22 months.

Relapsed or refractory indolent non-Hodgkin's lymphoma is an incurable disease. Although there are other treatment options available for physicians, there's only one approved agent, rituximab. Given that the disease is incurable and that there are limited treatment options for these patients, there's a need for new and innovative therapies.

The Zevalin regimen consists of the following. On day 1, patients receive an initial infusion of rituximab at 250 milligrams per meter squared to optimize biodistribution of subsequently administered radiolabeled antibody. This is followed by indium-111 Zevalin for imaging. Biodistribution of the indium-labeled antibody is assessed by obtaining images at 2 to 24 hours, 48 to 72, with an optional third image at 90 to 120 hours.

Approximately 1 week later, patients receive a second dose of rituximab followed by a single therapeutic dose of yttrium-90 Zevalin. The standard dose of yttrium-90 Zevalin is 0.4 millicuries per kilogram body weight. This is reduced to 0.3 millicuries per kilogram body weight for patients with mild thrombocytopenia. The maximum dose for any patient is 32 millicuries.

Radiolabeling of Zevalin with either indium-111 or yttrium-90 will occur at the nuclear pharmacy. A single point distribution system will be established for distributing the Zevalin kits and the yttrium-90 isotope for use with the kits. This will ensure that control is exercised over the yttrium-90 isotope for use with Zevalin.

Clinical development of Zevalin was initiated in 1993. Seven clinical trials have been conducted, six completed, with one, 106-98, still ongoing. The total body of clinical experience is 489 patients with an integrated safety database of 349 patients and an integrated dosimetry database of 179.

IDEC conducted two key trials that evaluated the safety and efficacy of Zevalin based on the following agreements with FDA. 106-04 is a phase III randomized trial against the standard of rituximab in patients with relapsed or refractory low-grade, follicular, or CD20 positive non-Hodgkin's lymphoma. The primary endpoint was overall response rate. The trial was designed with an 80 percent power to detect a difference in overall response rate between the arms. The secondary endpoint, time to progression, was to be clinically equivalent to the control.

The second key trial, 106-06, was a phase III trial in rituximab-refractory patients.

The product received fast track designation on June 5, 2000. A biologics license application was submitted to FDA on November 1, 2000.

The data we will present today demonstrate that Zevalin has significant clinical activity with acceptable toxicity. In a patient population that has an incurable disease requiring treatment, there's a need for other agents to treat non-Hodgkin's lymphoma. We believe that Zevalin represents a clinically meaningful advance in therapy for these patients.

We'd like to express our gratitude to FDA for providing guidance through the development of this product and working with us in a collaborative manner.

We're pleased to have in attendance individuals who have been instrumental in the development of Zevalin. These include a number of our investigators who are available to answer any questions.

These clinical scientists, as well as our clinicians and biostatisticians, are available today and have played important roles in the development of Zevalin.

Dr. Christine White, Vice President of Medical Affairs at IDEC Pharmaceuticals and a medical oncologist and hematologist, will provide further background on the product and review the safety and efficacy data that we feel support our indication. During the question and answer period, Dr. Pratik Multani, a medical oncologist and hematologist, and Dr. Bryan Leigh, a radiation oncologist, will join Dr. White. Now I'd like to introduce Dr. Christine White.

DR. WHITE: Dr. Nerenstone, advisory committee members, FDA representatives, ladies and gentlemen, it is my pleasure to come before you today to summarize the clinical development of ibritumomab tiuxetan, or Zevalin, radioimmunotherapy for non-Hodgkin's lymphoma.

My presentation today will be organized as follows. First, I will address some background information with regard to non-Hodgkin's lymphoma treatment alternatives, radioimmunotherapy, and Zevalin. I will then briefly discuss the phase I and II conclusions and our conclusions from our imaging and dosimetry trial. Since this data is extensively outlined in your briefing document, I will only briefly touch upon conclusions here. I will then, in more detail, summarize the efficacy results of the phase III randomized trial and the phase III rituximab-refractory trial, adding some additional information with regard to the efficacy in the transformed patients and the non-follicular low-grade patients. I'll then turn to integrated safety and give detailed information on our randomized trial safety, as well as the integrated safety summary of 349 patients treated with ibritumomab tiuxetan. Finally, I will summarize our conclusions.

Initially patients with low-grade or follicular non-Hodgkin's lymphoma are often sensitive to therapy. However, response rates and response durations decrease with relapses and successive courses of therapy. In 1986, Dr. Gallagher published the results of a study at St. Bartholomew's in London looking at response rates and response durations with successive courses of alkylator therapies. You will see on this slide that response rates diminished from 70 percent initially to 39 percent by the fifth therapeutic intervention, and response durations dropped from over 1.5 years to a matter of months by the fourth therapeutic intervention.

This slide also illustrates that these patients require multiple successive therapies.

Treatment of relapsed or refractory low-grade, follicular, or transformed non-Hodgkin's lymphoma typically involves multiple cycles of therapy per course of therapy. In this slide are summarized the results from the literature of single agent trials and multiple chemotherapy agent trials in relapsed or refractory low-grade, follicular, or in some cases transformed disease. And you will see that response rates range from 10 percent to 66 percent with the more aggressive therapies.

Toxicities of commonly used regimens can be formidable. Up to 59 percent of patients can develop grade 4 neutropenia, and this, of course, is recurrent with recurrent cycles of therapy. Up to 39 percent grade 4 thrombocytopenia, up to 40 percent of patients can develop grade 3 and 4 infections, and treatment related deaths, unfortunately, can be substantial with 16 percent, 17 percent, and even 23 percent of the patients dying of acute toxicity in these studies. In addition, some agents have grade 3 or 4 neurotoxicity in up to 10 percent of patients or renal or hepatic toxicity, both reversible and nonreversible, in up to 21 percent of patients.

There's no cure for relapsed or refractory low-grade, follicular, or transformed non-Hodgkin's lymphoma, and no regimen has been shown to be superior with regard to survival. Patients need additional treatment options. In the absence of cure or survival benefit, treatments that induce remission and prolong time off therapy are valuable.

The rationale for developing radioimmunotherapy in non-Hodgkin's lymphoma stems from the inherent sensitivity of non-Hodgkin's lymphoma to radiation. External beam radiation can be curative in limited stage non-Hodgkin's lymphoma, but cannot be applied to advanced stage disease because of toxicity. Antibodies that are radiolabeled can target radiation to the tumor cells. Radioimmunotherapy can kill both bound tumor cells and neighboring tumor cells in the same mass, overcoming the problem of access in poorly vascularized tumors or bulky tumors.

Critical factors for successful therapy with radioimmunotherapy include target antigen, antibody selectivity, choice of isotope, and stability of antibody-isotope linkage.

Zevalin, or ibritumomab tiuxetan, is the murine monoclonal antibody, ibritumomab, which is the murine parent of the rituximab engineered chimeric antibody. It binds the same epitope on CD20 and is covalently bound to tiuxetan, which stably retains 90 yttrium upon chelation. The CD20 antigen is expressed only on B lineage cells. It is important for cell cycle differentiation and initiation. And very importantly for immunotherapy, it does not shed into the blood, does not down-regulate upon binding, and does not modulate.

We have chosen 90 yttrium as the isotope to use in radioimmunotherapy for the following reasons. It has a half-life of 64 hours. The high beta energy of 2.3 MeV and path length of 5 millimeters make it advantageous in treating bulky tumors and poorly vascularized tumors. Because it is a pure beta emitter, patients can be treated as an outpatient in all cases without having to be restricted, shielded, or have any particular restrictions on their behavior or environmental impact.

I will now turn to the phase I/II study conclusions.

From our phase I study and our phase I/II study, we were able to conclude that unlabeled, pretreatment rituximab improved Zevalin biodistribution. The maximum tolerated dose was determined to be 0.4 millicuries per kilogram, 0.3 millicuries per kilogram in patients with mild thrombocytopenia. In all cases the maximum dose was 32 millicuries.

Toxicity was primarily hematologic and reversible and correlated with millicurie per kilogram dose.

Clinical variables, including baseline platelet count, a surrogate for bone marrow damage from prior chemotherapy and prior external beam radioimmunotherapy, as well as percent bone marrow involvement, a surrogate for specific targeting of Zevalin to the disease in the marrow, were more predictive of toxicity than was dosimetry.

Clinical parameters proved adequate for dosing.

The overall response rate was 82 percent in low-grade non-Hodgkin's lymphoma in our phase I/II study, and 42 percent in intermediate-grade predominantly diffuse large cell non-Hodgkin's lymphoma.

The median time to progression in the patients who received the 0.4 millicurie per kilogram dose was 15.4 months. These are the responders. And the median time to progression in those patients who received 0.4 millicurie per kilogram and achieved a complete response has not been reached yet with a range of 28.3 to 37.2-plus months.

Turning to imaging and dosimetry, this is a representative scan from a patient who was imaged with indium-labeled Zevalin. At 4 hours, there is Zevalin still in the blood pool, as evidenced by appearance of the iliac vessels and the cardiac shadow.

By 66 hours, this large periaortic and retroperitoneal mass, seen very well here on CT-scan, is specifically targeted by the Zevalin radioimmunotherapy and appears very bright. You can also see the liver shadow.

At 139 hours, there is retention of the Zevalin radioimmunotherapy in the large mass.

On SPECT scan, you can see that the targeting to the mass is very intense, much more so than the targeting to the liver, which is the major organ of metabolism.

Tumors imaged in all low-grade follicular and transformed patients. Radiation absorbed doses to normal organs was acceptable in all patients. There was minimal urinary excretion, 7 percent over 7 days time, which is 3.5 physical half-lives of the isotope. There was no correlation between hematologic toxicity and dosimetry parameters or pharmacokinetic parameters.

This bar graph shows the median 90 yttrium radiation absorbed doses to tumor and to normal organs. You will see that the median dose here to tumor is approximately 1,500 centigray with a range up to 24,000 centigray. The spleen, which is often involved with lymphoma, receives 1,000 centigray. The testes and the lungs were 950 centigray median and 220 centigray median, and we feel that this may be overestimated as our methods for dosimetry did not include the proper attenuation correction factors for these superficial organs. So, that is an absolute maximum. Median dose to liver was 520 centigray; to bone marrow, 143 centigray; and to kidneys, 13 centigray; total body, 58 centigray. This is the acceptable maximum dose to bone marrow, 300 centigray, defined in the protocol, and 2,000 to normal organs. And you can see that the normal organs received far less than the acceptable limits.

I will now turn to the efficacy results from our phase III randomized and our phase III rituximab-refractory trials.

There were approximately 200 patients in these two trials, the phase III randomized and the rituximab-refractory trials, and these patients were a representative population with regard to the greater population of patients with low-grade, follicular, or transformed non-Hodgkin's lymphoma. In this table, you can see that patients on the Zevalin arm of the phase III trial, the rituximab arm of the phase III trial, and on the rituximab-refractory trial were similar with regard to age, gender ratio, histologic classification, stage, and marrow involvement to the referenced group of 433 patients first defined in 1982 at the time of the definition of the International Workshop Histologic Classification, and also to 1,850 patients described in seven comparison studies in the literature.

Our phase III randomized study was conducted at these 27 institutions.

The study design was as follows. It was stratified by histology, IWF A, which was actually non-follicular low-grade, follicular, which was the second category, or transformed. And the patients received either the Zevalin regimen as previously described, that is, two doses of rituximab 250 milligrams per meter squared, the first followed by 5 millicuries of indium-labeled Zevalin, the second followed by 0.4 millicurie per kilogram of 90 yttrium Zevalin with a maximum of 32 millicuries. Or if on the control arm, the patients received rituximab, 375 milligrams per meter squared weekly times 4.

The primary endpoint prospectively defined was overall response rate. The study was designed with an 80 percent power to detect a difference of 25 percent in overall response rate with an alpha of 0.05. Overall response rate was determined by an independent lymphoma expert confirmation of response panel comprised of radiologists and oncologists expert in lymphoma who were blinded to what treatment the patient received and to the investigator-assessed response.

Secondary endpoints included duration of response, time to progression, time to next therapy, complete response, clinical complete response, partial response, and quality of life.

Patients were enrolled in the trial if they had bidimensionally measurable disease initially greater than 3 centimeters. Later the trial was amended to be greater than 2 centimeters. Less than 25 percent marrow involvement by biopsy, looking at the cellular space. A WHO performance status of 2 or less, and adequate hematologic function, including an absolute neutrophil count of 1,500 or greater, and a platelet count of 150,000 or greater.

Patients were excluded if they had prior myeloablative therapy or radioimmunotherapy, prior radiation to more than 25 percent of the active marrow, prior anti-CD20 therapy, total bilirubin or creatinine greater than 2, or a lymphocyte count circulating greater than 5,000.

The two arms of the study were well balanced. There was no statistically significant difference in these prognostic factors, and in particular, I will point to number of prior regimens, response to last therapy, marrow involvement, splenomegaly, bulky disease, and chemoresistance.

Patients were 60 years of age median in the Zevalin arm, 57 on the rituximab arm. But the age of the patients extended up to 80.

Most of these patients had follicular histology: 75 percent on the Zevalin arm; 83 percent on the rituximab arm. Smaller numbers of patients with non-follicular, low-grade, or transformed histologies were entered.

The median number of prior therapies was 2, with a range to 6.

43 percent of patients on the Zevalin arm and 36 percent on the rituximab arm had bone marrow involvement.

10 percent on Zevalin arm and 4 percent on the rituximab arm had splenomegaly.

18 percent on the Zevalin arm and 13 percent on the rituximab arm had extranodal disease, and about 45 percent of these patients had bulky disease greater than 5 centimeter lesions. 8 percent had lesions at least 10 centimeters or more.

Nearly half of these patients were resistant to their most recent chemotherapy and up to 64 percent were resistant to any prior chemotherapy. Resistance is defined for this protocol as no response or time to progression of less than 6 months.

This bar graph shows the response assessment by protocol-defined response criteria on the phase III randomized trial. There was a 73 percent response rate on the Zevalin arm, 47 percent on the rituximab arm, and this difference was statistically significant with a p of 0.002. There were 18 percent complete responders and 3 percent clinical complete responders on the Zevalin arm, 11 percent and 4 percent on the rituximab arm.

This table shows response by histology. Again, the majority of patients were follicular histology, and there was a 76 percent response rate on the Zevalin arm, as compared to a 47 percent response rate on the rituximab arm, with a p of 0.002 with regard to this statistically significant difference. There was no statistically significant difference in the non-follicular, low-grade, or transformed groups. The IWF A or non-follicular, low-grade group included small lymphocytic lymphoma patients, MALTomas, lymphoplasmacytic, and monocytoid B.

This table shows response rate in patients resistant to their most recent chemotherapy. The response was 64 percent on the Zevalin arm, 36 percent on the rituximab arm, with a statistically significant difference and a p of 0.045. Again, resistance to last chemotherapy was defined as no response or a time to progression of less than 6 months.

Looking at patients with regard to tumor bulk less than 5 centimeters or greater than 5 centimeters, in this larger tumor bulk category there was a 67 percent response rate on the Zevalin arm and a 45 percent response rate on the rituximab arm.

At the time that these protocols were developed, there was no standardized criteria for response for non-Hodgkin's lymphoma. For that reason, IDEC worked with the FDA to develop protocol-defined response criteria. Subsequently, however, the National Cancer Institute has convened an international workshop to establish NHL response criteria that would be standardized internationally. These have been published and are now being used, rapidly adopted to evaluate trials and to compare.

The LEXCOR measured response rate, using the international workshop response criteria, was 80 percent for the Zevalin arm and 56 percent for the rituximab arm. This difference was statistically significant with a p of 0.002. The complete response rate was 30 percent for the Zevalin arm and 16 percent for the rituximab arm, and this difference was statistically significant as well, with a p of 0.040. There were 4 percent unconfirmed complete responders on both arms.

This table summarizes duration of response, time to progression in all patients, and time to progression in responders. There was no statistically significant differences in each of these variables between arms. The duration of response on the Zevalin arm was estimated by Kaplan-Meier methods at 14.2 months; on the rituximab arm, 12.1 months. Time to progression in the responders was 15.4 months by Kaplan-Meier estimation on the Zevalin arm and 13.8 months by Kaplan-Meier estimation on the rituximab arm. At the time of this analysis, 32 percent of the patients had ongoing responses on the Zevalin arm and 19 percent on the rituximab arm, with a median observation time of 22 months.

This is a slide that demonstrates duration of response by histology. Again, looking at the follicular histology where the majority of patients were, the duration of response estimated by Kaplan-Meier methods is 18.5 months on the Zevalin arm as compared to 12.1 months on the rituximab arm. This difference is not statistically significant.

This is a Kaplan-Meier curve of time to progression in all patients, again not statistically significant.

This is the same group of all patients, and this is a Kaplan-Meier curve of time to next therapy, and there is a trend towards greater time to next therapy in the Zevalin arm as compared to the rituximab arm, with a p of 0.084.

In follicular patients, this is a Kaplan-Meier curve of time to progression. Again, there is a trend towards a greater time to progression in the Zevalin arm as compared to the rituximab arm, with a p of 0.062.

For follicular patients, this is time to next therapy by Kaplan-Meier analysis. Again, a trend towards longer time to next therapy in the Zevalin arm as compared to the rituximab arm, with a p of 0.058.

Looking only at patients who achieved a complete response or a clinical complete response, there is a trend towards longer time to progression in the Zevalin arm as compared to the rituximab arm, though this is not statistically significant in this small group. No median has been reached yet for the Zevalin patients.

To summarize the phase III randomized trial efficacy results, efficacy objectives were met. The primary objective of overall response rate was significantly higher, as determined by an independent blinded lymphoma expert confirmation of response panel, as compared to the control arm.

Secondary objectives were also met, including a comparable overall time to progression. In addition, although this trial was not powered to show differences in TTP, a trend towards longer TTP in the follicular patients and in the complete response and clinical complete response patients was demonstrated. There was also a trend towards longer time to next therapy in all patients, and the median time to progression responders has been 15.4 months by Kaplan-Meier estimation.

I will now turn to the phase III rituximab-refractory trial results. This trial was conducted at these 17 institutions.

Study design was as follows. 28 patients received the same regimen I described before: rituximab 250 milligrams per meter squared on two occasions, 1 week apart; the first being followed by indium-labeled Zevalin, 5 millicuries, the second being followed by 90 yttrium-labeled Zevalin, 0.4 millicurie per kilogram with a maximum of 32 millicuries. By protocol design, 29 patients did not receive indium-labeled Zevalin imaging or dosimetry.

The primary efficacy endpoint was a target overall response rate in 35 percent of rituximab-refractory follicular patients. Rituximab-refractory was defined as no response or a time to progression of less than 6 months to rituximab. The response was again evaluated by the independent lymphoma expert confirmation of response panel, again blinded to the treatment received and also to the assessed response by the investigator.

Secondary endpoints were duration of response, time to progression, time to next therapy, CR, CCR, PR, and a comparison of overall response rate and duration of response to the prior rituximab and to the last chemotherapy, as is demonstrated on this line graph.

Median age was 54 years with a range to 73 years.

This was a more heavily pretreated, more advanced disease population. There was a median of four prior chemotherapy regimens with a range to nine.

32 percent of patients had bone marrow involvement.

18 percent had two or more extranodal disease sites.

12 percent had splenomegaly.

Three-quarters of these patients had bulky disease, defined as greater than 5 centimeters. 44 percent had tumor lesions greater than 7 centimeters, and 19 percent had tumor lesions of 10 centimeters or more.

Two-thirds of these patients were resistant to last prior chemotherapy, again defined as no response or time to progression of less than 6 months.

The response assessment by the LEXCOR, lymphoma expert confirmation of response, panel using the protocol-defined response criteria, included an overall response rate of 59 percent, with a complete response of 4 percent. By international workshop NHL response criteria, the overall response was 74 percent, with a complete response rate of 15 percent.

Duration of response was estimated by Kaplan-Meier methods at 7.7 months; time to progression in all patients, 6.8 months; and time to progression in responders, 9.1 months.

Analysis was performed to compare Zevalin to the prior therapy, both rituximab and prior chemotherapy. The methodology was as follows.

The analysis favored one therapy if the patient responded to one therapy but not to the other, or if the patient responded to both but the duration of response was at least 3 months longer. This analysis was actually done for 3 months and 1 month. Both are in your briefing document, but I will present the 3-month more conservative analysis.

It was neutral if the patient did not respond to either or responded to both, but the duration of response was 3 months greater.

Here are the results of that analysis. The analysis favored Zevalin in 48 percent of patients, favored prior therapy in 9 percent of patients, that is, in the comparison to rituximab, and was neutral in 43 percent of patients. This difference was statistically significant with a p by sign rank of less than 0.001 and by sign of 0.011.

When comparing Zevalin to the last chemotherapy, the analysis favored Zevalin in 30 percent of patients, favored prior therapy in 30 percent of patients, and was neutral in 41 percent of patients. This difference was not statistically significant, which was a better than expected result in that the chemotherapy was the patient's third regimen and the Zevalin was the patient's fifth regimen on median.

To summarize the rituximab-refractory trial results, the efficacy objectives were again met. There was a significant overall response rate in the heavily pretreated, bulky disease, rituximab-refractory patient population.

Secondary endpoints were also met. There was a statistically improved efficacy demonstrated with Zevalin therapy as compared to the prior rituximab therapy, and there was no difference between Zevalin therapy and the prior chemotherapy, a better than expected result in that the chemotherapy was the third regimen and the Zevalin the fifth on median. The median time to progression in responders in this patient population of advanced disease, heavily pretreated, and bulky disease was 9.1 months.

I would now like to give some more details with regard to efficacy in the transformed and non-follicular low-grade groups.

Across our phase I/II, phase II, and phase III Zevalin trials, there were 15 patients with transformed histologies. Of those 15 patients, 6 responded for an overall response rate of 40 percent; 2 achieved complete responses for an overall complete response rate of 13 percent. Time to progression ranges from 0.8 to 37.2-plus months, and 3 patients, or 20 percent of this small group, are still in ongoing remission at 11-plus, 22-plus, and 37-plus months.

With regard to the non-follicular low-grade histology patients -- again, these are including small lymphocytic lymphomas, lymphoplasmacytic, MALTomas, and monocytoid B -- there were 16 patients in our phase I/II, phase II, and phase III Zevalin trials. 9 responded for an overall response rate in this small group of 56 percent with a complete response in 1 patient, or 6 percent. Time to progression has been 6.1 months to 12.6-plus months, with 2 patients, or 14 percent, in ongoing remission in excess of a year.

I will now summarize our integrated safety.

Adverse events are primarily hematologic with Zevalin. Nonhematologic adverse events are primarily grade 1 and 2. Zevalin is not associated with hair loss, severe mucositis, persistent nausea and vomiting, and other symptoms common with chemotherapy. There's a low incidence of serious infection and a low incidence of human anti-mouse antibody or human anti-chimeric antibody.

This bar graph demonstrates the most frequent nonhematologic adverse events in a population of 349 patients treated with Zevalin. Grade 1 is in yellow, grade 2 in light blue, grade 3 in orange, and grade 4 in green. You can see on this graph that nonhematologic adverse events were primarily grade 1 and 2. They included fatigue, nausea, chills, fever, headache, and throat irritation previously described with rituximab infusion.

There was no difference in adverse events between patients who were less than 65 years of age and greater than 65 years of age.

This bar graph looks at only the phase III randomized trial patients and, in particular, displays the most frequent nonhematologic adverse events in the Zevalin arm in yellow and in the rituximab control arm in orange. This center vertical line divides grade 1 and 2, which is on the left, and grade 3 and 4, which is on the right. Initially you can see that there are very few grade 3 and 4 nonhematologic adverse events.

With regard to grade 1 and 2 events, there were more events of nausea, vomiting, abdominal pain, cough, and dizziness and dyspnea, particularly in anemia patients, on the Zevalin arm. There were more events of pruritus and angioedema on the rituximab arm.

This graph depicts whole blood counts over time. In yellow is hemoglobin. This is the median hemoglobin for 349 patients. The yellow dotted horizontal line is a hemoglobin of 10. This red line is the platelet count for 349 patients, the median, and this red horizontal line is a platelet count of 50,000. The green line is the absolute neutrophil count median, and the green dotted horizontal line is an absolute neutrophil count of 1,000.

This is a scatter plot showing all blood counts on all 349 patients. There's a gathering of the data in what appears to be vertical lines because most blood counts were taken on weekdays and not weekends.

You can see that with regard to absolute neutrophil count, there's a trend downward with a nadir at week 9 and subsequent recovery. 5 percent of patients recovered to an absolute neutrophil count over 1,000 following the week 13 period. After week 13, patients who had recovered were not obligated to have blood counts drawn, so many of the recovered patients disappear on this graph following this line. There were 4 patients who did not recover. They included 2 patients who died, one of lymphoma and one of an event that I will discuss in a few minutes, and 2 patients who remain neutropenic and asymptomatic.

This is a scattergram of all platelet counts on 349 patients over time. Again, you see there's a trend downward with a nadir at week 7 and then recovery. Again, those patients who have recovered don't have their platelet counts taken, for the most part, after the 91st day here, or 13th week. 3 percent recovered their platelet counts to 50,000 after week 13, and again the 4 patients, 2 who died and 2 who have not recovered their counts. Patients who went on to subsequent therapy had blood counts no longer displayed in this graph after subsequent therapy.

This is the method that we used at the request of the FDA to determine duration of hematologic toxicity. This arrow here that I'm pointing to is the first date of grade 3 or 4 hematologic toxicity, and this arrow here is the last date of grade 3 or 4 hematologic toxicity. We determined duration by going to the blood count prior to the first date in grade 3 or 4 and looking at the blood count following the last blood count in grade 3 or 4, and this was the duration of hematologic toxicity. This is 1 week in this interval and another week in this interval for most patients.

The median nadir absolute neutrophil count for the entire population was 800 in ANC. 28 percent of patients had an absolute neutrophil count grade 3 nadir, and 30 percent had a grade 4 nadir. Using the method that I just described, the median days within grade 3 or 4 only for those patients who had a grade 3 or 4 nadir was 22.

The median platelet nadir was 41,000. 52 percent of patients had grade 3 platelet nadirs. 10 percent had grade 4 platelet nadirs, and the median days within grade 3 or 4 for patients who had grade 3 or 4 platelet nadirs, using the method I just described, was 24.

Median hemoglobin was 10.5, 14 percent grade 3 and 3 percent grade 4 nadirs, with median number of days, using the same method, of 14.

There was no difference in hematologic toxicity for patients less than 65 years of age or patients greater than 65 years of age.

We also conducted a phase II study that looked at dose reduction in patients who were mildly thrombocytopenic at baseline, in other words, those patients with poorer bone marrow reserve, more damage from prior chemotherapy, or radiation. In these patients, similar to slightly worse hematologic toxicity occurred even though the patients received a 25 percent less dose.

Grade 3 and 4 hematologic toxicity correlated with the percent bone marrow involvement. This was not surprising because the more bone marrow involvement, the more targeting of the Zevalin to the marrow. It also correlated with the number of prior therapies and whether the patient had purine analogs in the past.

Despite myelosuppression, the clinical consequences of myelosuppression were not great. This trial allowed growth factors to be used and transfusions to be given at the physician's discretion. Only 18 percent of patients were given growth factors. 13 percent were given G-CSF and 8 percent were administered erythropoietin. 20 percent received a red blood cell transfusion and 22 percent received a platelet transfusion. However, there were only 2 percent of patients who experienced grade 3 to 5 bleeding events.

The incidence of infection is displayed in this table. Looking at any grade infection, in the integrated safety experience of 349 patients, there were 29 percent who had any grade infection. Only 5 percent of these were grade 3 or 4. Only 7 percent of patients were hospitalized with infection.

Looking at the phase III randomized trial, 43 percent of the patients in the Zevalin arm and 20 percent in the rituximab arm had any grade of infection. The excess infections in the Zevalin arm were predominantly urinary tract infections and upper respiratory infections. Only 7 percent of patients on the Zevalin arm had a grade 3 or 4 infection, and 7 percent were hospitalized with infection. None of the patients on the rituximab arm had a grade 3 or 4 infection, and only 1 percent was hospitalized with infection. This was during the 13-week treatment period defined as 12 weeks following the Zevalin administration.

In the follow-up period, which was defined as from 13 weeks up to 4 years or when the patient went off study for progression of disease, there were 10 percent any grade infections in both arms. On the Zevalin arm, 4 percent of patients had a grade 3/4 infection which required hospitalization; on the rituximab arm, 1 percent which required hospitalization.

Median B-cells recovered to normal by month 6. Median T-cells remained normal. Median IgG and IgA remained normal. Median IgM, which began at very low normal at baseline, had a transient decline but recovered by month 6. The development of human anti-mouse antibody occurred in 1.4 percent of patients after treatment, and the development of human anti-chimeric antibody in 0.5 percent of patients.

5 patients have developed myelodysplasia or acute myelogenous leukemia out of 349, for a raw total of 1.4 percent. Myelodysplasia or acute myelogenous leukemia developed from 4 years to 14 years post-diagnosis in these 5 patients and from 8 months to 34 months post-Zevalin in these 5 patients, although this 1 patient who was diagnosed at 8 months in retrospect had blasts in the peripheral blood at week 4 after Zevalin.

All of these patients had extensive alkylator therapy in the past from 11 months to 18 months to 28 to 21 months of alkylator therapy, and this patient who could not have alkylator therapy quantitated had had prior chlorambucil, CHOP, ProMace-CytaBOM, and FAMP.

Chromosomal abnormalities were found in the 4 patients where they were looked for. In 3 of these patients, they were chromosomal abnormalities that have been described with chemotherapy-associated myelodysplasia and AML.

This table displays the annualized rate for time to development of myelodysplasia or AML. The Kaplan-Meier estimate from date of diagnosis was 0.6 percent and from date of first infusion of Zevalin, actually rituximab in the Zevalin regimen, was 1.1 percent. The number of events per person-years was 0.3 percent from date of diagnosis and 1.2 percent from date of first infusion. The literature quotes in 1,100 patients a 4 to 8 percent incidence in patients who have not had high-dose therapy with stem cell or bone marrow transplant rescue and a 1 to 1.5 percent per year for a 2- to 9-year incidence.

There have been 70 deaths among the 349 patients to date. 56 deaths have occurred secondary to progression of non-Hodgkin's lymphoma. 2 patients died in neutropenic sepsis following additional chemotherapy. The 5 patients who developed myelodysplasia or AML have died.

There have been two treatment related deaths. These were both traumatic intracranial hemorrhages in patients at platelet nadir. In one case the patient was also on therapeutic Coumadin for a history of chronic DVT and was taking self-prescribed ibuprofen.

There were 5 deaths that were due to unrelated or pre-existing illnesses. These included the following: pneumonia at 16 months post Zevalin, occurring when the patient was receiving CHOP chemotherapy plus rituximab; a patient who died of COPD after CVP therapy 29 months following Zevalin; a patient who died of respiratory failure, 2.7 months following Zevalin. This patient had preexisting idiopathic pulmonary fibrosis and had been on escalating doses of azathioprine. A patient who had a history of coronary artery disease, a coronary artery bypass graft times 3, hypertension, CHF, and COPD died of a cardiac arrest at 2.4 months following Zevalin. And a patient who was oxygen dependent when they came onto study and on 10 liters the day that they received the Zevalin therapy died at week 1.

The question has been asked as to whether patients can receive subsequent chemotherapy in the event of relapse following Zevalin. 139 of our patients have received subsequent chemotherapy following relapse. Response assessment was available for 40 patients. The others were still undergoing chemotherapy. Of those 40 patients, 50 percent responded to various types of subsequent therapy, and please note that some patients have been treated with single agent purine analogs and also with other aggressive therapies like ESHAP, DHAP, ICE, ProMACE-CytaBOM, CHOP-Bleo, and fludarabine-containing combinations, as well as transplantation and CHOP. When we look on the two arms for the Zevalin-treated patients and the rituximab-treated patients, there is only a small number of patients who have gone on to chemotherapy where response assessment is available, but there are similar rates of response in those two groups of patients.

10 patients have received transplants following Zevalin. Nine were autologous and one was allogeneic. 6 of the patients who had autologous transplants had their stem cells collected after Zevalin. 2 additional patients have had stem cells collected but have not yet undergone their transplants.

In conclusion, Zevalin represents a new class of targeted therapy. It is well-tolerated outpatient therapy and it is completed in 8 days. Adverse events are primarily hematologic, and their severity is related to the baseline platelet count, which is a surrogate for prior damage from prior chemotherapy and external beam radiation, and to the percent bone marrow involvement, which is a surrogate for specific targeting.

Nonhematologic adverse events have been predominantly grade 1 and 2. There has been a low incidence of serious infection, and treatment-related mortality has been less than 1 percent.

The incidence of human anti-mouse and human anti-chimeric antibody is less than 2 percent.

There have been rare cases of myelodysplasia but within the expected rate for this patient population.

Clinical benefit has been established. The overall response rate is statistically higher than the rituximab control in the randomized phase III study. There have been trends towards longer time to progression in the follicular patients and in the patients who achieved complete response, as well as longer time to next therapy in all patients.

Significant activity has been demonstrated in a heavily pretreated, bulky disease, rituximab-refractory population.

The median time to progression in the responders, both in the phase I/II trial and in the randomized phase III trial, has been 15.4 months.

I hope that the data that I've shown you today has convinced you that Zevalin therapy represents a clinically meaningful advance for patients with relapsed or refractory, low-grade, follicular, or CD20 positive transformed non-Hodgkin's lymphoma, as well as patients with rituximab-refractory non-Hodgkin's lymphoma. I thank you for your attention.

DR. NERENSTONE: Thank you very much.

We're going to open it now for questions from the committee to the sponsor. Dr. Sausville.

DR. SAUSVILLE: Could you clarify how you would see the indium portion of the regimen as playing into the use in the field, so to speak, and how one outcome or another with the indium would contribute to subsequent use of the yttrium-labeled product?

DR. WHITE: Yes. In our phase I/II and even in our phase III randomized trial, at the recommendation of the FDA, we included dosimetry. We believe that dosimetry is absolutely mandatory in the phase I/II and phase II development of a radioimmunotherapy. We included the dosimetry in order to summarize the radiation absorbed doses to normal organs.

Indium-labeled Zevalin was used as a gamma emitter in order to be able to produce the images and to obtain that dosimetry. Since we discovered and were able to confirm the clinical parameters were sufficient for dosing and the dosimetry estimated radiation absorbed doses to normal organs were acceptable in all patients, it is no longer necessary to do complex dosimetry.

However, we now do indium-labeled Zevalin imaging, and an image is obtained at two or three time points, as was described in the presentation. That image is examined to look for any evidence of contraindication for proceeding to therapy. The types of areas that one might look for there would be total urinary obstruction or altered biodistribution, something like that.

DR. SAUSVILLE: Right. So, I guess then you would have criteria written in to how it would be used potentially that would have fairly precise definitions of those sorts of criteria?

DR. WHITE: Yes, we do, and we have included them in our protocols and would also include them, of course, in a package insert. If you're interested, we can show those criteria.

DR. SAUSVILLE: In your experience, have you seen, shall we say, that all tumors tend to label or image as well as others? Is there a range of, shall we say, imaging that might suggest that this could be a variable in response?

DR. WHITE: Let me answer that first, and then we'll get into the imaging criteria.

First of all, let me say that we did tumor dosimetry in selected patients with selected tumors. It's more difficult to do tumor dosimetry logistically because the entire perimeter of the tumor is not always visible. Tumors in all patients with low-grade, follicular, or transformed lymphoma image. However, we know, from doing tumor dosimetry, that larger tumors have a tendency to get a trend towards a lower dose than smaller tumors do. The homogeneity appears approximately the same.

Joining me here at this microphone is Dr. Bryan Leigh, who's a radiation oncologist and worked with our dosimetry, and he's going to present the imaging requirement and might be able to comment a little bit further with regard to the intensity of targeting with regard to the indium-labeled Zevalin.

Before you start, I just wanted also to introduce Dr. Pratik Multani, who's a medical oncologist and was our safety officer for the past two years, who will join me with regard to safety questions.


DR. LEIGH: Thank you.

Let me just review what we see as the purpose of imaging. It's a visual evaluation of the Zevalin biodistribution, and that should be differentiate from complex dosimetry where you actually calculate the radiation absorbed dose to an organ. So, it's similar to an Oncoscint scan or a bone scan where you're looking at the image and making a decision based on the visual interpretation.

We've been asked by the FDA to include this as a safety measure that may detect unexpected areas of high radioactivity uptake that may indicate that the patient should not proceed to the yttrium-labeled Zevalin.

As you've seen, the indium is given approximately 1 week before the yttrium-labeled Zevalin and following an infusion of rituximab. The first image is obtained between 2 and 24 hours after the indium, the second one between 48 and 72 hours after the indium, and then an optional third image to resolve any ambiguities.

In a second I'll show you what is expected to be seen on the images, and we'll make this available through our medical information department. The radioactivity is typically seen in the blood pool on the first image, less so on the second image, moderately high to high uptake in the normal liver and the spleen, and low uptake in the lungs, kidneys, urinary bladder, and bowel.

There's an example of this on the next slide. Here you can see what a typical patient's indium scan would look like. There are anterior and posterior scans at three time points. The first time point you can see the activity in the blood pool, as evidenced by the blood vessels and the cardiac shadow. Later you see targeting of the activity to an abdominal mass and to inguinal lymph nodes as well as posterior cervical lymph nodes and even more targeting on the later image.

Working with the FDA, we've identified some potential items that could be discovered on the indium scan. Potential things that might indicate that the patient should not proceed to treatment might be if the blood pool is not visualized on the first image, with rapid clearance to the liver, spleen, and marrow by the reticuloendothelial system, that may indicate that the patient has an unknown HAMA causing rapid clearance. A second item is if there was increased uptake in the kidneys or the bladder, it may identify urinary tract obstruction, and also high uptake in the lungs or kidney, more intense in the liver on the second image might indicate that there's disease in the lungs or kidneys that we were not aware of.

I'd just like to end by pointing out that we have observed any of these altered biodistributions to date, but they still remain a possibility and we would like to include the imaging as a safety measure.

DR. WHITE: These are the criteria that are in our trials and would be included in the labeling.

DR. SAUSVILLE: So, there would be quantitative limits of how much could be taken up by these other sites then.

DR. LEIGH: No, there would not be quantitative limits.

DR. WHITE: This is a visual --

DR. SAUSVILLE: This is what I'm a little unclear about. In other words, you're saying a little bit of this, a little bit of that. How would we use this in practice I guess?

DR. WHITE: In actuality in our experience now with 489 patients and 200 patients where we have dosimetry, again we found acceptable radiation absorbed doses to normal organs in all cases. But we have only treated those 489 patients that have been submitted in the BLA. Of course, with a broader experience sometimes you can pick up something there at a very low incidence or very rare frequency. The imaging is only being included as an extra safety step for the nuclear medicine physician to take a look and make sure there's no dramatic altered biodistribution or dramatic complete urinary obstruction.

We would be prepared to further collect data on imaging as time went on to see how useful it is as we get into a broader, larger patient population.

DR. SAUSVILLE: Again, my comment is I think it was a really good thing to do in the initial studies and the dosimetry was useful. It's just that I think one would have to carefully consider what level of complexity it adds to its actual use potentially in a widespread way in contrast to what we learn.

Turning to another question that occurred to me, the patient population that you studied in your phase III's had not had prior bone marrow transplantation, as I understand it, with any type of preparative regimen.

DR. WHITE: That is correct.

DR. SAUSVILLE: So, do you foresee that population as needing to be either studied in greater detail or perhaps have some level of stratification in subsequent studies? How do you propose to deal with that substantial fraction of the population?

DR. WHITE: Absolutely. Our proposal would be that the labeling of this product at commercialization would prohibit the use in patients who have had prior bone marrow transplant. However, we are interested in further studying patients with prior bone marrow transplant or stem cell rescue with high-dose therapy as to whether those patients can receive potentially a lower dose or what the proper approach to those patients is. Actually Dr. Julie Vose in Nebraska is just beginning an investigator-initiated trial which will examine exactly this question. I think she just has entered her first patient or two.

DR. NERENSTONE: Dr. Blayney.

DR. BLAYNEY: Thank you. I have three questions.

Why do you think the incidence of HAMA, human anti-mouse antibody, is so low in your patients?

DR. WHITE: Our speculation is as follows. When we had our very first phase I study in 1993, that study differed -- and by the way, the safety from that study was not integrated in the 349 patients because the study used ibritumomab rather than rituximab as the pretreatment cold antibody. It also dose escalated the myeloablative ranges and had some other differences as well.

In that study, we had about a 12 percent HAMA rate. When we started using rituximab as the pretreatment cold antibody, our expectation was that we would deplete B-cells effectively and that there might be a lower HAMA rate because of that B-cell depletion.

In actuality since that time with all of those patients that we have described, there has been a less than 2 percent HAMA rate. So, we believe that it may very well be due to the prior rituximab. Of course, lymphoma patients develop HAMA at a lower rate than do solid tumor patients, but other trials using murine antibodies in lymphoma patients with other agents have described up to a 66 percent HAMA rate. So, I think it is the rituximab given prior to the Zevalin radiolabeled antibody.

DR. BLAYNEY: What guidance do you propose to give to physicians who may want to treat patients who have had previous external beam radiation therapy to bulk disease or who may ask about retreating patients who have previously received your labeled monoclonal?

DR. WHITE: Let me take those separately.

With regard to patients who have had previous external beam radiation therapy, our only guideline which would be in the package insert would be not to include patients who have had radiation therapy to more than 25 percent of their active marrow. Now, I must tell you realistically that was the one thing in our trials that didn't seem to become a problem with regard to excluding patients. I've only heard of a couple of patients who were excluded on that basis.

With regard to soft tissue radiation therapy, we will include in the package insert an actual table of the mean, median, and also range of radiation absorbed doses by millicurie administered so that a patient who had received external beam could have his radiation oncologist be aware of that information. We have not found a problem with regard to prior external beam radiation therapy or, for that matter, patients who have had subsequent external beam radiation therapy. The agency asked us specifically to look for any adverse events that occurred with subsequent external beam radiation therapy and a specific CRF form was designed for that purpose, and we have seen no problems up to this time.

The second part of your question was with regard to retreatment. We do not advocate retreatment at this time, and again in a package insert it would specifically say that this is for single course therapy only. Retreatment is another area that needs to be studied, and actually the National Cancer Institute, under a clinical trials agreement that they have with us, has moved ahead and approved Dr. Witzig and Dr. Wiseman at Mayo Clinic to look at a retreatment trial. That is not a retreatment trial at relapse but rather a retreatment trial that would look at a second dose of therapy approximately 12 weeks or so out when the patient was hematologically recovered and would be a dose escalation trial. But we feel that there needs to be considerably more study before we would have enough information to understand retreatment. It would be excluded.

DR. BLAYNEY: Finally, could you expand a little bit on the coordination? It strikes me if you're obtaining a radioactive isotope from one vendor and a monoclonal antibody from IDEC, getting these to places that may not be as experienced as your investigative sites represents a challenge for stability as well as preparation in the radiopharmacy lab.

DR. WHITE: Right. It's actually less complex than it seems. Most hospitals in the United States have a relationship with a commercial radiopharmacy. That's the radiopharmacy that brings them their technicium and their indium and their gallium for their regular use every single day. For all of those hospitals who have a relationship with a commercial radiopharmacy, it would be the commercial radiopharmacy probably who would call up and order. They would call one phone number. Once they called that phone number, a kit would be shipped -- this is a cold kit -- from IDEC to the radiopharmacy and from our supplier of yttrium, at a different location, the yttrium would be shipped just like any other isotope by a radioisotope manufacturer to the same address to arrive on the appropriate day.

For the indium component, however, since most of these radiopharmacies and hospitals have already relationships with the two suppliers of indium that are approved for supplying indium -- that would be approved and were in our clinical trials for supplying indium -- they would make a separate phone call to order the indium in the same way that they do for other indium needs.

Then at the radiopharmacy would be the compounding of the product, which is about a 5-minute process. We have had experience now -- in fact, maybe we can bring up that radioincorporation slide -- with 533, I believe, radioincorporations at many hospital and radiopharmacy sites, I think 40 or 50 in this group, and 98 percent of the doses, 523 out of 533, at these 40 sites exceeded the 95 percent release specification of the ITLC, which is a routine type of QC study at a radiopharmacy. The mean radiochemical purity is 98 percent with a standard deviation, and the median radiochemical purity was 98.6 percent. So, in actuality this will be easier and less complex than it seems.

DR. BLAYNEY: What about the stability after the isotope is mixed with the chelator?

DR. WHITE: Yes. It is stable for 8 hours after it is radioincorporated, and that is, of course, within the time frame that it would be used. The radiopharmacy, if it was off-site -- it could be on-site, it could be off-site -- it would compound that morning and then, on its usual run or a special run, drive it up to the hospital or outpatient nuclear medicine department that it was being used at. And then it would be infused after the rituximab was completed, and the rituximab could be completed either at the oncologist's office and then the patient come to the nuclear medicine department for the Zevalin component, or if there were provisions and the usual approach was at the hospital, then that could be done there.

DR. BLAYNEY: Thank you.


DR. PELUSI: My questions were along the same lines as Dr. Blayney's in terms of this coordination. So, you just mentioned then that the patient could come into the oncologist's office, then go to the nuclear medicine department, and then return back to the oncologist's office. Is that how you see that occurring?

DR. WHITE: Yes. Since the oncologist really has experience administering rituximab and has all the provisions for administering rituximab, the patient would go to the oncologist's office. They would have the rituximab infusion. Then they could either be hep-locked or the line could be taken out, and they would go across the street or down the street, or whatever, to the nuclear pharmacy where they would receive the indium-labeled Zevalin on the first day and seventh, eighth, or ninth day the yttrium-labeled Zevalin in the same way. Of course, we would expect that the oncologist would be the one following the patient after administration of the agent.

DR. PELUSI: I was just again looking at the safety issues in terms of some of the smaller community hospitals when many people from the oncologist's office are not necessarily in the closest proximity to some of the hospitals in terms of timing as well. That could be an issue.

DR. WHITE: Yes. The majority of our patients had treatment with the radiolabeled Zevalin within 4 hours of the completion of the infusion of the rituximab. We had a small experience at a greater time interval, but at least 4 hours are reasonable.

DR. NERENSTONE: Dr. Przepiorka.

DR. PRZEPIORKA: Thank you.

Can you tell us a little about the short-term or, if you have information, the long-term effects of the radiation dose to the testes?

DR. WHITE: Yes, I can. First of all, let me start by saying that our experts, both Dr. Stabin and Dr. Sparks at Oak Ridge Institute of Science and Education, who did our dosimetry, as well as Dr. Wiseman, who is both an oncologist and nuclear medicine physician and professor at Mayo Clinic, tell us that the methods that we use overestimate the dose to the testes because they assume that the organ that you're estimating is in the center of the body, and they assume that there has been attenuation by the tissues on top of the organ. Well, of course, that is not the case with the testes. I think that we may, as time goes on, do some more complex dosimetry to better get at that dose.

That dose, of course, is a significant dose, although low-dose biologic radiation from an isotope is not the same as external beam. Still, with that dose, one might ask the question as to whether there would be an effect on sperm or on hormone levels. A lot of lymphoma patients who have had chemotherapy beforehand will be oligospermic to begin with and may have lower levels of androgens.

However, what we proposed to do is, in some of our just planned and newly begun investigator-initiated IND trials, to look at baseline both sperm counts and androgen levels and look again after Zevalin to see if there's any biological effect and if it's important to have any therapeutic intervention in these patients.

DR. PRZEPIORKA: You have given us the dose of the drug in millicuries per kilogram. Can you give us an idea of how much antibody that is in mgs per kg?

DR. WHITE: Yes, I can. It's less than 2 milligrams of antibody.

DR. PRZEPIORKA: Your predose of Rituxan was based on some studies you had done in the phase I/II, if I read that correctly, where you found that with 100 milligrams per meter squared, you got 92 percent targeting to the tumor. Then in your briefing document you said, so we decided to use a dose of 250 milligrams per meter squared, but in treatment you actually used 500 per meter squared because they get two doses of a humanized antibody very close together.

Do you have any dosimetry to determine whether or not that huge dose of Rituxan you give before the therapeutic dose reduces your targeting?

DR. WHITE: Yes. Let me answer that question and maybe we can pull up the pretreatment antibody slide.

First of all, what we compared in our very first phase I study was no pretreatment antibody. This was done by Dr. Susan Knox at Stanford with 1 milligram per kilogram, or approximately 70 milligrams, of cold ibritumomab, and that was compared to 2.5 milligrams per kilogram, or approximately 175 milligrams, of cold ibritumomab. In that study, dosimetry and imaging did improve with increasing pretreatment antibody dose.

Now, as I had said a little earlier, we wanted to move to trying the rituximab rather than the ibritumomab prior to the radiolabeled Zevalin. So, we began with a similar dose of rituximab, about 170 milligrams, but this time we were dosing in milligrams per meter squared just because that is the convention for rituximab. So, 100 milligrams per meter squared. Then we escalated further to 250 milligrams per meter squared or 425 milligrams. Here we saw no difference. So, in other words, there was an improvement as we went up, but here there was a plateau, no difference in dosimetry or imaging.

So, we needed to choose one of these doses in order to improve biodistribution and minimize HAMA, and we decided to choose the higher dose for a greater contribution to therapeutic effect. But this was not just a single dose. These were before each Zevalin dose in every case, in this case, in this case, and in this case. So, this is what we had explored.

Did that answer your question?

DR. PRZEPIORKA: Directly. Thank you.

You had indicated that there were several patients who had blood stem cells collected after Zevalin. Can you tell us a little bit about whether or not they were transplanted and whether or not they engrafted or whether or not they were actually able to collect stem cells after Zevalin?

DR. WHITE: Yes. We know of 8 patients who have had stem cells collected after Zevalin to date, and 6 have already been transplanted and all engrafted well. 2 are waiting for the right time to undergo transplant. We have not heard of a single instance -- and we have inquired -- of a failure to be able to harvest stem cells. As I said, in those 6 patients who had stem cells harvested after Zevalin and 3 who had already had stem cells banked before they ever had Zevalin, all 9 of those patients engrafted well.

DR. PRZEPIORKA: Finally, you included in your patients under the definition of refractory those who had no response to prior therapy, as well as those who failed within 6 months. Now, that is a common way to describe a refractory patient for standard chemotherapy, but it may not hold to be true with Zevalin. Are the response rates in those two groups the same?

DR. WHITE: I actually have a slide on that too. That would be in the 106-06 trial, responders versus nonresponders. It was about a duration of response of about 4 months because we used a TTP of less than 6 months as our parameter, and we divided it down and took a look at that, rituximab nonresponders versus the ones who were the short responders with a TTP of less than 6 months. And there were comparable response rates, a slightly smaller number here, 51 percent versus 67 percent. In the rituximab responders group, though, there were 17 percent CRs and there were none in the rituximab nonresponders. So, there may be an effect there.


DR. LEVINE: Going to the transformed patients, you've got a total of 15 patients with transformed lymphoma and that's one of the indications you're interested in. On your randomized trial, the Rituxan group, 3 out of 4, which is interesting, 75 percent response rate on the Rituxan, not different than the Zevalin. Can you discuss that in terms of your request for indication here?

DR. WHITE: Yes. Moving forward, when we planned that trial, we expected to have larger numbers of transformed patients. We did have transformed patients on our dose-reduced trial as well and just a couple in our phase I/II trial, and you saw the total experience of those 15 patients with a 40 percent response rate and a 13.3 percent complete response rate. It is a small population of patients, without question.

These patients, though, are very, very needy patients. By one series, the median survival is 7 months, so they transform and die rapidly. Some are chemotherapy resistant. So, it may be that when a physician weighs the potential risk and benefit of a treatment, as physicians always do, that they may find in that type of a patient who was critically in need of therapy that timing was critical, that they may want to choose a therapy that had a higher chance of a response and that that outweighed the toxicity. So, we wouldn't want to preclude these patients, although we certainly acknowledge that the body of evidence is considerably smaller.

As far as the rituximab response, we were, of course, pleased to see that the rituximab response was good, but it was just 4 patients.

DR. LEVINE: Were any of those transformed patients any of the patients who subsequently had bone marrow stem cells or peripheral blood stem cells harvested, or were any of those individuals those who had subsequent chemo? And how did they tolerate that?

DR. WHITE: None were in the transplant group, but I do not have at my fingertips what subsequent chemotherapy they specifically had. We could maybe try to take a look at that and see if we can come up with something by the end.

DR. LEVINE: Another question. Was there any difference in the prevalence of HAMA in those who had Rituxan prior in your trial with the Rituxan-refractory patients? Any difference in the patients who received the product after having had Rituxan earlier?

DR. WHITE: I see your question.

DR. LEVINE: Assuming that people will have multiple such products over time.

DR. WHITE: Yes. Again, the patients who had prior Rituxan, if they had developed an antibody to prior Rituxan, it would have been a HACA, and that rate is known to be less than 1 percent.

But across all trials, each time we resummarized the HAMA rate, when we had some trials open and some trials completed, sometimes not, ever since we switched to rituximab, it has always been less than 2 percent, and I do not believe there was any preponderance on any one trial.

DR. LEVINE: I'm sorry. One more. It was a similar kind of question as to the real Rituxan resistance. You also defined chemotherapy resistance as resistance or a response less than 6 months. If you look at the patients who really were resistant to chemo and their subsequent response to Zevalin, do you have those data?

DR. WHITE: I'm sorry. I do not have those data.

DR. LEVINE: Thank you.


DR. SLEDGE: Three questions. First, can you show us survival curves in the randomized trial?

DR. WHITE: Yes, I can. Can we bring up survival curves?

Thankfully, there have not been that many deaths yet on the randomized trial, but we do have survival curves. You have to temper your examination of these survival curves, which you can see are exactly the same, by the fact that over half of the patients on the rituximab arm who did not respond to rituximab went on to receive Zevalin on another trial.

DR. SLEDGE: Thank you.

Second, do you have any data on crossover response to rituximab for a patient who has had prior Zevalin?

DR. WHITE: I am aware of a couple of patients who have had prior Zevalin and have gone on to rituximab who have responded and also a couple who have not. I don't have any definitive, quantitative data on that.

We did collect information on the first subsequent therapy, but not all subsequent therapies. And the numbers of patients who had rituximab as their first subsequent therapy was low.

DR. SLEDGE: Third, actually a question for some of your investigators here. This committee always wrestles with the issue of clinical benefit, and typically we've not thought of response per se as being a clinical benefit but rather a survival endpoint or some symptomatic benefit. What's the clinical benefit here, if I may ask some of your investigators?

DR. WHITE: Yes. Let me open and then I would like to call some of our investigators to the --

DR. SLEDGE: Actually I'd like to hear from the investigators.

DR. WHITE: Okay. Thank you. I'm sorry. Dr. Witzig is here. Dr. Witzig from Mayo Clinic has experience with more than 100 and now with the open label trial approaching 150 patients, and I think maybe he can speak to net clinical benefit, maybe followed by Dr. Leo Gordon who also has a very large patient experience.

DR. WITZIG: I think the big clinical benefit is it's a single-dose therapy, and these patients were all heavily pretreated and many of them have had many chemotherapy recipes. Now they come in and get a one-day treatment with virtually no side effects like nausea, vomiting, hair loss, and the next day they go back to work and go back to their usual activities. So, from a quality of life standpoint and patient acceptance standpoint, it's been remarkable.

In addition, we're seeing a lot of complete remissions, much more than Rituxan, 30 percent versus 16 percent. As you've seen on some of those curves, those CR patients are really doing quite well for a number of years and not requiring any further treatment.

So, I see it as a major advance in the treatment of lymphoma. It offers these people something new and it's very well-tolerated and a very high patient acceptance rate with a high CR rate.

DR. WHITE: Thank you.

And Dr. Leo Gordon, who is Cancer Center Director and Chief of the Department of Hematology and Oncology at Northwestern University.

DR. GORDON: Thank you. I can pretty much echo Dr. Witzig's comments. I think these are patients who are progressing on prior therapy, symptomatic, and when you're seeing a 30-some percent complete remission rate and an overall response rate of 70 to 80 percent, those are symptoms that are disappearing. These are people who are requiring therapy and with really 2 weeks of treatment with remission that lasts for a long time. There is significant clinical benefit from the investigator and from the clinical oncologist standpoint with what I think is fairly acceptable minimal hematologic toxicity. So, I think there's a significant benefit.

DR. WHITE: Thank you.

Maybe Dr. Sandra Horning, who is the Chair of the Lymphoma Section of the Eastern Cooperative Oncology Group, could also approach the microphone.

DR. HORNING: Well, I would really echo the comments that have been made previously and just encourage you to think about the patients with these disorders and primarily the low-grade lymphoma patients for whom, it is true, we have not defined curative therapy, but the opportunity to have a series of therapeutic options that are effective that can be used sequentially over time, particularly those that offer good quality of life are very meaningful.

In fact, there are data from our own institution that indicate that in the 1990s and up to the present that patients with follicular lymphoma are actually living longer without a defined curative therapy, and I think it's a tribute to the entre of monoclonal antibodies. I think that Zevalin adds to this armamentarium.

DR. SLEDGE: Thank you.


DR. ALBAIN: I've been in the question queue for a while and my question was the same as Dr. Sledge's regarding clinical benefit.

So, just to follow up, are you seeking conventional approval or accelerated approval for this application?

DR. WHITE: Maybe our regulatory staff member, either Leslie Shelly or Alice Wei, can address the regulatory approval.

DR. SHELLY: We will be seeking conventional approval.

DR. ALBAIN: So, with that in mind, do you have any hard quality of life data or other supportive data to go along with what we've heard from the three investigators regarding translation of this very remarkable response rate into other measures of clinical benefit?

DR. WHITE: If we can bring up a quality of life slide.

We attempted to have a quality of life secondary or tertiary endpoint and collected FACT-G quality of life instrument data. Unfortunately, compliance was not 100 percent. Only 45 patients out of the 73 on the Zevalin arm and 36 out of the 70 on the rituximab arm actually completed baseline and follow-up questionnaires so that we could compile this data.

The baseline level on FACT-G was lower for the Zevalin arm than the rituximab arm. In both cases, there was an improvement in FACT-G. The improvement was statistically significant for the Zevalin arm, not for the rituximab arm, but we did not emphasize this data as we felt, along with our discussions with the agency, that these incomplete samples were not adequate to truly focus on this data as a quality of life endpoint.


DR. GEORGE: I have a couple of design and results questions. On the TTP endpoint, I believe it was stated that this was a secondary endpoint looking at equivalence. The problem I see, was this defined ahead of time? What did you mean by equivalence?

DR. WHITE: Yes, it was prospectively defined. At the time that this phase III protocol was designed, in collaboration and in agreement with the FDA, we chose overall response rate as an adequate and appropriate primary endpoint. At that time we knew that the target population of 150 patients would not permit a comparison of TTP, so TTP in all and TTP in responders was a secondary endpoint. And it was specifically stated in the statistical section of the protocol that the trial was not powered to compare TTP, but instead that we sought clinical equivalence, not statistical equivalence, and that was defined in the protocol prospectively as a median plus or minus 1.5 months. So, it was felt at that time that the overall response rate was an appropriate endpoint.

DR. GEORGE: I was just doing some back-of-the- envelope calculations. You haven't ruled out a decrement in this, of course. You haven't ruled out a decrement in the time to progression by the usual kind of standards.

The other question I have concerns the refractory trial. There also you had a target overall response rate of 35 percent.

DR. WHITE: Yes, and that was just chosen clinically. Just in discussion with our 17 investigators who were involved in that trial, we said in a rituximab-refractory population, what type of a response rate would you like to see as a clinically meaningful, important response rate in the rituximab-refractory population, and a target of 35 percent came from that discussion.

DR. GEORGE: So, it was historical just by survey.

DR. WHITE: Yes, in terms of known response rates in this relapsed and refractory disease.

DR. NERENSTONE: Ms. Krivacic.

MS. KRIVACIC: Can you talk about the AEs, or adverse events, that you possibly saw during some of the growth factors that were given to the patients for supportive care?

DR. WHITE: I'll ask Dr. Multani to address that question.

DR. MULTANI: 18 percent of patients received some sort of growth factor after Zevalin therapy. 13 percent of patients received a neutrophil growth factor, G-CSF, and 8 percent of patients received erythropoietin. Some patients received both. We didn't have adverse events that were attributable to the growth factors themselves, however.

DR. NERENSTONE: I have a brief question. I'm a little bit concerned about the thrombocytopenia that we're going to be seeing. I think the community oncologists are very comfortable with neutropenia and we have lots of antibiotics and we know what to do with that. The mean nadir platelet counts of 41,000 in your nonhematologically impaired patients and down to 24,000 or significant numbers for a prolonged duration would make some of us in the community concerned.

In addition, your two fatalities were intracranial hemorrhage.

Are you going to have any recommendations about concomitant Coumadin therapy in these patients?

DR. WHITE: Yes. In fact, after the events of intracranial hemorrhage, after the first one, we notified the physicians of that event, and when the patient had the traumatic intracranial hemorrhage who was on Coumadin and unfortunately also on self-prescribed ibuprofen, we did amend all of our trials to exclude patients who were on Coumadin. And we would propose that in the labeling we could either exclude or at least certainly caution physicians that, if at all possible, patients should be taken off of Coumadin at least at the time when the nadir is expected.

The nadir, by the way, is pretty reliable in terms of the timing down and timing up, so it can be anticipated.


DR. TAYLOR: Along those lines, though, did you give prophylactic antibiotics during this time? You have a very high percentage of grade 3 and 4 neutropenia.

DR. WHITE: Prophylactic antibiotics was at the discretion of the physician, and there were 10 percent of patients who were given prophylactic antibiotics.

DR. NERENSTONE: Dr. Blayney.

DR. BLAYNEY: It strikes me that one other way to accomplish what you have set out here is, because you got some response rate with your humanized antibody, rituximab, and you're shrinking bulk tumors with radiation, if I say to you why can't I get the same response with rituximab/external beam radiation and not have my patient pay the price of a 20 percent chance of platelet transfusion and 20 percent chance of red cell transfusion and the nuisance of growth factors, et cetera.

DR. WHITE: I might ask one of our investigators to also address this. But with rituximab, of course, there is a 48 percent response rate and a 6 percent complete response rate. External beam radiation to a single mass or more than one mass would be expected to further shrink that. But it would not cause 30 percent of the patients to enter complete remission most likely.

Maybe Dr. Witzig or one of the other investigators might want to address that specific question.

DR. WITZIG: Well, I think almost 80 to 90 percent of our patients had stage 3 to 4 disease and a lot of them had marrow involvement. So, giving them external beam radiation would not have controlled the entire disease.

In addition, one of the remarkable things we've noticed is even in patients with large abdominal masses, you can deliver a high-dose radiation this way with virtually minimal toxicity to adjacent bowel or bladder, other normal organs, which in my experience would not be possible with traditional external beam radiation therapy.

DR. BLAYNEY: Thank you.

DR. WHITE: Any other investigator that wants to address that.

(No response.)

DR. NERENSTONE: Well, thank you very much. Not seeing any further questions from the committee, we will break. If people can be back for our next presentation at 10:15. Thank you.


DR. NERENSTONE: If the committee could please take their seats.

DR. TEMPLETON-SOMERS: Excuse me. I think most of you already know that there's been a terrorist attack in New York and in the Washington, D.C. area. It sounds like several different locations and many, many people are injured or killed. All flights are grounded. So, nobody is going anywhere for a little while. The beltway is closed. Government is closed. Those of you that the FDA has paid for your travel in will obviously pay for your hotel as long as we need to and any other travel problems.

I'm not sure what else there is to say. We'll finish the Zevalin session since we are part way through, and then we'll re-evaluate the situation at noon and see if everybody is here for the Gliadel session. Thank you.

DR. NERENSTONE: I'd like to continue with Dr. Meredith who is going to be talking about normal tissue effects of radionuclide therapy.

DR. MEREDITH: Thank you.

We don't know as much about the tolerance of normal organs to radionuclide therapy as we do to external beam. This has just not been as well studied. So, in my presentation today, I'm going to make a comparison as to where it is today.

Before I give you some numbers that actually compare the two, I'd like to give you some qualifiers as to how these numbers were derived. The external beam data that I would enumerate is a tolerance dose 5/5 or 50/5. This means a 5 percent or a 50 percent risk of severe, late complications by 5 years. In comparison to the radionuclides, I've listed virtually any toxicity that's been noted, whether it was acute, very transient or whether late.

The numbers of patients for external beam is obviously much larger in most reports than that for radionuclides. Many of these are derived from studies with less than 50 patients. In the external beam data, most of this is derived from patients who had possibly surgery but no other modalities, for the most part. And for the radionuclides, most of these have failed numerous other therapies, including multiple chemotherapy regimens.

For external beam data, this is highly fractionated at 2 gray per day 5 days per week, whereas most of the radionuclide data is a single dose. And we know that a single dose of external beam is more toxic than fractionated in most instances.

It's high energy external beam data versus the radionuclides is a lower energy. External beam is also a high-dose rate compared to a low-dose rate and usually an exponentially decreasing dose rate with radionuclides.

For the external beam, I will restrict this to whole organ toxicity, although we know in most organs that partial organ irradiation is much better tolerated and at higher doses than whole organ toxicity. For the most part, whole organ toxicity is applicable to radionuclide therapy.

To give you some numbers, bone marrow toxicity has been the dose-limiting toxicity in most radionuclide studies. You can see from radionuclide information that very little dose to the bone marrow, from some of our studies using lutetium-177 labeled antibody, results in some toxicity. With I-131 labeled antibodies that have a circulation time of about 2 days, the MTD has been about 185 centigray. All these numbers are listed in centigray, and these numbers are for non-marrow targeting therapy and for diseases that did not have significant marrow involvement.

For thyroid, we know that there's a great difference between the TD5/5 and the 5/50 for external beam. We don't have real good information on the radionuclides. Part of this is that the TSH can increase especially after I-131 therapy, and patients are placed on Synthroid. So, we don't really know how many of these may become hypothyroid if they weren't treated.

The kidney is one of the more radiosensitive organs and you can see there's not a lot of difference between the dose for a TD5/5 and a TD5/50 for external beam radiation. Again, this is for severe complications, whereas for the radionuclides, I've listed studies where any toxicity has been noted. This is 2 patients who received greater than 2,170 or 1 patient that received somewhere up to 3,100. In this case it was less than 5 percent of the patients in these studies and there was some increased creatinine. We don't have long-term follow-up.

In terms of bladder, it's more radioresistant than many other organs, and for radionuclides some hemorrhagic cystitis has been seen somewhere between 4,000 centigray and a much higher dose. There will be more information on this shortcoming in the literature from studies of holmium.

The lungs are relatively radiosensitive. I listed here a total body radiation single-dose tolerance, which you can see is about half that of fractionated radiation. When this has been looked at as a dose-limiting organ for myeloablative studies of radiolabeled antibodies, we see that the place where we're seeing toxicity is about the same actually as with external beam, and these are instances where no chemotherapy was given at the time of the radioimmunotherapy, although most of the patients had had prior chemotherapy.

For the brain, we don't have a lot of good information from radionuclides in terms of whole brain. In fact, in many instances where treatment has been given either to a cavity after a tumor was resected or into the tumor itself, the dose to the normal brain is felt to be quite low and complications have generally been some instances of edema and rarely headache or seizure. In each case you can see that when radionuclide is given into a tumor resection cavity, relatively high doses can be tolerated and very high doses have been tolerated without significant toxicity when the radionuclide is injected directly into a tumor.

I've listed here that the meninges are probably more resistant than normal brain because for external beam you really can't separate the two and so I could not find a separate number for meninges.

Our information from radionuclides here generally comes from intrathecal administration, and this is usually after external beam radiation. Complications have been few with relatively high doses. This has mainly been transient aseptic meningitis.

As much as 160 millicuries of I-131 labeled antibody can be given as a single dose, and cumulative doses of over 300 millicuries. From some of these, the surface dose was calculated to be close to 6,000 centigray.

We really don't have radionuclide reports for spinal cord.

For stomach and intestine, these have about the same toxicity from external beam and with radionuclide therapy. Here I've listed some nausea. In most cases there are a few reports of nausea at less than 2,700 and generally it's mild until you get about the level of 6,000, which you can see is greater than the normal toxicity from external beam. There are some cases of grade 4 diarrhea, and in those cases patients got a relatively high dose, in fact much greater than tolerated by external beam.

In terms of bowel serosa, again for external beam you can't separate it from the mucosal surfaces, but there is data to suggest that the serosa itself is more tolerant than the mucosal surfaces. For radionuclide therapies, we know that there is more toxicity with P-32 than has been tolerated by targeted therapy with radiolabeled antibodies. This may be because you have a more uniform dose with the P-32 and this is more targeted to tumor areas. In fact, we know that 8,000 at tumor deposits can be tolerated with rare complications, including adhesions or some GI complaints.

Liver is one of the dose-limiting toxicities in myeloablative studies. And again sometimes at relatively low doses, mild nausea has been noted, but in general, in terms of dose-limiting toxicity, this has been around the dose of 2,400 by yttrium-90 radioimmunotherapy and higher doses, at least for some patients, with I-131 therapy. When the radionuclide has been combined with other regimens, including chemotherapy and total body radiation, as preparation for myeloablative therapy, you can see that the tolerance is a little less, although if you add the external beam plus the radionuclide therapy, it turns out to be about the same as a radionuclide alone.

Pancreas. I did not find something in the literature for the TD5/5 or 5/50, but from general practice I can tell you that the TD5/5 is greater than 4,500, and if you get up to about 6,000 even to small areas of the pancreas, you get in trouble. We don't have much information from radionuclides in terms of the whole pancreas, but very high doses have been given to small areas of the pancreas when this is injected into the tumor itself, a P-32 regimen giving more than 1 million centigray to small areas. This has been well-tolerated and dose fractionation of this has been tolerated.

I'm going to quickly go through a few other aspects here in terms of how accurate are these radionuclide dose estimates and can we compare doses between studies.

First off, radionuclide dosimetry is generally much less accurate than external beam. I'll go into how accurate are tracer studies, that the calculated dose is not necessarily the dose that correlates well with the biologic effect, and how accurate are comparisons between radionuclide dose estimates.

In general, radionuclide dosimetry is just less precise than external beam. I believe Dr. White gave an explanation this morning. As an example, I just said if you had a parenchymal lung tumor, there would probably be no attenuation correction for radionuclide dosimetry, whereas for external beam, you would take into consideration that there is less attenuation in the lung tissue itself and more dense tissue at the chest wall.

Also, with external beam radiation, as soon as you turn on the machine, you get full, immediate dose to all organs, whereas with radionuclide therapy, in general there is a buildup of organs and then an exponential decrease.

You would think that tracer studies where a small amount of a radiolabeled agent is given and dosimetry and imaging performed would be the best indicator for the later therapeutic dose. And we have a few studies in which tracer studies were done and then dosimetry was repeated with the therapeutic study, which gives us a basis for a direct comparison in individual patients. I've listed some examples here. One is from our own institution where we used I-131 LYM-1 for non-Hodgkin's lymphoma, and the variance between the tracer dose and that for the therapeutic was between .9 and 1.38. This was using the tracer 7 days before the therapeutic administration.

The University of Washington, which has done myeloablative studies using I-131 labeled antibodies for non-Hodgkin's lymphoma and leukemia, have found a correlation here between .67 and 1.15. This is the example I gave for their effective biological clearance from lung. They have done this for multiple organs, and they're on the same ball park.

Another example is from an indium-labeled antibody used for CEA positive tumors, and the reported data was a concordance between .6 and .99 for most normal organs.

In a study conducted in breast cancer patients at M.D. Anderson using I-131 labeled CC49 and the patients received interferon after the tracer dose, it was noted that there was a mean increase in the whole body residence time and this was felt to be an interferon effect. However, I would like to notice that the percentage difference here between the tracer and the therapeutic is in the same ball park as the others, and in this case they can predict that the whole body residence time will be increased; whereas on the other tracer studies, it's not predictable on an individual basis whether your dose estimate will be an under-estimate or an overestimate.

I'd like to say here that calculated dose, that is, the number given for radiation exposure is not the same as the biologic dose because we know that there are a lot of physiologic and biologic interactive factors. For radionuclides, among these include a very heterogeneous distribution at the cellular level at least, even if it looks relatively homogeneous on imaging. There are dose rate effects. There's an effective range of radiation, such as some radionuclides may have a range of four or five times penetration in tissue as others. Radiation biologic effectiveness or relative biologic effectiveness varies and there are other characteristics that need to be taken into account.

Among biologic factors that affect tolerance, we know that some things are important, and there may be different degrees of impact from the various factors. But some of these can include age, prior therapies, the time since prior therapy, the disease status, is the patient anemic, do they have marrow replacement from disease, from fibrosis, other factors. There are genetic and physiologic factors or conditions that are important. Such as, hypoxia can decrease radiation effectiveness. There are also genetic and physiologic factors that can affect repair from radiation damage.

Some of the factors that are felt to be important have been analyzed at least briefly, and adjustments for biologic factors have been found to improve correlations between radiation doses and effects. One of these studies looked at prior chemotherapy and the time since prior chemotherapy, and found that in general if it was less than 3 months from prior chemotherapy, the patients did not tolerate the radionuclide therapy as well.

Wessels looked at a number of factors, including age, gender, prior radiation, prior chemotherapy, and found that if he developed a formula that took into effect some accounting of this and applied this to the radiation dose, that the correlation between toxicity and the reported dose was improved with r value going from 0.57 to 0.80.

There are also agents that can impact on the biologic effectiveness of radionuclide therapy that don't contribute to dose estimates. There are a number of agents or factors that can contribute to this. Chief among them in the studies may be chemotherapy, but also other biologic response modifiers such as radiosensitizing agents, including BuDR; cytokines, interleukin-1 and interleukin-2, have been studied to some extent; and growth factor inhibitors such as antibodies to the epidermal growth factor receptor or anti-tyrosine kinase inhibitors.

As an example of this, a study of radiolabeled chimeric L6 antibody was given as therapy for breast cancer xenographs in a study published by the DeNardos. In this case, the patients received the radiolabeled antibody alone. There were 79 percent responses but no cures, but the addition of Taxol 6 or 24 hours after this, the same dose, resulted in an increase in response rate and nearly 50 percent of the animals having cures.

How accurate are comparisons between radionuclide dose estimates? There are a number of things that one can consider in terms of trying to compare doses from different studies and different institutions. Some things that one may want to take into account are variance in dosimetry methods, such as did they use measured organ volumes such as in the myeloablative studies at the University of Washington, or are these phantom studies based on a single model such as one-size-fits-all for a male of adult size, one-size-fits-all for a female adult?

Do the calculations use computer programs? Are they the same programs, such as MIRDOSE 2 or MIRDOSE 3?

Was attenuation correction applied for the regions of interest or a transmission scan technique used?

Was there background subtraction?

What was the frequency and appropriateness of data collection? Some examples are if the peak concentration is missed in a normal organ such as liver, the dose may be lower than normal. If there were no early scans and there was an assumption that there was immediate full dose, then the dose estimate will be over that which actually was received.

How accurate are the doses reported in the literature? Barry Wessels looked at this and finds that in the 1980s, in terms of reporting marrow toxicity based on the radiation dose reported, it was about a 700-fold difference. By the 1990s, it was down to 200 percent and now it's about 30 percent.

In his recent analysis, he actually took data from seven institutions and recalculated it using each institution's own data, and his calculations varied from what was reported from minus 35 percent to plus 6 percent.

An example of this also, in terms of looking at methods, such as the bone marrow, the AAPM task group report by Dr. Sgouros looked at a concentration factor of .19 for radioactivity in the blood as that portion of the blood in the marrow that should be taken into account for the dose estimate. He came up with a dose estimate of about 200 centigray versus another report of a similar agent where they looked at the radioactivity in the whole blood, and the dose estimate was approximately three-fold greater.

Dr. Mills has asked me to look at when does imaging or dosimetry have a great impact, when should it absolutely be done. So, some things that may make a difference.

One is when there is good correlation of the data between organ toxicity or antitumor effects with the radiation reported.

The second is when there's a normal organ that can be somewhat accurately assessed and may be dose-limiting such as in the myeloablative studies for organs such as liver and lung.

As a couple of examples of this, this comes from a myeloablative study at the University of Alabama at Birmingham, and you see here that each of these bars represent dose to individual patients to the whole body, liver, or spleen. You can see that despite different size of patients and different amounts dosed, that there's not much difference in the whole body dose received, but there's a great difference in the individual organ doses for liver and spleen.

This example also comes from the University of Washington studies using high-dose I-131 labeled antibody for lymphomas and leukemia. In this case, they looked at the amount of radiation per unit of injected activity and compared this to the injected activity per body weight. You can see there's a large spread here. Now, the tracer study showed about a 30 percent plus or minus difference. Here you have a very wide range for individual patients.

Other areas that it may be very important to do imaging and dosimetry may be when tumor is adjacent to a normal critical organ and there's a high chance that the tumor will receive a high dose that would affect the normal organ.

Areas where distribution may be highly variable. One antibody against CEA studied in this regard had variable distribution between patients, and it was noted that colorectal patients generally had twice the clearance rate as those with other types of diseases.

Another thing would be when the distribution is unknown for a particular agent and the distribution could be critical. An example here is some of our work at the University of Alabama at Birmingham giving radiolabeled antibodies into the peritoneal cavity for ovarian cancer. You can see on the left that when a trace element was put in, all the activity is loculated here in the lower part of the abdomen and there's nothing up in the rest of the abdomen. This patient has catheter revision, and you can see on later image that this is distributed well through the abdomen.

In this one you can see that these sort of look like loops of bowels. This is another patient that had a test dose before radiolabeled antibody. Indeed, the catheter had eroded into the bowel, and this patient could not be treated until that catheter was removed, the area healed and a new catheter placed.

There's concern about the addition of radioimmunotherapy or radionuclide therapy with external beam. There's not a lot of information in the literature, but some of it goes back to the early days of radioimmunotherapy when hepatomas were treated. In this instance, 2,100 centigray were used of external beam radiation concomitant with Adriamycin alternated with 5-FU and the patients also received Flagyl. Then they received 2 months of intensity increased chemotherapy. Subsequently they received radiolabeled anti-ferritin which delivered a calculated dose of 400 to 1,000 centigray to normal liver. So, if you add this dose to their external beam, this was well tolerated and is within the reasonable numbers that were noted with external beam radiation alone.

Other information comes from myeloablative studies, and I've taken this first one as an example from leukemia studies at the University of Washington where patients received between 76 and 612 millicuries of I-131 anti-CD45 antibody. This was delivered in conjunction with cyclophosphamide and 1,200 centigray total body radiation as preparation for bone marrow transplantation. The MTD of the radiolabeled antibody was at 1,050 centigray, and again this is added to the external beam given concomitantly of 1,200 centigray.

Some of our studies at the University of Alabama at Birminghman treating breast and prostate cancer patients with I-131 labeled CC49 therapy included a higher dose of total body radiation with the cyclophosphamide. In this case, the liver dose was slightly less than that reported at the University of Washington. Some liver toxicity was seen. However, the patients that had liver toxicity are those that also got thiotepa and those that had had many cycles of chemotherapy prior to this intervention. Again, the toxicity seen was increase in liver function tests and these were very transient.

In terms of tissue tolerance to retreatment, from animal studies it's found that some acutely responding tissues can tolerate a full course several months later. In terms of late responding tissues, there's virtually no recovery, however, in these studies for the heart, the bladder, or the kidney. There was partial recovery for the skin, mucous membranes, the lung, and spinal cord.

We don't have quite as clean data for human studies, but we do have a fair amount of practical experience in giving retreatment of external beam radiation. Some of these examples include head and neck studies at our own institution, as well as others, where patients initially receive a tolerance dose to the head and neck of at least 70 gray. At least 6 months later, they may be treated as salvaged and can get pretty close to a full dose of radiation again with concomitant chemotherapy. This has resulted in reasonably good 2-year control, and there is some increased fibrosis, but this has been done without severe complications such as fistulas or necrosis of bone.

Spinal cord. Most people would usually tell you that it remembers all the radiation and cannot be retreated. However, most recent information from animal studies and selected human data indicates that there is some partial recovery with time and that after about a full dose of around 40 gray, some patients have tolerated almost the same dose again without developing myelitis.

In terms of prostate cancer, many patients have been treated with a full dose of 70 gray or higher and then salvaged for local recurrence by implants which deliver more than 9,000. Again, this has been done without severe toxicity when a time period elapses between these two.

Nasopharynx cancer has been treated twice. In most instances, patients get a full dose the first time and then if they relapse. The M.D. anderson has found that you can give a total of 10,000 centigray with only a 4 percent complication rate, and Massachusetts General has used another 6,000, sometimes with brachy therapy, to result in a 50 percent survival without a large complication rate.

There's not as much data about radionuclide retreatment, but there are some instances in the literature. I've listed here that strontium-89 has been given more than five times. I believe in some instances in the literature it's actually been given more than 10 times. Generally this has been separated by at least 6 weeks and often much longer periods.

Some information about the agent under study today from Stanford University. 3 patients with unfavorable characteristics received a second dose of Y2B8 and some received a total dose of 70 millicuries with a maximum of 40 at one administration.

Our own experience at the University of Alabama. We have a few patients with various radionuclides that have received retreatment, and in general with all of these things, what we could say is that there is a trend of somewhat longer recovery, mildly increased toxicity with the retreatment, especially if it's done at short interval, but for the most part, these have been tolerated.

To summarize some of these points I've tried to make today, more radionuclide data is needed to be able to improve dose and toxicity relationships. This may come about by improved data collection, processing methods, that will increase the accuracy and perhaps the standardization between institutions.

There are a number of modifiers that need to be taken into account in terms of the biologic effect as opposed to just the radiation number calculated. These include chemotherapy, as well as other radiosensitizing agents, cytokines, other agents, such factors as prior therapies, the disease status. All of these things can affect toxicity or tumor response, but they do not change the dose estimates.

Thank you.

DR. NERENSTONE: Thank you very much, Dr. Meredith. Does the committee have questions for Dr. Meredith before the FDA?

(No response.)

DR. TEMPLETON-SOMERS: Is Louise Peltier or anybody else from Guilford Pharmaceuticals here? I'd like to talk to you please. I'll be outside in the hall.

DR. NERENSTONE: The next part of our morning is the FDA presentation. Dr. Bishop.

MR. OHYE: Excuse me.

DR. NERENSTONE: I'm sorry. Mr. Ohye.

MR. OHYE: I don't have a question for Dr. Meredith, but I have a small request on behalf of the industry since I'm their industry rep.

If you've had time to assemble a formal bibliography, I would ask you to submit the bibliography to the Executive Secretary so that I can obtain same for distribution to interested parties in the industry.

DR. NERENSTONE: You mean a bibliography from Dr. Meredith's talk.

MR. OHYE: Yes, Dr. Meredith's.

DR. BISHOP: Dr. Nerenstone, members of the committee, once again good morning. I'm sure many of you find it very difficult to focus on these presentations in view of today's news. I will try to keep my remarks brief and to the point.

Over the next 30 minutes, I will focus on the most relevant efficacy and safety study results from the studies that were presented in the biologic license application for Zevalin. In part, this is because our time is limited but also because we do agree with the analyses that were performed on the primary and secondary efficacy endpoints that were presented to you by IDEC this morning. So, I will not try to duplicate this morning's presentation but again only focus on those relevant study results that will be salient to this morning's discussion and the questions to the committee.

First, the regulatory history. The results of five clinical studies were submitted to the agency in support of the proposed biologic license application. Very briefly, to remind everybody what the proposed indication is, Zevalin is being proposed for the treatment of patients with relapsed or refractory low-grade, follicular, or CD20 transformed B-cell non-Hodgkin's lymphomas and for the treatment of patients with Rituxan-refractory follicular non-Hodgkin's lymphomas.

The IND was submitted in 1992 and presented here on this slide are the dates that the five clinical studies were launched. Please note that the 106-04 study and 106-06 study, which are the two phase III pivotal studies, were initiated in 1998. The 106-98 study, which is the open access trial, the study that is currently ongoing, was initiated in December of 1999.

Fast track designation was granted in June of 2000. The BLA was received by the agency in November of 2000. Our first action was in May of 2001, at which point we issued a complete review letter to the company.

Two months later, the company responded, and this triggered the class 2 response initiating another 6-month clock, and the next action date is January 8, 2002.

First, the efficacy results. The Zevalin BLA contains two major studies: one major controlled efficacy study and one supportive trial in the refractory setting.

First the efficacy study, study 106-04. This study was a randomized study having an active control, rituximab. Subjects enrolled in the study were stratified by histology, IWF A's, the folliculars, and then the transformed. The primary efficacy endpoint was superior overall response rate as defined in the protocol and as evaluated by an independent group, the LEXCOR group. This LEXCOR group was blinded to the study assignments.

The overall response rate for the study 106-04, which is the primary efficacy endpoint, was achieved in this trial. Zevalin had a response rate of 73 percent; rituximab, 47 percent; with a p value, a Cochran-Mantel-Haenszel test, stratified by histology, p value of .002.

Looking at subgroup analyses for overall response rate in this study, what we have learned is that few subjects with IWF A histology or transformed histology were enrolled in either the Zevalin arm with 9 subjects each for these categories and 8 subjects for the rituximab with IWF A's, 4 subjects with the transformed in the rituximab arm for the transformed.

Represented here are the number of individuals that had overall response rates with the corresponding percentage. What we learned is that follicular subjects, which are the majority of the subjects that were enrolled in this trial, indeed had a response rate of 76 percent, 42 responders, in the Zevalin arm, as compared with 47 percent, or 27 individuals, in the rituximab arm.

In the IWF A group, 6 of 9 subjects were responders, representing 67 percent of the individuals, and this was compared to 3 out of 8 individuals in the IWF A categories in the rituximab arm.

For the transformed, only 5 out of 9, or 56 percent, were responders in the Zevalin group versus 3 out of 4, or 75 percent, in the rituximab group.

The median duration of response for all responders, 53 subjects in the Zevalin versus 33 in the rituximab, was 14.2 months for the Zevalin-treated subjects versus 12.1 months in the rituximab-treated subjects.

Within this group, 25 individuals, or 47 percent, were censored. 23 of these individuals are ongoing responders. 1 individual has been lost to follow-up, and 1 individual has expired. Similarly, 42 percent of the subjects in the rituximab arm are censored, and some of these individuals being ongoing responders.

If we look at duration of response and break this down by subgroup analysis, what we find is for the IWF A's the median duration of response was 9.8 months for the Zevalin-treated subjects. Because 67 percent of the individuals are ongoing responders in the rituximab, a median is not provided.

For the follicular subjects, the median duration of response was 18.5 versus 12.1.

And for the transformed individuals, the mediation duration in the limited number of subjects, 5, was 6.8 months and in the rituximab, 11.7 months.

Now shifting to the supportive trial, the trial 106-06, which was a nonrandomized trial in the rituximab-refractory follicular, B-cell non-Hodgkin's lymphoma. There the primary efficacy endpoint was overall response rate, again as evaluated by the LEXCOR group. The LEXCOR group was again blinded to the investigator's assessment of response.

In this patient population, a prospectively agreed upon overall response rate target of 35 percent and a duration of response comparable to prior rituximab would have been considered acceptable evidence of activity.

The primary efficacy analysis revealed an overall response rate, again protocol-defined response criteria, by the LEXCOR evaluation group, for the entire study population of 59 percent.

Now, 2 individuals within the study population did not meet protocol definition of follicular histology, and I have presented here the results for those 52 individuals that did meet the protocol-defined follicular group. In these individuals, the response rate was 58 percent. Because those 2 individuals really do not affect subsequent analysis, I am going to present to you results that include all 54 subjects.

The duration of response for the Zevalin-treated individuals, 32 responders, was 7.7 months as compared to their prior rituximab therapy. So, this is the median for the entire group. Looking at the median response for all 17 individuals who had a documented, although short, time to disease progression, the median was 4.0 months. This compares to the median for Zevalin of 7.7 months. This was the protocol-defined analysis that would have compared the duration of response of Zevalin to the rituximab.

During the review cycle of the material that was submitted in support of the license application, the FDA asked IDEC to perform an additional analysis looking at duration of response for the Zevalin therapy compared to the prior rituximab therapy, using each individual as their own control. Again, as was presented this morning, therapy was considered to be favorable towards Zevalin if the duration of response to Zevalin was at least, in our analysis, 1 month. The morning's data was at least 3 months longer. Again, the same thing with the rituximab. The alternative would be true, that if duration of response to rituximab was at least 1 month longer than Zevalin, the therapy would have been considered to favor the rituximab.

Looking at this analysis, what we find is 54 percent of the subjects would have been considered to have a duration of response that would favor the Zevalin therapy as contrasted to the 9 percent of the individuals whose duration of response would have been considered to favor the rituximab therapy.

In general, Zevalin therapy has, we believe, demonstrable and durable antitumor activity in the follicular subjects. However, there is limited data in the IWF A's and in the transformed subjects that would preclude, in our opinion, definitive conclusions. And we seek the committee's advice in terms of how they see those two subgroups fitting in in terms of Zevalin therapy.

Now, turning to safety, the dominant safety concern with Zevalin therapy probably relates to the observed rate of cytopenias. So, let me first begin by reviewing for you the hematologic toxicities.

Represented here are the grade 3/4 neutropenias and the grade 4 neutropenias. Grade 3 neutropenia is an ANC of equal to or below 1,000. Grade 4 is an ANC of 500 or below.

What we have done is looked at the first 90 days following initiation of therapy. There 214 subjects out of the entire integrated safety analysis, or 392 subjects, representing 55 percent of the individuals, had a grade 3/4 neutropenia. The median duration of this grade 3/4 neutropenia was 25 days. Please note that the range is wide. Approximately 3 percent of the individuals included in this group did not have documented neutrophilic recovery. Now, some of those individuals had confounding factors such as going on to additional therapies.

Similarly for thrombocytopenia or platelets, 57 percent of the study population had grade 3/4 thrombocytopenia and the median duration of the 3/4 thrombocytopenia was 27 days. Again, let me emphasize that the range is wide and the plus sign here represents that some individuals, approximately 9 percent of the study population, did not have documented platelet recovery or back to baseline. Again, some of these individuals had confounding factors such as going on to additional therapy.

Graphically the data is represented in this slide. Time 0 is the time at which Zevalin therapy was administered. What we have seen here is a predictable decline in neutrophilic count at approximately 30 days from the onset of therapy. This decline was durable for approximately 3 to 4 weeks prior to seeing a recovery. Some individuals have protracted recoveries and some individuals, as I have mentioned, did not have documented recovery, although this represented a small percentage of the population.

Putting this into context with additional effects of Zevalin therapy on the immune system, all subjects had B-cell depletion, and as was presented this morning, the median time to baseline recovery was approximately 6 months. There was a transient IgM decline which also went back to baseline within 6 months, and IgG and IgA remained normal. The reason I present this slide is, number one, to emphasize that this is similar to the profile that we see with rituximab alone and also to put it into context of the neutropenias that we see and then the incidence of infections.

114 out of the 358 data set, looking at infection, or 32 percent of the population, had a total of 183 events. 8 percent of these individuals had grade 3 or 4 events. Represented graphically on this slide is the percentage of subjects that had at least one of those events. So, looking at bacterial for all subjects, 54 percent of individuals had at least one bacterial infection, 15 percent viral, 11 fungal. In 67, the infection was not otherwise specified.

Looking at the breakdown according to NCI CTC grade for severity, the majority of the infections were grade 1/2's with a minority of the subjects having grade 3 and 4, again 8 percent of those subjects having grade 3 and 4.

Now, shifting to thrombocytopenia, 224 individuals, or 57 percent of the population, had documented grade 3/4 thrombocytopenia, again a predictable course where we see a rapid decline at approximately day 30 and a sustained thrombocytopenia for approximately 3 to 4 weeks prior to seeing recovery. There is, again, a number of individuals that have protracted thrombocytopenias and a small percentage, approximately 9 percent, again having a documented recovery back to baseline.

Looking at the study 106-05, which was introduced this morning in Dr. White's presentation, which included individuals with a baseline platelet level that was below 150 but above 100, what we learned is that when individuals have already a low baseline platelet at the time of receiving Zevalin therapy, a higher number of these individuals, or 87 percent of them, will incur grade 3/4 thrombocytopenia, again with the same predictable course where we see by day 30 a decline and then approximately a 3 to 4 weeks' duration of thrombocytopenia prior to recovery.

I have scaled this axis up to 210 days to emphasize that a number of these individuals can have protracted thrombocytopenia, this representing approximately 12 percent of the population. As was pointed out this morning, Dr. White indicated that a number of dots here would be missing because, as per protocol, once somebody had recovered from their hematologic toxicity, it was no longer required to continue to monitoring these individuals.

Putting this into the context of incidence of bleeding, 18 percent of the individuals enrolled had at least one bleeding event. 7 of these subjects had a total of 12 grade 3/4 events. Let me cover those 12 events for you.

2 individuals had intracranial bleed that resulted in death. One of those individuals with intracranial bleed also had a vaginal bleed and ecchymosis that was at least grade 3. 5 subjects had gastrointestinal bleeding. Of these, 4 of them were documented as GI bleed, 1 of them hematemesis, and 3 of them melenas. One of those subjects, 1 of the 5, had a GI bleed, hematemesis, and melena occurring during the same episode.

IDEC has performed some exploratory analyses looking at cytopenias and potential risk factors predictive of cytopenias. Represented here are some of the results of those exploratory analyses. It appears that baseline bone marrow involvement, the number of prior regimens, especially when fludarabine was used, and baseline platelet level would all be predictive of grade 3/4 hematologic toxicities.

Now, shifting over to the non-cytopenic adverse events, represented here in this table are the most common adverse events that were documented in subjects enrolled in all of the studies submitted in support of the BLA. Asthenia was the predominant adverse event, followed by nausea, infections, chills, fever, abdominal pain, dyspnea, headache, increased cough, and pain. There were other nonhematologic adverse events that are not listed here, but all of them were below 15 percent.

Please note that the incidence of grade 3/4 nonhematologic adverse events was, for the most part, low in the study population. Probably the highest number was with infections, representing approximately 8 percent of all subjects. There were 3 percent of individuals who had grade 3 asthenia, fever, and abdominal pain. 2 percent of the individuals had dyspnea. All the other nonhematologic adverse events that are not listed here either had no events that were grade 3/4's or less than 1 percent of the subjects had events that were grade 3/4's.

Looking at the comparative study, the 106-04 study comparing Zevalin therapy to the rituximab therapy, portrayed here in the bar graph are the most common nonhematologic adverse events for these two arms. Notable are asthenia and nausea, long with infections, pain and abdominal pain, where were more commonly seen in the Zevalin therapy. Other common adverse events portrayed here are chills, fever, and headaches.

What I have done in this bar graph is highlight for you some of the notable adverse events in terms of having a numeric difference between the two study arms. Increased cough, dizziness, dyspnea, peripheral edema, arthralgia, anorexia, anxiety, and ecchymosis were more common in the Zevalin-treated subjects. Please note that the majority of these events are grade 1/2 adverse events.

Pruritus and angioedema again were more common in the rituximab-treated subjects.

Secondary malignancies were observed in the Zevalin-treated individuals. 3 acute myelogenous leukemias, 2 myelodysplastic syndromes were noted in the entire study set. One individual was also documented as having a meningioma. The onset of the secondary malignancies was 8 to 34 months following Zevalin therapy and approximately 4 to 13 years following the lymphoma diagnosis.

Prospectively the time points for the HAMA and the HACA's sampling were probably inadequate to assess the true incidence of the HAMA and HACA. The reason being is at the time that these trials were designed, I think the agency had not anticipated that HAMA and HACA formation could actually appear 6 months post therapy. We have reasons to believe that HAMA and HACA formation could be documented up to a year following initiation of such therapy. Suffice it to say that currently IDEC has incorporated longer time points in the ongoing studies, and because of these ongoing studies, I think that we will have to wait to really find out what the true incidence of HAMA and HACA response is for Zevalin-treated subjects.

But in the integrated safety analysis for which we have data on 211 subjects, there were 5 individuals who had positive HAMA titers at any given point during the therapeutic course. 2 of them had positive baseline HAMA titers. 3 of them developed titers post treatment. 3 subjects also had positive HACA titers at any given point in the Zevalin treatment. 2 of these subjects had positive baseline HACA titers, and 1 individual developed HACA titers post therapy.

Looking at the adverse event profile from these individuals, they are not at all outstanding compared to the rest of the population.

70 of 349 individuals, representing 20 percent of the population, have died. 58 of them were due to progressive disease and 12 of them due to other causes. We have already talked about the two intracranial hemorrhages which we believe was related to the documented thrombocytopenia in these subjects. There were 5 myelodysplastic/AML subjects who have also subsequently died. 3 individuals died of pulmonary complications in the context of preexisting pulmonary disease such as COPD. 1 individual has died of coronary artery disease and had a cardiac arrest, and 1 individual had a pneumonia subsequent to salvage therapy following Zevalin therapy.

Overall Zevalin therapy can be characterized by a high incidence of cytopenias, and as mentioned, 55 percent of the individuals had grade 3/4 neutropenia and 57 percent of individuals had grade 3/4 thrombocytopenia. And the median duration of these neutropenias was approximately 3 to 4 weeks.

To review for you, the most serious adverse events included the hemorrhages, including the 2 individuals who have died, the myeloid malignancies with the 5 individuals who have died. There was a percent of individuals who had grade 3/4 infections, and although most of the allergic reactions were grade 1/2, I think there is sufficient concern from the agency that we would categorize them as troubling adverse events.

Briefly I will now shift to dosimetry and biodistribution. The agency has received data on 179 subjects that were assessed for biodistribution imaging. There were five imaging time points obtained for these individuals. In summary, there was sufficient diagnostic quality imaging provided to assess the dosimetry for multiple organs, as well as imaging for known tumor sites.

The MIRDOSE 3.1 software was utilized to analyze these studies. Regions of interest for multiple organs with localization of radiolabeled antibodies, such as the heart, lung, liver, small intestines, spleen, and testes, as well kidney and bone marrow looking at the sacrum area, was performed.

Represented in this table is a subset of these analyses. First, the spleen, marrow, and liver, which I think are traditional organs commonly looked at in terms of maximum dose and potentially reflecting target injury to these organs. Represented here is the median dose for 32 millicuries to these organs. So, 1,350 for the spleen, 90 to the sacral region of interest for red marrow, and 547 centigrays for the liver.

The next three categories are categories that were of interest in the post-submission analysis performed by the FDA. During this review, it was uncovered that the testes were receiving a median dose of 950 centigray. As you've heard today, I think we would agree that it is possible that this number overestimates the measurements in the testes because of the limitation of the software. However, even considering the limitation of the software, we believe that substantive doses are likely within the testes, and although we cannot see the ovaries, they're possibly also involving the female gonads.

Represented here are median centigray dosages to the upper large bowel and lower large bowel. The reason that these are presented to you is because in almost all of the images that were reviewed at the FDA, we do see lymph node aggregates within the bowel imaging with the indium-labeled 2B8.

Correlated with the imaging of the GI tract is the numeric appearance that there's a greater number of GI toxicities within the Zevalin-treated therapy represented in the blue graph compared to the rituximab-treated individuals in the red bar graph. So, nausea, abdominal pain, and vomiting were more common in the Zevalin-treated individuals. Please note that the majority of these GI toxicities were grade 1 and 2.

The FDA has also performed worst case scenarios using the existing data that was submitted in the BLA. In the modeling that we have performed, we have estimated that it is possible that adjacent normal tissue could have as high as 8,000 centigrays up to 1.1 millimeters into an adjacent structure so the potential dose that an adjacent tissue could receive could be as high as 8,000 centigrays up to 1.1 millimeters into that structure.

We have also modeled the data looking at alterations in the biodistribution and also obstruction of the clearance route for the Zevalin therapy looking primarily at potential outlet obstructions and delivering sufficiently high dosages to the kidneys where injury could be caused.

So, overall assessment pertaining to the dosimetry and biodistribution, I think we would agree that normal organ dosimetry supports the use of a fixed dose of yttrium-labeled Zevalin. In addition, the biodistribution we believe is necessary to assess normal organ and tumor site localization.

Another comment that I would like to make is that currently we feel that there is inadequate data to assess the safety of additive localized radiation effects from external beam radiation therapy and Zevalin therapy. And I think Dr. Meredith alluded to that in her talk, that that additive effect could be serious and could result in significant morbidity.

So, briefly let me conclude that we believe that there is sufficient data to demonstrate durable antitumor activity, as was documented with overall response rate in both the efficacy studies, the 106-04 trial and the 106-06 trial. We believe that Zevalin therapy is associated with a significant hematologic toxicity in the majority of the subjects, and this can result in serious morbidity in the minority of subjects.

As compared to the Rituxan therapy, Zevalin was associated with a superior overall response rate. There was similar duration of response and time to disease progression. Zevalin also showed a 58 percent overall response rate in the rituximab-refractory individuals.

Data is, however, limited in the non-follicular subgroups and data in these subgroups is also limited for subjects who have not received prior Rituxan.

Thank you.

DR. NERENSTONE: Thank you very much.

I'm going go open it up now, questions from the committee to FDA. Yes, Dr. Taylor.

DR. TAYLOR: Could you clarify again? On the second malignancies, you saw five and they were all in the Zevalin, none in the control group?

DR. BISHOP: That is correct. The 5 individuals that I presented in my slide are from the integrated safety analysis in the Zevalin-treated subjects. I am not aware of rituximab-treated individuals.

Dr. White, do you have a comment to that effect?

DR. WHITE: There was a single patient on the rituximab control arm who developed a pancreatic cancer.

DR. BISHOP: That's correct, but none of the myeloid malignancies. I'm not aware. Again, that study arm was only with 70 individuals as compared to the entire integrated safety analysis of 358 individuals in the Zevalin.

DR. NERENSTONE: Just information for everyone who's wondering, to bring us back to the real world for just a moment. I'm getting things brought up to me, and as they come up to me, I will pass them on. One is that Camp David has been hit, and another is that there is a biological warfare threat. So, those are both confirmed I guess on CNN.

I know this is hard to keep to us concentrated, but since I know, I think it is fair for people to know.

And this afternoon's discussion will be canceled and postponed.

Dr. Levine.

DR. LEVINE: A question just going back to the myelodysplasia. What was the follow-up from the time of diagnosis or the number of treatments given in the Rituxan group versus the Zevalin group? I'm trying to figure out the myelodysplasia and is that due to Zevalin or is that due to all the other treatments and all the other times? So, how long was the follow-up and how many treatments given in the Zevalin group versus the Rituxan when you have, whatever it is, 5 versus 1, as far as the AML or myelodysplasia follow-up?

DR. NERENSTONE: Who wants to answer that?

DR. BISHOP: In the control study, the median number of therapies for both groups was, I do believe, two prior therapies in the Zevalin-treated and the rituximab. This number increases, I do believe, for the overall analysis, and I think it's three prior regimens for the overall integrated safety analysis.

I am not aware -- and maybe Dr. White would like to supplement this. Looking just at the 73 individuals in the Zevalin-treated arm and the 70 rituximab, I don't believe that any of these Zevalin-treated individuals were the individuals with the myeloid malignancies. Is that correct, Dr. White?

DR. WHITE: If you restrict the analysis only to the phase III randomized trial, there was one pancreatic cancer on the rituximab arm and one myelodysplasia on the Zevalin arm. The median number of prior therapies on the various trials were either two or four prior therapies, with a range up to nine. And the median observation time for the safety population after Zevalin was about 2 years. That's median.

DR. NERENSTONE: Did that answer your question?

DR. LEVINE: No. It still doesn't answer it. In other words, I just want to know whether the Zevalin-treated patients have been followed longer than the Rituxan followed patients. That's my question. How long has each group been followed to see what's going to happen?

DR. WHITE: I understand now. The Zevalin-treated patients, if you count the entire group of patients that we submitted data on, 489, have been treated from 1993 to the present. However, the majority of those patients were treated more recently with a median follow-up of about 2 years.

The rituximab-treated patients that were in the 106-04 trial were only treated on that particular trial which began in 1998, and I would have to estimate that the median for that group of patients would be less, but we would have to calculate it to know for sure.

DR. NERENSTONE: Dr. Sausville.

DR. SAUSVILLE: I have two sets of questions. The first does relate a little bit to the toxicity issues. A striking feature is the duration and the risk for morbidity that is of potential concern in the broad application of this type of therapy. So, I guess I would ask you, I guess primarily the FDA, in your analysis of the data provided with the dosimetry, was there any evidence that the people who had the more severe myelosuppression by one index or other had either a different feature to their dosimetry, a smaller tumor mass, any evidence of different behavior of the product?

That leads to the second set of questions. Are we clear that the robustness of the product elaboration and what is admittedly a somewhat more complicated procedure than we usually undergo is sufficiently reproducible that some wobble in that process might not be related to these toxicities?

DR. BISHOP: I will let Dr. Mills address the dosimetry issues.

DR. MILLS: From the standpoint, we saw no characterization amongst these subgroups that you're describing. The dosimetry unto itself, we were looking across in the representative organs, and we were not able to discern that there was any difference in these response curves that you're talking about. Our concern frankly from that standpoint was to look at the normal organ dosimetry, and we did not break out any differences to them. We did not look across board. And I'll ask Dr. Bishop if he could comment in terms of the response characteristics you're talking about for the toxicity.

DR. SAUSVILLE: Specifically did the people who had more toxicity have less tumor volume, a lower mass?

DR. BISHOP: Again, I think it's important to note that the majority of the toxicities that we had seen were grades 1 and 2, with probably the most significant toxicity -- the grade 3/4 toxicities observed being infection in 8 percent of the study population. Clearly we believe that the incidence of infection was higher in the comparative arm for the Zevalin-treated individuals than for the rituximab-treated individuals.

Now, having said this, the other nonhematologic adverse events that we were seeing tended to be, when looking at grades 3 and 4, of low frequency.

Now, pertaining to your question, addressing whether or not tumor bulk may have been related to a pattern of toxicity, we are unaware that there was any predictable patterns that were observed in the data set that would relate tumor bulk with observed adverse events, whether they were hematologic adverse events or nonhematologic adverse events.

DR. NERENSTONE: Dr. Blayney?

DR. KEEGAN: Could we answer, just for a moment, Dr. Sausville's other question?

I believe your question pertained to whether or not in the manufacture of the product, there were concerns about the robustness of the manufacture. And Dr. Shapiro can elaborate if you want, but she has said that to our satisfaction, the product can be reproducibly manufactured within their own specifications such that we did not observe great variability in the trial that would suggest that there was some looseness there or that makes us more concerned about the use of this in the community.

DR. MILLS: Maybe I want to express one extension from that. The difficulty that you're going to see in terms of this product is you've had a large multi-center experience, but you've not gotten it out in the community hospital. Part of the concern that you'll have, in terms of reflectiveness, is one of the elements within the biodistribution imaging that would represent for you another safety element, that indeed the ability to prepare the indium-labeled 2B8 product and actually observe the expected normal biodistribution is a safety element that indeed in the community hospital setting they'll be able to verify. And if that fails, that's evidence they should not proceed. One of the concerns that you may have is that you'll be able to identify that on an individual basis.

DR. SAUSVILLE: So, again, I guess in the proposed labeling for this product, are there criteria that would tease that out? And do we have sufficient quality control out in the community that this type of thing --

DR. MILLS: In the presence of an altered biodistribution if a product fails in terms of breakdown, it's readily apparent to a nuclear medicine imaging physician where you see a significant loss of the blood pool evaluation. So, as a result, this is a fairly course and relatively straightforward element. You have basic standard production elements also, but there's another fail safe in terms of looking for that expected biodistribution and vascular component that would fail if indeed it was in a colloid form which would break down.

DR. SAUSVILLE: Again, just to pursue the point just briefly, where you're thinking leads, at least me, is that I have no doubt that a nuclear medicine physician in a research facility that's used to doing clinical trials with imaging of this sort would be very comfortable with the call that you're making. On the other hand, I think it would be incumbent upon somebody to make sure that that level of expertise is --

DR. MILLS: I think that information can be readily transmitted to any facility which would be using this type of product, even in the community setting. But from our standpoint, realize that you have such products as Oncoscint and Prostascint which are already out in the community. So, there is a fair understanding, in terms of the expected biodistribution of a radiolabeled antibody product on a diagnostic basis, and that's what you have with the indium.

DR. SAUSVILLE: I guess that's the key point. It's a diagnostic rather than therapeutic basis.

DR. MILLS: That's right.

DR. NERENSTONE: Dr. Blayney.

DR. BLAYNEY: First of all, I'd like to say that your review was well organized and it was a pleasure to read. Thank you.

Second, you alluded to an adjacent organ toxicity of 8,000 rads or equivalent of 8,000 centigray. What organs did you simulate in your analysis?

DR. MILLS: In the analysis, we took several different models, pulmonary artery, simulated a small nerve such as the vagus nerve, also simulated the pericardium. Those would be elements where we would be concerned. The pericardium would be a structure which, indeed, if you put one of the example tumors, which we had one almost as high as 25,000 centigray, adjacent to the pericardium in a theoretical model, we could put as much as 8,000 to 10,000 rads into that adjacent pericardial structure. So, there is a potential theoretical risk.

We did not see it in the clinical trials, but one should understand there is a potential risk to put that much radiation a millimeter away from the margin of the tumor. So, it's a concern for the morbidity to identify that it is a potential risk and especially in these early evaluations.

DR. BLAYNEY: The second question is, do you anticipate any inclusion in the label about patients who have received previous external beam or involved field radiation therapy as a warning or precaution to physicians who might be administering this product?

DR. MILLS: I would anticipate that certainly we're going to have to have advice because we know that we have limited data and that data is not adequate for us to really come down to fair conclusions from it. We've had a small, but good experience with this product to this point, but we need to have extensively more follow-up and more evaluations in the community setting as well to be able reflect what's going to happen in these patient populations. So, obviously a warning, an indication, information for the attending physician to realize that this lack of data is there and to be aware of this concern.

DR. BLAYNEY: I don't know if it's appropriate to your simulations. Even though such a high dose hasn't been observed, I notice the dose range of adjacent organs is quite wide in the sponsor's material. Is it appropriate to warn physicians that may be treating large or even small mediastinal tumors adjacent to the pericardium, vagus nerve, and some of these other vital structures you've talked about?

DR. MILLS: It's quite apparent from doing these simulations, that the more this information we draw into the package insert, I think the better we're going to be. You're going to go from a small experience to a larger experience in the community, and for them to have adequate information and to realize what the theoretical model is is a very significant concern for us in terms of making sure we get adequate information out there. We don't have a lot of data yet in terms of it. As you've heard this morning, the dosimetry community is itself evolving as we speak in terms of these concerns and understanding how we're going to assess them.

DR. BLAYNEY: Maybe that community will read the package insert more vigorously than the oncologists.


DR. MILLS: I am certain they're going to have full attention to it. At a recent meeting, they had quite a bit of attention already to it.

DR. BISHOP: If I may make one additional comment to that effect. I think having the indium 2B8 step incorporated into the overall therapeutic administration also provides the opportunity for a nuclear medicine physician, should he or she decide to do so, to perform some dosimetric analyses, especially if there was an adjacent area of particular concern. I think that having that in there, although we do not envision this being a requirement at this point, we also understand that there is that option should one decide to do so.

DR. NERENSTONE: I have a quick question or not so quick. It was a little disappointing to see that only one randomized trial is being submitted for full approval and that the basis of that one randomized study is really response rate. As we talked about, clinical benefit has really not been shown in a statistically validated way other than testimonials from the investigators, and we certainly see these patients and understand that.

The second trial that was submitted as a phase III, as you're quite aware, is really a phase II study in Rituxan-refractory patients.

I understand your attempt to get at patient improvement by your post hoc analysis that you suggested of the duration of response with the patients acting as their own control. I'm a little bit concerned about that because that was an unblinded evaluation of duration of response. Would you discuss that a little bit as to how robust you think that indication is?

Certainly it makes us reassured that it's not worse than Rituxan. Do you really think we could say that it's better than Rituxan given all those concerns?

DR. BISHOP: Certainly some of the questions that we have before you today are seeking such advice from the committee. So, I'm not sure in terms of the clinical review team having really formulated an opinion on the latter portion of your question.

Let me try to address, in terms of the initial drug development program and some of the agreements that may have been reached over time with the agency.

We recognize the limitations of the studies that are before us, and we certainly share your concern about the number of subjects that have been studied in pivotal trials, especially when we understand that low-grade non-Hodgkin's lymphoma has a high prevalence when compared to other lymphomas.

The study population that we're looking at is, indeed, a relapsed or refractory patient population who have had multiple prior regimens, as high as nine prior regimens. Although the median in the Zevalin-treated patients in the randomized trial was two prior regimens, I think the range was up to five.

These individuals tend to have fewer and fewer options as they continue to relapse with their illness. I think that we recognize that there's also a shorter and shorter overall response rate with sequential therapies.

This probably influenced the basis for our acceptance of at least looking at activity in terms of overall response rate. We were never satisfied that this would be sufficient by itself. I think it's important to try to extrapolate from the data some sense of clinical significance, and that I think is important for the committee to consider whether or not there is such evidence in some of the exploratory analyses that have been done in the two studies.

And we do agree that the second trial, which is labeled as a phase III, is really a supportive phase II study, and we understand the limitation of that trial as well.

DR. NERENSTONE: In this setting, just to ask on behalf of the committee, if the decision is that there is not enough information to support full approval, would accelerated approval with subsequent phase IV commitment be an option for this committee to recommend, or is it an up and down on full approval?

DR. SIEGEL: The company has to request accelerated approval before we can grant it. Sometimes we ask if they're interested when we get such advice and they can come back and request it.

Accelerated approval would be approval based on a surrogate with a plan for confirmation.

Interestingly, in my experience with cancer drugs, it's not uncommon that an initial approval may be based significantly on response data in a treatment-refractory population and that the commitment for confirmation is a head-to-head trial perhaps in an earlier stage of disease.

Here we have a head-to-head trial in an earlier stage of disease. What you're correctly pointing out is lacking is a clear-cut clinical outcome as opposed to a response rate outcome finding. So, if the committee were to deem that was useful advice, it would be helpful for us also to hear discussion about what sorts of data the committee would find useful in confirming the clinical benefit and what sorts of trials the committee thinks would be appropriate to conduct in the future.

DR. NERENSTONE: Are there other questions from the committee to FDA? Dr. Lippman.

DR. LIPPMAN: To clarify Stacy's comment, again in the early discussions between the FDA and the company, were the discussions of one trial versus two? Was that something that was worked out up front and deemed acceptable to do one randomized trial?

DR. KEEGAN: The discussions focused around a single randomized trial with good supportive data and durable overall response rates in the refractory population. That was the agreement with the agency. We did not require two randomized controlled trials.

In this disease, we didn't feel it was an appropriate standard to hold them to a survival endpoint in that it was a late stage of the disease in terms of where they were in treatment and the fact that there is no data that there is any therapy that actually confers a survival advantage. So, we felt that was too high a mark.

I will say that there was some thought that the time to progression data should have or would have confirmed -- there would have been an advantage confirmed on time to progression if there was, in fact, an overall response rate that was higher with a similar duration of response. And we were somewhat expecting to see that in this study as sort of internal consistency. What we found was that, in fact, in both arms the patients who had stable disease as their best response, that category of nonresponders, in fact, had a very prolonged of stable disease. So, we really have not observed a time to progression advantage, which is something we were expecting in this trial and we thought would have been a strong confirmatory secondary endpoint.

DR. NERENSTONE: But we were also told that the study was not powered to look at that.

DR. KEEGAN: It was powered to look at substantial differences in time to progression but not to specifically exclude a smaller difference. Obviously, every study is powered, to some extent, to see differences if they're in fact very large, but we did not require them to identify a very specific difference but only to give us a qualitative evaluation.

DR. NERENSTONE: Dr. Sausville.

DR. SAUSVILLE: I think the other complication here, which is I think unusual, is that you're not really comparing two different therapies. In the Zevalin arm, you basically get one way of looking at it, almost half of the treatment that's in the other arm. So, again this leads us into somewhat interesting waters in terms of whether or not the usual rules for comparison or expectations for comparison are reasonable I think.

DR. SIEGEL: That concern is reflected, you might note, in some of the wording of the questions. We have a regulation regarding fixed combination therapies, which arguably this falls into, and from a scientific point of view it has a lot of similarities, which would suggest that you need to demonstrate a contribution of each drug in the combination.

Here there's, I would venture to say, although I haven't independently researched it, very little if any data about what this dose of rituximab, this 250 twice, would do if done alone. I think the most we could say with comfort is that it's probably not any more effective than the approved higher dose of rituximab.

So, I think one of the things we're interested in determining is whether one can say from these data -- and I think this gets at the heart of what you were saying -- that this treatment offers benefits beyond that which would be seen by rituximab alone because from that, we can then deduce that the radiolabeled component is contributing to the therapeutic effect.

DR. SAUSVILLE: But by that way of thinking, it becomes in a sense a toxicity tradeoff. In other words, one of the features that's attractive about rituximab as a single agent is that it's very, very safe in terms of the usual things you worry about. On the other hand, while it is true that this could be considered less safe by some criteria, there is the evidence that was presented that there might be at least subsets of patients or individual patients who obtain a prolonged response.

DR. SIEGEL: Right. From the point of view of that particular regulation, I would say what needs to be established is that it has a benefit beyond that of rituximab alone, not that that benefit is outweighed by the risk. So, we don't need, in order to able to approve it, to say it's a superior therapy overall to rituximab. We need to be able to say that there's a contribution.

I think then, though, we also, as with any approval, will look at the risk/benefit and say do the benefits of this therapy outweigh its risks. So, maybe I'm agreeing with you but wording it differently in light of our regulation.

DR. SAUSVILLE: But that's where in a sense the second trial, which is the comparison against, as it were, prior response to either chemo or rituximab -- although it's a different way of looking at things, it clinically addresses a very common scenario in the care of these patients because to have a meaningful response rate in a rituximab-refractory setting is noteworthy.

DR. SIEGEL: Right, we think so. In a lot of drug development, looking at refractory early in development, one of the approaches is in fact an open label study in treatment-refractory patients. Analysis of those studies is often, as in the analysis you've seen presented, based on a presumption that simply repeating or adding yet a different related chemotherapy regimen would not yield a response better than the response to the prior cycle of chemotherapy, that based on a fairly broad, as I understand it, data set -- you all know better than I -- in a variety of diseases that recurring rounds of chemotherapy tend to give diminishing returns. So, if you see something that surpasses the preceding cycle, you can presume that that's better than you would have received by simply another drug or repeating the prior drug.

Now, there is certainly less data about repeating rituximab in this setting, but one might -- and this is sort of at the heart of the question -- be willing to presume from that study that a repeated round of rituximab would not give a better response than the prior round of rituximab did. If that presumption is correct, there are regression to the mean issues. There are a lot of issues here that you could question. But if that presumption is correct, then this comparison showing a significant response rate, as well as the one showing more durable responses, suggests a significant activity.

DR. SAUSVILLE: Right. But to be clear, there would be very little clinical enthusiasm, given the definition of Rituxan-refractory that was used here, in retreating patients with Rituxan.

DR. SIEGEL: And that's why it's hard to do a controlled study in that setting. You almost have to do this design because nobody wants to randomize to a treatment that hasn't worked in the past.

DR. NERENSTONE: Are there any questions from the committee to FDA?

(No response.)

+ DR. NERENSTONE: If not, then what I'd like to do is open the discussion of the committee. I don't know if Dr. Bridges, Dr. Sausville, or Dr. Levine would like to start out. They're our invited consultants. Comments? Dr. Levine?

DR. LEVINE: I'm not worried about what was last stated because the data was very careful in the real nonresponders to Rituxan. That was a 51 percent response rate on the Zevalin. So, I'm very comfortable thinking that the Zevalin will work when Rituxan doesn't.

One of my concerns relates to the broader indications, including the transformed cell. I have real difficulties on that one for several reasons.

Number one, it's a total of 9 patients, which you can't say anything. An example of the fact that if the numbers are too small, it doesn't mean anything, is the concept that they've got a 75 percent response rate to Rituxan alone in transformed cell lymphoma, and if that's valid, then I've been treating my patients wrong all these years. So, I think the numbers are just way too small.

In addition to the fact that the numbers are too small, their data show that in fact Rituxan was superior to the Zevalin, if you want to believe those small numbers. So, I think we need a lot more information on the transformed cell group.

On the IWF A group, I could make the same comment. The numbers are about the same, i.e., very, very small. Personally I don't have the same problem with that group, however. They have shown in those very small numbers that in fact the IWF A group treated with Zevalin did do better, did have a higher response rate than the same group treated with Rituxan alone. That's more reassuring to me. In a biologic sense, I'm comfortable as well that it's the same biology, as far as the disease. It would make sense to me that it might respond. So, I don't feel as strongly on the IWF A group as I do on the transformed group.

Moving around a little bit, I am concerned about the AML/MDS, not to the extent of believing that this product should not to go forward, but certainly with the belief that this must be very carefully looked at over the years to come.

That kind of colors my view as it relates to the pediatric indication as well. I'm worried about it. I don't think there's a major problem here because follicular lymphoma is so unusual in children. So, I don't think it would be a big hit to say, no, let's just leave the children out of this for a moment until we see the long-term effect of the radionuclide conjugated product, especially considering that the product would be used in people who had failed other regimens.

So, I'm uncomfortable about the indication for children. I'm clearly uncomfortable about the transformed. I could be swayed in either area I guess on the IWF A, and then wanting follow-up data on the MDS/AML.

There may be other comments and maybe I could come back and say those later when I think of them.

DR. NERENSTONE: Thank you.

Dr. Bridges.

DR. BRIDGES: The one issue of external beam toxicity with this treatment would be a concern to me that even though limited and there's no clinical evidence about these high doses to adjacent structures, that somehow that would have to be communicated to the radiation oncology community. There would have to be, I think, a vehicle there.

I think if there were possibly some modeling -- an additional -- in maybe several paraspinal tumors in the periaortic region adjacent to the spleen with your modeling capability to educate the radiation oncology community on the potential dose there.

Those would be the two factors that I would think could be looked at.

I don't know. There was no comment about spinal cord. It was omitted as a normal tissue, an organ that was looked at, as far as the dose to the spinal cord in this study. And I don't know if it was a limitation of the computer model or picking a site in the spinal cord to dose, but I think that would be something that needs to be put out to the radiation oncology community.

DR. NERENSTONE: Dr. Sausville.

DR. SAUSVILLE: I agree with the comments of my esteemed colleagues and I particularly agree with the characterization of the relatively small database to generate enthusiasm for the nonfollicular lymphoma subtypes.

I guess I also, on the other hand, think that there are certain elements to the optimal use of this product that can only emerge from further clinical trials.

It remains unclear to me, and I guess it really wasn't looked at in any of the material that we were presented. CD20 is assumed to be the same in its expression level in all comers. I think that's uncertain even within the world of follicular lymphoma. I think it will be interesting as experience emerges, if there are additional studies, of whether that can be honed in on with some greater precision.

I've already commented on the concerns that I have that if this is used in a general sense, that very clear ability to define the capacity of the receiving, more community-oriented oncology setting be supported in their optimal application of the product.

Beyond that, though, I think the sponsor is to be congratulated for tackling a difficult problem and clearly reaching some very interim fascinating and potentially valuable outcomes for some patients.

DR. NERENSTONE: I have a question that probably is going to have to go back to the FDA. We're not allowed to consider cost in our evaluation, but certainly if you look at difficulty in giving this medication, especially in the community, it's really not one dose. It's really Rituxan, then at least two follow-up imaging studies, then Rituxan and the therapy.

According to the sponsor, they had no problems after the imaging to change their decision to give the dose.

Is there going to be central monitoring of these initial images to know at what point they are no longer needed, or is this just going to be in perpetuity that we have to do these indium studies? I think the complexity it adds and the cost it adds really can be considerable.

DR. BISHOP: I'll lead off probably with a comment relating to one of your questions which pertains to what additional studies can be performed in terms of making us comfortable that the imaging portion of the Zevalin therapy is no longer necessary. I think that's one of the questions to the committee today. So, we're certainly looking for input from you in that regard.

It is not entirely accurate to say that there's no experience in which imaging studies did influence clinical decisions in the overall study experience. We are aware of at least three instances in which clinicians did make decisions, based on the images that they were seeing, not to continue with the yttrium administration. None of those instances turned out to be because of unsafe dosimetric evaluations when these images were subsequently reviewed in a centralized area. But there were some concerns with at least 2 of the patients that they may be unsafe dosages to certain organs. Again, this was very early in the overall experience with Zevalin.

And there was a third case in which there was at least a distribution of the indium-labeled antibody that was not concordant with the known distribution of lymphoma in which the physician subsequently decided not to go ahead with the therapeutic administration of Zevalin.

DR. NERENSTONE: Would the sponsor like to add anything?

DR. WHITE: Yes. Just a point of clarification on two items.

The first item, with regard to Dr. Bishop's description of the three cases where imaging was used or dosimetry was used in a decision to move forward or not to move forward, the first of those cases was in the 1993 106-01 trial where the dosimetry specifications were different, and it required that the ratio between the dose to the tumor and the dose to the highest organ was of a certain magnitude. In that patient, because of a liver dose of 900 relative to the tumor dose -- this was an intermediate grade patient -- the dose was not given. So, it's a little different situation.

In one patient, the imaging was looked at with regard to a SPECT scan because of underlying retroperitoneal imaging, and the decision was made to go ahead and treat.

And in the third patient, there was a decision by the investigator, because of a single functioning kidney, not to go ahead and treat, although it met dosimetry requirements.

Another point of clarification with regard to the CD20 distribution. If you would like, Dr. Horning is prepared to address this as well. But to our knowledge, there is greater than 95 percent, virtually all of the patients with low-grade nonfollicular lymphoma, do express CD20 and also the majority of patients with transformed lymphoma, although the intensity of expression in the low-grade nonfolliculars can be lower, particularly in those who have small lymphocytic lymphoma.

DR. SAUSVILLE: Right. So, you raised an interesting issue though. I actually agree with those numbers, but if you were one of the 5 percent, is that a problem? Number one.

And number two, there's the intensity issue. Being positive might mean 5,000 per cell or 500,000 per cell. The dose that's going to be given to the tumor is different.

DR. WHITE: Acknowledged.

One last point of clarification. In the rituximab-refractory trial, the comparison to the prior therapy was prospectively defined in the protocol prior to opening of the protocol. The additional different methodology was defined by FDA later on. But there was a prospectively defined comparison that was also performed and also had the same exact results.

DR. NERENSTONE: You're talking about the duration of response question?

DR. WHITE: Comparison of the overall response rate and duration of response to prior rituximab in prior therapy.


DR. KEEGAN: Dr. Nerenstone, to get back, we believe that it's possible that there may be a body of evidence that might be convincing in terms of telling us when one might omit that initial imaging step. Our concern is that the database is so very small that we don't have a very good estimate of how often, in fact, alterations in biodistribution might occur, although it certainly must be in a relatively small number in a carefully selected population, as we've seen in the study. Whether that represents the patient population at large that may be exposed to this drug is again another issue that we have.

So, what we're seeking from the committee is what are the types of patient populations we should focus on and what level of alteration of biodistribution might be considered acceptable for this type of therapy so that we could decide what the total numbers of patients and the types of patients are that should be studied to gather this data and reassess.

For example, at this moment, we couldn't say that we could exclude an incidence of 1 percent altered biodistribution even in the population that's screened in this studies. Is 1 percent of altered biodistribution acceptable? Or missing that, would that be acceptable or would that possibly not be acceptable.

So, we would need some discussion of what we should be focusing on. And we raised some of these issues like prior exposure to murine proteins and prior exposure to chimeric proteins possibly as well, Rituxan in particular, as well as whether there's some incidence of altered biodistribution which you would find unacceptable, so unacceptable that we should not consider removing the imaging.

DR. NERENSTONE: But there's nothing so far to think that you might.

DR. KEEGAN: I think the experience is too small at this point, but the question would be, at some point, would there be a robust enough experience that would give us a level of comfort that the incidence could be lower than 1 percent, say, and would that be sufficient in your minds to consider removal of the imaging step as a screening procedure.

DR. SAUSVILLE: I think this gets back to the point that's come up in different ways throughout the morning, that although the imaging was used and was certainly reassuring that everybody imaged the tumor, and there was acceptable biodistribution, we nonetheless really didn't see hard and fast criteria that were applied prospectively. I think that's something that I presume the agency would work out with the sponsor if this goes forward in a way that addresses some of the concerns related to distribution and one might even say lighting up of a mass in the first place. In the unexpected event there was reasonable distribution, but the mass didn't reach a ceratin level of brightness, as it were, one could imagine that as a criteria not to go forward as well.

DR. MILLS: I think also the committee should consider the concern that we've heard this morning. It's the potential that we may develop subsets in terms of the tumor distribution and where it's imaging and other potential organs at risk adjacent to it. You may see in the next several years criteria that would align for further evaluation with biodistribution imaging, maybe even dosimetry for tumor distribution within the mediastinum adjacent to the pericardium or in the retroperitoneal area or adjacent to the paraspinal region as these models are evolved. We've had a limited amount of data on a limited number of patients, and much of this information will evolve.

That's why at this present time, working through the biodistribution imaging and working through the medical community, as well as through the sponsor to gather further data, I think is going to be in everyone's interest to be able to gather that information and then come back with a more informed opinion in terms of where these elements may or may not be necessary in patient subgroups.

DR. NERENSTONE: So, this is not an issue unique to this one product. This is going to be an ongoing issue as we evaluate more of these monoclonals.



DR. SIEGEL: I would just add to the discussion, since your question mentioned and the complexity mentioned the use of the cold rituximab, that as our discussion has talked about and our question addresses what data might be necessary to allow us with comfort to think about not using the initial imaging or biodistribution step, that that is a different question from what might allow us not to use that initial rituximab dose which may be contributing to the therapeutic effect, may be lowering immunogenicity by depleting B-cells, may change biodistribution of the radiation by changing B-cells, we would need, I think, data about a single day type therapy before we would consider that --

DR. SAUSVILLE: As was pointed out by one of my colleagues to my right, it's possible that there's an element of interference actually with --


DR. SAUSVILLE: And that's a matter of future trials.

DR. SIEGEL: That synergizes.

DR. NERENSTONE: Dr. Lippman.

DR. LIPPMAN: I'd just like to ask Dr. Sausville to clarify for me. You seem to have a lot of concerns with the available data. Clearly it's extremely provocative. It's an active drug. That's not the issue, but you seem to be framing things that you're very concerned with the application of this agent as it's being evaluated now, not the idea that future studies would ask different questions. Normally when that comes up or often, that's the basis of accelerated approval. We think that it is going to lead to clinical benefit, but there are a number of concerns remaining, particularly in the context of one randomized trial.

So, I guess I was wondering how concerned you are with this. Again, this also gets at the issue that Dr. Nerenstone raised about how important are these subsequent studies to the safety of these patients.

DR. SIEGEL: Before we move on with that, let me just provide a little bit of clarification about the accelerated approval regulation. Accelerated approval does not change the standard of evidence or the standard of proof that, in effect, is present. That remains the same, substantial evidence from adequate and well-controlled trials, as it is for a regular approval. It just allows that standard to be applied to evidence based on other than the ultimate clinical outcome and in some cases surrogate endpoints other than clinical endpoints altogether.

So, just so we're clear, because there are potentially two ways of looking at this, accelerated approval may be applicable if we're convinced about an effect on response rate and not about an effect on clinical, but it's not to be used because, well, we're almost convinced about an effect but we're not quite sure about that effect per se. So, it doesn't lower the standard. I'm not suggesting your comments implied otherwise. I just want to make sure that we're all on the same page.

DR. LIPPMAN: No, I didn't imply otherwise. This is clearly an active drug. There's no debate even from someone who doesn't treat these kinds of patients. But it was very clear.

But I am concerned about the clinical benefit issue because of the concerns that Dr. Sausville raised. In an accelerated approval, there are mandated phase IV studies to look at some of these things, continue trials. That's what I'm trying to get the level of that issue from one of the experts.

DR. SAUSVILLE: Well, to comment, I think that part of the questions that I've explored in the course of the morning was this issue of the nature of the product because it is breaking new ground in what potentially would become widely available, namely, a targeted radioisotope for treatment as opposed to diagnosis. And our FDA colleagues have looked at this pretty thoroughly at this as a product. So, if they're convinced, based on the evidence at hand, that at least in the centers in which it has been used in a more research oriented sense, that it performed well, I take that as very encouraging and would actually encourage its more wide dissemination.

But I think what I tried to emphasize is that because these concerns in product utilization exist, is the isotope bound to the antibody, when we get more experience with biodistribution in outlying sites, whether the biodistribution curves and the agreed to label comparable populations of masses are actually looked at, that that be folded into a consideration of how best to use it and to really, as I said before, support the broader community in using it. That's my concern. It's not that there's a concern about the nature of things. It's how it's going to be ultimately translated.

With respect to the issue of benefit, in the rituximab-refractory population it's very provocative and impressive data even though it represents the level of response. I share Dr. Levine's concerns that in the other histologies we don't have as firm a notion of what potential value it might have simply because of the numbers. That's something that, again, might be part of future trial endeavors that could lead to an expansion of the indication on more solid evidentiary grounds.

DR. NERENSTONE: Dr. Przepiorka.

DR. PRZEPIORKA: A few comments. First, I want to commend the company on having an open access program folded right in and keeping it open during the regulatory review period. It should be a model for the rest of the pharmaceutical industry.

I am somewhat concerned about the toxicities, the hematologic toxicities specifically. But in fact if you look at the response rate of combination chemotherapy, the only ones that would give you something similar to this is going to be ICE or CVP which has similar hematologic toxicity and a lot more nonhematologic toxicity. So, I think if a doc knows how to deal with prolonged neutropenia, it will be fine, but if they don't, somebody has to teach them before they're going to be able to use this drug.

Having said that, I noticed in your backgrounder that a lower dose actually gives the same response rate in your phase I/II trial, and at some point someone should consider whether or not a lower dose should be explored with less hematologic toxicity as well.

I am concerned about whether or not the bone marrow should be harvested before this drug is given as we do with any other kind of radiation. I was not happy to hear any detailed information about CD34's or the number of apheresis to collect blood stem cells in patients who were treated. That would have made me more comfortable about making any comments about getting harvests done before this drug.

I was surprised to see that there were no complete responses in patients who were truly Rituxan-refractory since the indication being sought was for Rituxan-refractory patients, that all the CRs were in patients who were relapsed after a short period of time after receiving Rituxan.

On the other hand, the response rates in the follicular lymphoma patients are spectacular, and I have no doubt that it's the radiation and not the antibody that's doing this. That's pretty clear.

I think I am convinced that there will be a clinical benefit seen, if more patients were put on the study, since the curves for time off chemotherapy and time to progression, which are two hard and fast measures of clinical benefit, are pretty wide, and if the numbers were greater, the p values would probably reach significance.


DR. LEVINE: I just wanted to comment as well on the concept of clinical benefit versus response rate. From basically a history of caring for these patients, I accept it's not one dose one time. On the other hand, it's about a week, and given the fact that there's a week of treatment and the rest is yours, it would seem to me that this is a major clinical benefit, and I personally would be translating the response rate into a clinical benefit.

The objective data that they do have is small -- i.e., not all the patients answered the quality of life instrument and so forth -- and yet that correlates with the gut sense of what we're all hearing, what the people have written and said to us. So, frankly, I'm not at all worried about the concept of clinical benefit. I think it's there. They aren't being treated and the tumor responded very nicely.

DR. NERENSTONE: Dr. Blayney.

DR. BLAYNEY: Thank you.

Again, I think this is a good drug and the sponsor I think designed some clever trials to prove it.

I'm concerned. I think the myelodysplasia is going to be more of a problem than we've seen today. The 2-year follow-up is relatively short in the lifetime of lymphoma patients who in this country now have received a lot of alkylators. So, it is what it is, but I think we need to all be aware of that.

The clinical benefit has not been established by formal means. I think we've heard today both surrogates for it in terms of time off chemotherapy and also testimonials from very experienced investigators using it. But I'd remind the committee, and also in our duty of just trying to fulfill our regulatory requirements, that we haven't seen a clinical benefit demonstrated in the hard and fast rules. Nonetheless, I think I agree with both Dr. Levine and the other investigators that there is likely to be clinical benefit conferred by this agent.

DR. NERENSTONE: Any other comments? Dr. George.

DR. GEORGE: I would like to complain a little more about the sample size, I guess, because both the sponsor and the FDA were in on this from the very early days.

Some of this could have been anticipated. The design was to pick up a 25 percent difference in the overall response rate, which is remarkably similar to what was observed. So, it's fine. It reveals a strong statistically significant difference.

However, if you take a negative view and look at what kind of response rate might you have excluded for sure, the lower confidence bound is about 9 or 10 percent. If that had been observed, it might have, of course, colored what you're thinking, if that is the truth, not what is observed.

The same way with the time to progression. It was looking at clinical similarity, which means I think the eyeball test. You know, it looks close to me or maybe even a little better. But again, if you look at what you've excluded, there's still a reasonable possibility that in fact it's worse, even though it looks better now, in terms of time to progression even.

The third point about the sample size issue or the size of the studies is the issue of the histology. About 80 percent were follicular and the rest were roughly equally split into low-grade and transformed in the first study. This I think could have been anticipated or at least been guarded against some. If you do a trial and you really think you're going to be concerned about the results in all of the subgroups, as opposed to simply stratifying for the purpose of getting the balance and doing a little more efficient test -- that's one of the questions before us today. Is there any effect here in some of these very small subgroups?

This could have been anticipated some. You get these small numbers. It doesn't matter what they are. You're not going to find a difference. I'd just point out in the transformed group, I think there were 9 on the Zevalin and 4 in the other group, and 3 of those 4 responded. That's fine. But what if it had been 2 of 4? Well, then suddenly it's worse.

This could have been anticipated, and when you're at the end of the trial and you did it in a stratified way and you didn't anticipate it ahead of time, it seems to me you either buy the whole package or you don't with respect to the response rate.

One minor point about this I just noticed -- it's a question I should have asked earlier I guess. This first trial was stratified, but was the randomization also blocked in terms of numbers in each group? I guess it was.

DR. LANDIN: I'm Rick Landin, the biostatistician, and yes, we did block it.

DR. GEORGE: What was the block size?


DR. GEORGE: Too bad it was 9 and 4. If you're going to have a small group, it would have been better to have it a little smaller.

These are just random gripes about putting us kind of on the spot here now with these uncertainties.

DR. KEEGAN: With regard to the sample size, we did agree that if they demonstrated a 50 percent increase in the response rate or the delta of 25 percent, which they did, that that would seem to be fairly robust evidence of superior activity. At the time they designed the trial, we all recognized that they did not have clear ideas of what their targets were going to be in the study, as I recall, had a built-in interim analysis for re-estimation of effects in sample size, so that the study would have actually likely have been altered if they were seriously off on this.

When the interim analysis was performed and it suggested that they were adequately powered for the delta they were seeking, which, as I said, was about a 50 percent increase in response rate over Rituxan, the sample size as selected remained.

You're right that if we had concerns about the transformed data set, again it was essentially a risk that we allowed the sponsor to take by saying if the transformed patients behaved similarly and you have a large number of them, then the trial will be successful, and if not, then the trial will be unsuccessful and that is your risk to take.

In fact, what happened is we simply feel that we don't have a lot of information, and there was some level of discomfort with how well one should extrapolate to that data set.

With regard to the confidence intervals, we had some lack of clarity. The confidence intervals around the response rates, in fact, don't overlap. The difference is small. We've not powered studies based on --

DR. GEORGE: That's true. I'm just pointing out the confidence interval on the difference is a lot smaller then the observed difference, obviously. It's a wide confidence interval.

DR. KEEGAN: It's a wide confidence interval with the small numbers. Correct.

I guess we would like again to get back to the time to progression. Again, this may reflect some of our lack of understanding about the effects of Rituxan and how prolonged they were since that in itself is an active agent. We didn't have a wealth of data at the time. But, in fact, when we didn't require that the time to progression data -- that the study be powered to also show a difference in time to progression, in fact, as you look at the curves, the data do show a fairly healthy trend in terms of the difference, particularly again in that follicular subset.

I believe the company may also have the data on the combined nonfollicular and follicular subset as well, excluding the transformed. I don't know if it's quite as strong, but the trend did seem to be there and to be supportive. So, we considered that in addition as we looked at this application.

DR. NERENSTONE: Dr. Lippman.

DR. LIPPMAN: The comment that Dr. George made made we think again about other kinds of situations we've had on this committee. When a study is done like this, it is usually an all or none. You design the trial. It's randomized. And did it work in that population or not overall? Since this is going to come up in the labeling recommendation, I'd like to discuss this, although Dr. Levine made a very compelling case, if really the biology is there.

But in a sense in a small subgroup -- because we're not talking about a question of ER status in half of the patients -- a small subgroup of patients, to remove that group and say that this is not recommended would be a little inconsistent with some of the other aspects. The way this trial was designed, there are some differences. They're small numbers, but there was activity in both groups. So, I'd like to get Dr. George's thought on that because it seems to be a major issue in what we'll discuss in the labeling.

DR. GEORGE: My response is clear from the statistical point of view: you include them or you don't. You prove this as stated or you don't unless there is something that has come up that is so compelling from a biological viewpoint that should have been thought about ahead of time, that you shouldn't have had these patients in on this study because they're just completely different and you couldn't have expected them to do the same.

In this case you did. You kind of threw them in there. Maybe you thought that overall they would be a little different in the overall effect but you stratified because you just wanted to have balance and do a little more efficient statistical test. But the result is an overall result. That is, it works or it doesn't work in this group of patients.

DR. LIPPMAN: So, my question is, what has come up since the design of this trial that was approved by the FDA and now that would lead us to select out that small population where we saw substantial responses in both groups?

DR. SIEGEL: I would say, first of all, that the FDA has -- and I'd ask my colleagues who were more involved in the design of this trial, which I wasn't -- consistently identified that subset as a distinct subset that needs distinct data in its own right. As Dr. Keegan said, there was a risk taken as to whether those data would be adequate.

I think there are biological and regulatory reasons that are compelling to look at, at least the transformed subset, differently. One is that this therapy targets CD20, and those patients do not uniformly express CD20, and as a result they had a different admission criteria. I'm not sure if this came out or not, but those patients all required prescreening to be shown to be CD20 expressors as an entry criteria. That wasn't the case in other patients.

And these are not new factors since the trial was designed, but it's certainly a critical issue regarding interpretation of this trial. Part of this therapy is rituximab therapy, and rituximab is approved for follicular but is not approved for a transformed set. The sponsor, as I recollect, didn't seek that approval and didn't provide data in that group. I think this committee felt and the agency felt at that time that also was a separate indication requiring independent data, not so separate that data on other types of NHL are irrelevant by any means, but separate enough that one needs to look specifically at that population.

So, I would argue that this is by no means a post hoc subset, nor was this trial designed without, I think, all parties realizing that there was a real possibility of significant response differences in that subset.

DR. LIPPMAN: Since I guess it's not based on new biology, it puts the committee in a little awkward position because it would be nice if this was clarified up front in the protocol and this group was going to be looked at differently. To just put them into a study that is not a huge study -- it's an excellent study, outstanding, and moderate size. But there are a lot of these patients. If that was a concern, it would have nice if that would have been taken into the prespecified analysis plan.

DR. SIEGEL: They were stratified separately, were they not? I'm sure the analysis plan speaks of a separate analysis. I haven't seen it, but I'd be shocked if it didn't, if that's what you're implying. They were stratified separately.

DR. WHITE: The reason that CD20 positive was required for the transformed subset was because at the time of the design of the trial, there was a single patient who had received Rituxan and at transformation became CD20 negative. This was published by Stanford University. At the time, people didn't know how often the transformed patients could potentially be CD20 negative at transformation.

Subsequent to that, Dr. Maloney who was at Stanford at the time -- and I don't know if Dr. Horning is still with us here to address it -- it's been looked at. In fact, these patients are virtually always CD20 positive.

Now, it was stratified at the time. We had information in intermediate grade diffuse large cells. There were one or 2 patients in that group that were transformed, but the information that we were going on with regard to the activity in transformed disease was based on similar antibodies, but not the identical antibodies. So, we stratified so that we wouldn't be in a position where we had all the patients that were transformed on one arm and not on the other just in case there was any difference and then you wouldn't be able to tell whether it was because of that or not.

One last comment and that is, as you've seen from our briefing document, we also in a prospective way looked at every single prognostic factor that has ever been published in the literature with regard to lymphoma, breaking down the bulky disease, breaking down the bone marrow involvement, breaking down the extranodal disease sites, breaking down the demographics, breaking down splenomegaly, et cetera. In some of those groups that are really small, we didn't show a statistically significant difference.

But our understanding was, as spoken by the statistician on the committee, that one understands that when you start to do subset analyses like that, that you may find a small subset where you're not demonstrating a statistically significant difference, but that the presumption is this isn't the entire population you're studying. They have in common that they have low-grade follicular or low-grade follicular, transformed at least in part of the body to a higher grade, and that especially to an immunotherapy or a biologic therapy, one would hope that the range of responsiveness that you would see would be similar among those types of histologies. Obviously, some patients have poorer prognostic factors than others and that can sometimes influence response. So, that was the thinking at the time of design of the study.

DR. NERENSTONE: Dr. Lippman, a follow-up?

DR. LIPPMAN: Can I rephrase the question a little bit? What would you have wanted us to see in this small subgroup of patients that would have made it more convincing to you that we wouldn't separate these out? In other words, we knew it would be a small subgroup. They were allowed to be included in the study, which should be evaluated on the whole. There were major responses in both groups. It wasn't significantly different. So, can you give me a scenario where we wouldn't be having this discussion and we'd all agree that it should be used? 100 percent response in both groups?

DR. SIEGEL: I'm not sure there's any scenario or we wouldn't have this discussion. You have a different scenario, for example, in the IWG A and the transformed. I think it's worthwhile having the discussion. I think whether you consider transformed part of the same indication and you don't require data in transformed to give that indication or whether you consider it a separate indication is, I think, an issue that ought to be discussed regardless of what the data show. The fact that the data did not show a trend toward a higher rate in this subpopulation or to other better outcomes I think adds information to that discussion.

But I'm not exactly sure what you're asking. I think we're asking you for advice.


DR. LIPPMAN: I guess what I'm trying to clarify is that normally in every study that we review here, we look at the group as a whole unless there's some biologic change or whatever. We don't dissect out these small subgroups. I guess what I'm saying is nothing has changed in the biology. It was allowed to be included and the differences were not significant and there was activity. So, that's what I was getting at.

DR. SIEGEL: Maybe I can ask the company. I assume that there was a prespecified analysis or analysis by histological type. Right? This was a question of concern, getting a confirmation. So, it existed and the design rolled in, but it also existed in their prespecified hypotheses that we wanted to check to see where we stood in that regard.

DR. LIPPMAN: And they weren't significantly different.

DR. SIEGEL: Well, there's not a statistical significance between 3 of 4 and 5 of 9? Is that what you're asking?

DR. LIPPMAN: Right. That's the obvious point with the small numbers.

DR. WHITE: It was prospectively designed as a secondary endpoint that we would analyze by histology, by gender, by age, by bone marrow involvement. We had a list of variables that we had used in our prior rituximab studies and were all of the ones we find in the literature, LDH, et cetera. So, like with those variables, we did do a prognostic variable analysis and looked again with the Breslow-Day and the Cochran-Mantel-Haenszel tests.


DR. GEORGE: Yes, I just wanted to add one more thing on this point. My point was for the committee for our regulatory considerations, if we do anything differently with this group, we have to be very clear it's going to be based on some kind of a priori biologic notions that this is completely different not from the data. This data is totally inconclusive on this point, and we knew it would be from the beginning.

DR. SIEGEL: I would add that there is the design of this trial, but there are also trials that precede and follow it. So, if your assumption is that there isn't a biological difference and notwithstanding the fact that rituximab hasn't been studied in transformed, then I guess we would revisit the decisions regarding that population for that and for future trials as well. So, you can advise us what's appropriate with the study design.

But I think also bear in mind that we're talking about development in this field, and I would like to be clear. If that's the advice of the committee that this shouldn't be considered a separate population, that we need separate data and we should just presume the same, then that's advice that will be valuable advice and impact other regulatory decisions as well.

DR. NERENSTONE: Dr. Sausville.

DR. SAUSVILLE: Yes, I would just comment in relation to this discussion. Unfortunately, we don't have ways of recognizing subtypes of breast cancer, lung cancer, et cetera that have accepted differences, as it were, in biologic behavior, at least as accepted differences. I think in the case of lymphoma, we do have. Even though they're all CD20 positive and therefore were certainly appropriate from a scientific point of view to be in the study, from a clinical point of view, there are nuances in the behaviors that now we are faced with asking is the number of patients, as I interpret your question, that have these sufficiently large that we can feel confident about conclusions related to the subset. I think the discussion that went on previously emphasized that we probably can't really feel confident about, for example, the transformed, and I think that's more or less a reality of the types of lymphoma presentations that we have.

I think the questions do actually provide an opportunity to comment on whether or not that should be considered in making the final labeling indications.

DR. LIPPMAN: But I think my point is that normally to exclude a small subset, you have to feel very confident that they're different, not that confident that they're the same.

DR. SAUSVILLE: I feel very confident that transformed lymphoma patients behave very differently.

DR. LIPPMAN: To this treatment.

DR. SAUSVILLE: Biologically different. In terms of this treatment, we don't know.

DR. LIPPMAN: Well, but that's the point because they were included in the trial. You normally have to feel very confident that they're going to respond differently to remove them and not analyze them the way the trial was designed.

DR. SAUSVILLE: I mean, I feel confident that we don't have enough of them to know. That's the only thing that I know.

DR. NERENSTONE: Dr. Blayney.

DR. BLAYNEY: I too am troubled by the small numbers, but I think several things we ought to keep in mind. One is that they were included, we're told, because they had the CD20 histology on the cell surface. So, there was some screening and there's some reason to think they biologically are similar to the other much larger groups.

Secondly, this is a therapy that's designed for near misses. Often the transformed lymphoma have the low-grade in close association with an intermediate grade or what we think of as this transformed histology. But this therapy has a tissue penetration of 5 millimeters, we're told. So, there is some near miss use of this as a therapeutic agent for closely approximated cells that may not express the CD20 marker.


DR. LEVINE: I guess I just want to repeat again that I am uncomfortable really on the transformed group. They may have CD20 positivity, but it's a mixed bag biologically. Some of those patients have additional chromosomal aberrations or molecular aberrations. Some of them, in fact, are de novo transformed. It's a very mixed bag.

My concern is not just that we have very little data, which is true. The data we do have on one of the slides is on page 32. Of a total of 15 patients, the overall response rate, 40 percent. I can't say that that's different, but it's not 90 percent.

The other piece of information that I didn't have that would be important to me is if the patient with transformed lymphoma is treated and doesn't respond, then what we would do clinically is multi-agent chemo. That's exactly where you're going to use the multi-agent regimens and so forth, and that's exactly where you're going to do bone marrow transplant. I had asked the question, how many of the transformed patients then went on to chemo and how did they respond to chemo or what were their harvests like, because that becomes very important information clinically. So, it's what you would do beyond the Zevalin in those patients who don't respond to it.

DR. NERENSTONE: But I'm not sure what the sponsor has to say really matters because it's very small patient numbers, which is getting back to your point.

DR. WHITE: We did analyze that data since you asked that question. We brought our database with us and I can answer that question now.

Let me just say since 5 percent are transformed at 5 years but 90 percent by the time of death, patients transform more over time. So, by definition, they have more chemotherapy over time. In a way, it's sort of a surrogate for additional chemotherapy. Maybe that may be related to the chromosomal abnormalities and the poor prognosis.

We did have 5 patients in the transformed group that went on to additional chemotherapy. 1 went on to DHAP. 1 went on to CHOP. 1 went on to ESHAP, 1 methotrexate. 1 was just Decadron. So, actually four chemotherapies. None of them responded to any of those interventions.

DR. NERENSTONE: I just have one question again for really our lymphoma experts. My concern is because we have so little prolonged efficacy data in terms of time to progression or the standard that we do use, which is survival, I have no problems with this in the Rituxan-refractory patients. My questions is when this gets out into the community, it will be another option for physicians to choose instead of Rituxan. We have some data that if they fail Rituxan, they can get the new monoclonal and have a decent response rate. We don't know, once they fail the new monoclonal, if they can cross over to Rituxan. There's no data about that. So, in fact, there's a possibility that you could actually be decreasing survival. We've been told by the lymphoma experts who gave the drug that we know that these patients get multiple sequential treatments.

Again, I have no question that it's an active agent. We don't know yet where in the queue it belongs. Is that something that we should be worried about in terms of approval?

DR. KEEGAN: I think that's why we asked the question. We think that it's something that really needs to be discussed where it stands in the queue and whether or not this should be available as an alternative to Rituxan as performed in the 04 trial.

DR. NERENSTONE: What about the lymphoma consultant?

DR. LEVINE: It's a rough question. I don't know the right answer, to be honest with you. The response rate in the patients who had failed Rituxan was 60 percent. They've not been looked at head to head, but it looks like it's a little bit less, in fact, in the patients who were treated with the Zevalin alone, i.e., who have not had Rituxan before. There the response rates were 70, 74, whatever it was, a little bit less.

I don't think it would be wrong to ask for Rituxan first. The big issue to me is the long-term radioactivity and the long-term toxicity to the bone marrow and the myelodysplasia. That would be the conservative approach. I wouldn't be upset by that kind of approach. It's probably what would be done in the community in a general sense.

DR. NERENSTONE: Dr. Sausville.

DR. SAUSVILLE: I would agree. No one has any illusions about curing these patients, and I think the idea is to afford minimal intrusion onto lifestyle and minimal risk of toxicity. So, my own view of the queue is that this would be used, at least from the data we have, after Rituxan failure as its most obvious point of potential benefit.

Now, we obviously can't legislate that. Once it's out in the community, people are going to have their ideas about this. Again, that's part of the product labeling and part of how it's ultimately marketed.

DR. NERENSTONE: Well, actually the application is not for only in Rituxan-refractory patients. So, the application is really in previously treated.

DR. SAUSVILLE: Right. And the questions make a distinction here. We will have the opportunity to convey varying enthusiasm, I think, when we answer the questions.

DR. NERENSTONE: If there are no further comments, then why don't we get to the questions.

In the two clinical trials, the Zevalin therapy was associated with durable objective tumor responses, as well as a high proportion of severe and life-threatening hematologic toxicity of prolonged duration. Zevalin is a combination of both Rituxan and a radiolabeled monoclonal. Approval for this product requires demonstration that both components contribute to benefit and, therefore, there should be a determination that Zevalin provides benefits beyond those provided by the Rituxan alone.

In the setting of treating chemotherapy and Rituxan-refractory patients -- so, this is really now specifically the Rituxan-refractory patients -- do the data support a determination that the clinical benefits associated with Zevalin extend beyond those that could have been realized by retreatment with Rituxan?

I think the comments have been yes. Any further? Do you want a vote? Can we have hands up, hands down, or do you want a count? We have a count now.

So, 1a, does the data say that the clinical benefits with Zevalin extend beyond those that could have been realized by retreatment with Rituxan? We need to go around the room. Dr. Sledge?


DR. NERENSTONE: Everyone is voting, including our consultants, except for Mr. Ohye.

DR. BRIDGES: Dr. Bridges, yes.














DR. NERENSTONE: We haven't had a unanimous vote in two days.

Do the benefits associated with Zevalin use, clinically significant tumor shrinkage, considered together with the toxicity, both hematologic and other, support a determination that Zevalin is safe and effective in this setting? Again, we're talking about the Rituxan-refractory patients.

Dr. Sledge?
















DR. NERENSTONE: In patients who have not failed Rituxan, has Zevalin been demonstrated to provide benefits beyond those attributable to Rituxan alone? And I'll make things difficult. I'll start with Dr. Carpenter. Any comments? If you want to vote and add comments at that time, feel free to do that. Dr. Carpenter?






DR. NERENSTONE: I'm going to abstain.










DR. NERENSTONE: I'm going to change mine to a no.

The question we just voted, has Zevalin been demonstrated to provide benefits beyond those attributable to Rituxan alone in patients who have not failed Rituxan.


DR. NERENSTONE: The vote is 8 to 8, 8 yes, 8 no.

Again, in patients who have not failed Rituxan, is the net clinical benefit of Zevalin, as compared with Rituxan, higher overall response rate, absence of a clear difference on time to progression or overall survival and higher toxicity, sufficient to recommend approval for this patient population?

Dr. Przepiorka?

DR. PRZEPIORKA: Just to clarify, is this full approval?

DR. NERENSTONE: Right now we have before us full approval.

Dr. Carpenter, would you like to start again?

DR. CARPENTER: Let me think about this one for a minute.


















DR. KEEGAN: Dr. Nerenstone.


DR. KEEGAN: Since the committee has not recommended I guess under 2b approval for this indication --

DR. NERENSTONE: Well, let me just read the final count. Yes, 6; no, 10.

Now your question?

DR. KEEGAN: The only study in which patients with IWF A or transformed were studied was in the randomized controlled trial. So, we wouldn't need any votes but just some general discussion on this area, in particular, additional studies.

DR. NERENSTONE: In terms of the third question? The question, as written, is in the randomized, active controlled study, 106-04, which is what we decided wasn't enough for full approval, a small number of subjects with low-grade nonfollicular non-Hodgkin's lymphoma or CD20 positive lymphoma that had undergone transformation to a more aggressive histology were enrolled. The clinical behavior and level of CD20 expression in low-grade nonfollicular lymphoma and low-grade lymphoma that has undergone transformation may be sufficiently different from that observed in low-grade follicular non-Hodgkin's lymphoma to preclude extrapolation of the clinical results. The data obtained in these subgroups across other studies have not been as rigorously confirmed for histologic diagnosis or documentation of tumor response and duration.

There's the table that we've reviewed.

And it goes on to say that the Rituxan is approved for the treatment of chemo-refractory low-grade nonfollicular non-Hodgkin's lymphoma, the IWF A group. Although the data for Zevalin in this group are quite limited, the response rate was high, duration of response was similar for the patients who received Zevalin as compared to those who received Rituxan.

Please discuss whether the data are sufficient to determine that Zevalin has benefits beyond those of Rituxan and there's a net clinical benefit of Zevalin for chemotherapy-refractory low-grade nonfollicular NHL. In particular -- and I think this is really your question -- does this subpopulation require independent data, or can we lump them all together with the limited number of patients with IWF A to support a determination regarding the IWF A patients?

And if the data are insufficient, discuss the design of additional studies that would be acceptable.

DR. SIEGEL: I guess as Dr. Keegan was pointing out, this question is substantially different given the advice regarding the front-line trial. It should be pointed out, if it's not clear to this committee, that the trial for use in Rituxan-refractory patients specifically excluded patients with transformed or IWF A. So, it was only for follicular. So, if we were to approve in refractory -- well, the way the questions are worded is perhaps not targeting that particular, but if there's a feeling as to whether they should or shouldn't be together, more guidance on that I think would be useful.

DR. NERENSTONE: Dr. Sausville.

DR. SAUSVILLE: The approach that was taken here to stratify them is, in general, an appropriate one. I think the issue is whether or not you want to have additional understanding of the subgroup as a disease and consider that in relation to what you see. This is where, quite frankly, I think our database is a little bit less secure about the magnitude and intensity of CD20 expression. They clearly should be broken out as a separate group.

I think that in subsequent studies attention to the efficacy of the targeting is going to be key in understanding the true level of efficacy of this agent in comparison to the other. And that's how I would do it.

DR. NERENSTONE: I guess my response would be, because you're going to get into the transformed patients as well, that in that subgroup of patients looking at a phase II study with a response rate would be sufficient if this monoclonal antibody in other subtypes where you have many more patients available for study where there is a linkage between response rate and clinical efficacy and benefit endpoints, you don't need to redo a whole phase III, but a phase II with sufficient numbers to get a response rate would be a compelling supportive document to allow a broader indication.

DR. SIEGEL: I don't know that we could consider actually an indication for transformed NHL in patients who failed Rituxan since we haven't approved Rituxan for transformed NHL.

DR. NERENSTONE: No. I'm saying if there are other studies that you get a first indication, then I think a supplement looking at just those patients in a phase II would be appropriate. You don't have to do a large phase III. That would be a supporting indication.

DR. SAUSVILLE: Actually to elaborate on that, you could make the entry criteria for the phase II whatever you want in terms of prior Rituxan treatment.

DR. NERENSTONE: Dr. Przepiorka.

DR. PRZEPIORKA: The other concern that I have is if we start saying yes to follicular and no to IWF group A, we may end up having to do this for a lot of other protocols with hematologic malignancies coming down the line. And since we treat them fairly similarly and we haven't proved that this treatment is different and we don't have an expectation that it would since they are both low-grade and CD20 positive, so the mechanism of action shouldn't be different between the two groups, and more importantly, they're a very small percentage, which will give you an idea of whether or not you'd be able to actually do a study within a reasonable period of time, I would suggest not having to do something different about that particular group.

DR. NERENSTONE: How do you feel about the transformed patients?

DR. PRZEPIORKA: Oh, those are a completely different group, and I don't know that you can make any conclusion from this data. I would agree that it's nonconclusive, and I was surprised to see it in here at all.


DR. LEVINE: I would agree. It seems very reasonable to me to include the IWF A patients in with the follicular patients in this application and would feel very different about the transformed.

DR. NERENSTONE: Yes, Dr. Redman.

DR. REDMAN: Not as a lymphoma expert, but as a clinical trialist, at a previous meeting or two meetings back, we had a trial that was totally negative and the industry was trying to support a stratified arm as being positive specifically for that group. I really look at this as the converse. If it wasn't decided beforehand, they stratified to make sure the risk factors were equal in all arms, and the majority felt that it shouldn't be approved for this indication. But if we had approved it and then gone and nitpicked on subgroups, I don't think that's appropriate the way the trial was originally designed.

Bowing to the lymphoma experts, if that's a problem, then the trial should have been designed differently, but this is the way it was designed and this is with the approval of the FDA.

DR. NERENSTONE: Dr. Lippman.

DR. LIPPMAN: Yes. That was exactly my point. You indicated it much more eloquently than I did. I don't treat these patients as a head and neck doctor. The issue is as a clinical trialist I have a concern with designing a study to answer a question and then pulling out a small subgroup of patients and saying they're different. That should be something that's indicated a priori.


DR. GEORGE: Just to add one more time something to this point. Let's be clear about what the stratification does in clinical trials. The reason you do stratification is to have balance and to provide for a little more efficiency in the overall test. It is not generally to answer the question separately in each stratum. That is explicitly not the purpose. And I think we do get into trouble if we, after the fact, start looking at it as one of the purposes and say, well, we stratified it, so we should be looking at these results. Most likely we shouldn't even be doing anything like computing a p value within each of the stratum. Really, it's the overall that counts. You can look at differences as sort of some information, but it's not a generally acceptable thing to do.

DR. NERENSTONE: Did the FDA get enough sense of the committee?

Can we go onto number 4? Do we need to go on to number 4? Pat, do we need to go on to 4?


DR. NERENSTONE: All right. The initial step, the step 1 administration of the Rituxan and the indium-labeled monoclonal, is an essential component of the therapy. There are no data on the safety and effectiveness of the Zevalin using only one dose of Rituxan, the elimination of step 1, and an inadequate safety database in patients who received Rituxan alone without radiolabeled material in step 1.

They're worried about the patients who have preexisting anti-murine antibody and that might be different than that observed in clinical studies. No other screening, i.e., HAMA, has been adequately evaluated to identify patients at increased risk of altered biodistribution.

Then they're also worried about alterations of clearance for mechanical reasons or based on tumor proximity and that may provide information on radiation dosimetry to assist in assessing cumulative doses for future planned radiotherapy.

The agency seeks advice on the additional post-marketing studies to better assess the utility of using indium-labeled monoclonal for determination of biodistribution, as a component of step 1, in optimizing the safety and effectiveness of the Zevalin. What types of studies and other data should be collected to determine the safety and effectiveness of deletion of the biodistribution assessment while retaining the first dose of Rituxan?

Dr. Przepiorka.

DR. PRZEPIORKA: I think the first step that's necessary in order to answer any questions that we could possibly ask about this is do we have the software to accurately measure what we want to measure. I think my sense from Dr. Meredith's presentation is we don't have that available. So, unfortunately, I would not be able to say yes or no or what to do without knowing what it is that we can do or what it is that we actually have done, since all of our correlations between the dosimetry and toxicities were based on possibly faulty calculations.

DR. KEEGAN: Dr. Przepiorka, could I clarify that there are two different things you could do with this initial step in the imaging. The first is a rather qualitative imaging assessment which requires no software other than the radiologic film and a radiologist to be looking at it and to get a general assessment. I think we were focusing somewhat on that more than on the dosimetry question because we would agree with you that the dosimetry is really not well developed enough to make very accurate predictions about the dosimetry with this type of a radiation source.

But we were also concerned about the issue that there are qualitative differences that can be detected on imaging. For instance, the outlet obstruction issue that we mentioned is quite frequently detected on other kinds of scanning that we see commonly where it's just an imaging study, and it could also be seen here where it might be predictive or a gross alteration in biodistribution which might be indicative of some alteration in clearance which may or may not be immune mediated. I don't think we have a good understanding of all of that.

So, if you could separate out the dosimetry issue even from just a gross biodistribution assessment in your response, that would be helpful as well.

DR. NERENSTONE: I just have one question. Everybody keeps getting back to urinary obstruction. For clinically significant urinary obstruction, for those of us who use cisplatin all the time, there's a much cheaper way of doing it than a radioisotope scan, and that's called a creatinine. Is there an indication that the creatinine may be normal but the biodistribution may be altered when it's just an obstruction with normal creatinine?

DR. MILLS: Part of the issue comes from, one, the kidneys are very sensitive in this system in terms of the radiation effect that you may elicit, and if you get slow clearance or incomplete clearance from the kidneys, number one, that's a sensitive organ.

Number two is we have also with other radionuclides just recently demonstrated significant problems with the bladder, and hemorrhagic cystitis has been another occurrence.

So, while your creatinine may not be affected necessarily, the residence time within these other organ structures may preclude safe distribution. So, if we have an image study that we're already performing at the present time, the biodistribution, it's another sensitive indicator, not necessary to add an additional type of evaluation. And then you may have other evidence such as evaluation of the adequacy of the preparation of the dose. That's another element that could be also assessed.

But the other concern, of course, is to know where the distribution of the tumor sites that are localized where the radiation oncologist may have additional information or understanding or want to be able to understand further where there should be applicable concerns in terms of their treatment fields too.

DR. NERENSTONE: So, do you foresee then the indium imaging as always being needed because many of these patients are going to go on to subsequent radiation therapy?

DR. MILLS: Too broad of a question. I think when we start out with this limited data set, my concern is to gather information for the various community hospital settings. However, I could imagine that you will find subgroups in the future, especially because you have other imaging modalities where you may know the distribution of the tumor, we may have assessed that some of these safety concerns have been able to be relieved with a broader experience. So, I wouldn't say forever, but my concern is that there may be patient populations that you will see in the clinical setting that you may never leave biodistribution imaging; other groups that you may say, no, this group is safe in terms of being able to evaluate them. We just don't have that body of evidence to have that confidence yet.

DR. NERENSTONE: Dr. Sausville.

DR. SAUSVILLE: I guess the question, as asked, is a fairly open ended one. There's obviously, from the research standpoint, lots of interesting things that you could think about doing here. Could maybe you elaborate on what you would regard as easily obtained parameters with what you are likely to have in hand and how these might, in the agency's mind, have an impact in defining the further use of this?

DR. KEEGAN: I think our concern, because I'm not sure, as you say, the imaging is the only way to identify outlet obstruction, that there might be other modalities. But in particular, I think we don't feel that we have truly assessed in an adequate population what might be the incidence of abnormal biodistribution that would suggest that the product is not going to go where it's intended and that it would be inappropriate to administer it.

That gets back again to the question. If the incidence of that happening in a particular well-selected patient population was less than a certain amount, we could collect data up until we have an adequate experience to exclude that altered biodistribution doesn't happen in more than .5 percent or .2 percent or .002 percent of the population who might be inappropriately treated if the committee felt that way.

So, I guess what we're asking is, is there some level of safety data or incidence of an adverse event that would be so uncommon -- in this instance, altered biodistribution -- that you don't think that it would be necessary to prescreen patients to look for it?

DR. SAUSVILLE: But the interesting question that you raised is do we know what a normal distribution is and how do we get that.

DR. KEEGAN: Again, this is based on our experience with monoclonal agents to date. Our thought is that there is a fairly clear pattern of normal distribution and when there has been evidence of an immune response, that the biodistribution is so drastically altered that it's a fairly gross finding and fairly easy to detect. So, we think that normal biodistribution can be described and abnormal biodistribution can be described.

DR. SAUSVILLE: I think everyone would support the idea of continued data to address this point because, as was pointed out, if you could eliminate it, I think it would be a lot easier to use, recognizing that you still might need to determine what the role of the two versus one addition of rituximab in the regimen is. On the other hand, if you can define subsets in which it would be the bellwether of either success or failure, that would be equally important.

DR. KEEGAN: Is there a particular incidence of abnormal biodistribution that would likely preclude efficacy that you would find acceptable to miss, not to seek, through an imaging evaluation or through a biodistribution evaluation?

DR. NERENSTONE: Well, given that this is going to be a concern of more than this one application, my suggestion was that you convene a panel of imaging experts who are used to looking at dosimetry because I think it's probably not something that most of the medical oncologists are very comfortable talking about, setting limits in future studies. I think it's a good question and an important one, but I'm not sure this forum is the right place to address it.

DR. SIEGEL: These studies can address the extent to which this early biodistribution study will generate information that might impact whether the patient is treated or what concerns there are, safety or efficacy-wise.

But one of the things it won't address -- and I wonder if there's some insight -- is whether it will provide useful information about future management of the patient with external beam radiation. I don't even want to go with future radiation therapies that don't now exist, but with future external beam radiation. Is this a population that has a significant probability of later receiving external beam radiation, and if so, would the information from where this drug dosed the patient be useful in planning and designing that later radiation?

DR. NERENSTONE: Dr. Bridges, you're a radiation therapist.

DR. BRIDGES: Yes. I think it's a very important point. I expect that if this is approved, there will be a push in the community, and I'd ask the medical oncologists if there would be use of this prior to potential radiation therapy and that radiation would highly be likely to come on later at a point in the course of the patient's care. Because the response rate with radiation to a lymphoma mass is approaching 90 percent, so it's really the best single agent we have. But many times we reserve it for later use because we want to get the systemic problem taken care of.

So, I think it's going to be important to be able to identify the dosimetry issues related to the critical structures like particularly spinal cord, kidney, and things like that. So, I think it's going to be paramount that we do get more dose analysis done.

There's got to be a clarification to the medical oncology community somehow that this is an issue, that when you get a patient, you've got to, in your review of systems, ask have you had this treatment. I mean, obviously we do. And then it's got to be something that's in the package insert, that there's a big, bold precaution, and it's got to be communicated, even at our national meeting level, to make sure that this is an issue for us. Because it's not been raised before significantly, as far as I know, in the community.

DR. SIEGEL: Would you anticipate then, if you had a patient that you were going to give external beam to who had previously received this product, that you would or might want to go back and look at the indium imaging to see where the radiation from --

DR. BRIDGES: Clearly, the data you provided today -- if the patient had a paraspinal mass with a cord compression and I had the risk that he had already gotten 8,000 rads to the superficial spinal cord and now I'm contemplating 4,000, 5,000 more, it would be very important for me to know. The concern I would have -- if people are aware of it, they're going to look and they're going to get the indium study. They're going to look at it. They're going to verify where the tumor mass was, but if they're not aware of it, it's not something that we normally think about in consideration of radiation after other radioactive treatments.

DR. NERENSTONE: Dr. Blayney.

DR. BLAYNEY: To get back to the biodistribution question, I think this is going to be tedious enough to use for medical oncologists that it will generate substantial pressure on the company to perhaps come to you and present their data. And for us to pose a hypothetical 2 percent or some number at this point, based on limited experience, I think is asking too much. So, my advice would be to keep an open mind, and if there's a subset that this sponsor or other sponsors can identify where it's no issue, then you might want to go ahead and approve abandoning the imaging dose.

DR. NERENSTONE: But I think someone has to be keeping a record of what's going on because we're going to be out in the community and those people who give it are going to give it, and then they go on and they get it. And if nobody is looking at these results in a centralized way, we're not going to have any idea at the end of the day what we're doing there.

Dr. Sledge.

DR. SLEDGE: In contrast, I've got to ask. You've treated close to 500 patients, according to the sponsor, and the biodistribution issue basically hasn't been an issue in those 500 patients. Let me ask the agency what reasonable trip wire is going to be required if you've already got data on 500 patients. 2,500, 25,000? How many more do you think you need?

DR. KEEGAN: Well, in fact, we don't have dosimetry information on close to 500 patients. It's more on the order of low 200's. Right?

DR. SLEDGE: No, but that's what you have safety data on, 500 patients. It's reasonable to suggest that this has not been a major problem to date with 500 patients.

DR. KEEGAN: I guess our concern is that alteration of the biodistribution may alter both the safety and efficacy profile, safety we may not have seen or may not have observed, particularly if it was just one person or two people. I guess what it goes down to, if we have a sufficient number, if we conclude that this might occur in less than 1 percent of the population, again --

DR. SLEDGE: Don't you think you're already pretty close to being there?

DR. KEEGAN: We may be close to being there. We don't have the correlative biodistribution data except for about 200 folks, though. So, we're not quite at the 1 percent rate.

DR. SLEDGE: From a kidney, ureter, bladder standpoint, this certainly seems safer than cisplatinum, for instance.

DR. SIEGEL: I guess there are a couple other differences between use in clinical trials and use in practice that have come up in our discussion. These patients were screened for lack of anti-murine antibodies. It has not been proposed that that screening be done, but one could study whether such screening would be a reasonable alternate. We know from, at least other products, that such antibodies can cause radical changes in the distribution of labeled antibody. I assume there were other screening parameters regarding kidney function and other factors that may or may not be applied the same way in the community.

But I guess the reason I asked that other question was not so much to change the topic, but if in fact these indium images are going to be important in guiding future radiation therapy of the patient, then even knowing that they're not important for whether the next dose is safe, that raises the issue of whether they have a critical role in the patient management in any case.

DR. NERENSTONE: Dr. Przepiorka.

DR. PRZEPIORKA: I think I know the answer to this question, but I'm going to ask just to make sure I do. What we want to know is how much radiation the patient got to whatever organs we need to know, and that historical information is critically important when you're having the patient going on to more radiation. But the scan that actually tells you what radiation was given was the one after the yttrium, not the one after the indium. So, the question is, can you do a scan with the therapeutic drug?

DR. MILLS: The indium image acts as a surrogate for the yttrium. The yttrium is essentially a pure beta emitter and therefore it's not imageable, but the indium has been an acknowledged surrogate in this situation.

From the standpoint here, I can imagine the application of looking at those biodistribution images over several time points where an estimation of dosimetry can be utilized and can be evaluated by the treating radiation oncologist to further assess and determine an approximation of the radiation that's been given in a certain area. So, dosimetry may possibly be a retrospective analysis in certain clinical settings.

DR. NERENSTONE: Just to recapitulate, we voted for approval in the Rituxan-refractory follicular population. Do you want further discussion about other approval?

DR. KEEGAN: I guess the only question I want is a little bit more clarification of an acceptable clinical benefit endpoint that the sponsor might consider in terms of designing new trials. I know survival has been brought up several times, and I was wondering if there are other endpoints that the committee would recommend or would think might be appropriate. I noticed that time to progression, time to next treatment are also other kinds of endpoints that have been looked at in these studies, and whether the committee feels that only survival or some other endpoints might also be better predictive of clinical benefit for the population.


DR. LEVINE: I would suggest several things. First of all, somewhere in the paperwork it stated that the patients who came on to the study required treatment. So, that would mean to me that there was a symptom somewhere down the line or a sign. So, simply number one, what are the tumor-related symptoms and signs prior to study, what are the tumor-related symptoms and signs post.

One of the easy ones -- you don't see it so often in low-grade, but I'll just say it anyway -- the systemic B symptoms, fever, night sweats, weight loss. That should be captured.

I would certainly go back to your verified instruments, the quality of life instruments, and just make sure that the patients really do fill them out and so forth.

So, those three things would be very useful and helpful in a clinical sense to show improvement.

DR. NERENSTONE: I guess there's one more question about the pediatric population. I know that Dr. Levine did address that in some of her earlier comments. Does anyone else want to comment? Dr. Przepiorka.

DR. PRZEPIORKA: Unfortunately, I'll have to disagree with Dr. Levine and clarify.

The question the way it's written is not exactly the way it went at the pediatric meeting. I believe we indicated that follicular lymphoma in the pediatric patients was a different disease than it is in the adult patients. That's very clear biologically and molecularly. There wouldn't be enough follicular patients in the pediatric age group to do a study.

But I think the sense of the folks on the committee was that if the drug has a target and is not just a broad cytotoxic, that it may be of value to study this in the pediatric population if that target were available in a similar or other malignancy in the pediatric group because once it gets out into the community, it's going to get used by the pediatric docs with no information whatsoever.

The history of Rituxan now is that it's being used for anything that's CD20 positive, not just follicular lymphomas. There are lymphomas in the pediatric population. Albeit there is a risk of leukemia or MDS when you give radiation plus chemotherapy, the general gestalt is it's a risk/benefit analysis that has to be made and that kid will never grow up if they don't get something. So, I don't know, unless the COG people say, no, we have 20 more drugs for lymphoma that have a 90 percent response, that we should turn this down right off the bat.


DR. LEVINE: I guess I agree and disagree. I totally agree this should be studied. There's no question that it should be studied. My problem with it related to long-term toxicity with very little data in a population that are not age 60 when this begins but rather age 6. I would very much be in favor of additional study in the pediatric age population.

DR. NERENSTONE: As opposed to a waiver of those studies.

DR. LEVINE: As opposed to waivering, saying, fine, it's there, start using it.


DR. ALBAIN: I wanted to come back again to the accelerated approval question. You answered her, I think, no, that you didn't need it. My understanding would be you have an indication voted yes for the refractory population. So, you then would not want a recommendation regarding accelerated, yes or no, for the non-Rituxan-refractory population?

I'm just concerned here that the pivotal trial that we saw today was in that population and that there was a statistically significant benefit for response, which for all the other compounds we review, would qualify for an accelerated decision.

In the context of that, I was walking in during that vote for 2a, and I wanted the record to reflect my vote to be yes for 2a, please. I was on the wrong question when I sat down.

DR. SIEGEL: If there might be significant sentiment that people would vote differently on a question of accelerated approval -- and I think even more relevant is really the 2b issue. I think the benefits in terms of response rate are there. The issue is is there net clinical benefit sufficient to recommend if you wanted to put accelerated approval --

DR. ALBAIN: Some of us are distinguishing, though, maybe falsely so, but I voted, for example, no on 2b because it was a conventional approval vote. Whereas, for an accelerated vote, I might vote differently for that question.

DR. SIEGEL: Right, I understand. I think it would be useful to reask 2b in the context of accelerated approval provided that if the committee feels favorably, I'd like to hear from the members as to what should be the appropriate study to confirm that there is clinical benefit to this usage.

DR. NERENSTONE: Dr. Lippman.

DR. LIPPMAN: Yes. I'm really glad this came up because I was really having a problem with this. The bar was set very high for 2b and I think that's why I brought up this whole discussion of accelerated approval. There is no question I would vote to approve this on accelerated approval, and I'd hate the label to not reflect that. We need more information and we should get it, but right now, I'd hate for patients to be deprived of this agent in this setting based on this incredible trial. It was really well conducted, and what we saw was highly significant.

DR. NERENSTONE: My understanding is that we do not have an accelerated approval application before us. The company can come back to you with that, but a vote today would not grant them accelerated approval.

DR. SIEGEL: Well, your votes, of course, even if we had the application, don't grant anything.


DR. SIEGEL: I didn't mean that to sound so flippant as it came out. My apologies.

The other part of that statement was if the committee were to vote that they recommended that this drug receive accelerated approval, that advice would be extremely valuable advice should the company, which wouldn't surprise me at all, decide that they might be interested in submitting such an application. So, the fact that we don't have that in hand I don't think means that it's an issue in terms of -- the committee's vote on that question I think would be quite useful.

DR. NERENSTONE: Then I think it's probably worth -- the sentiment of the committee is they want that vote and the question is do the data presented support approval for an accelerated approval for Zevalin for the prior treated population. Which population? Follicular lymphoma?

DR. LIPPMAN: The pivotal trial population.

DR. NERENSTONE: The pivotal population of the large phase III study.

So, the official wording is do the data presented support accelerated approval for the pivotal population of the large phase III study. Dr. Sledge, would you like to start out?

DR. SLEDGE: What is pivotal population?

DR. SIEGEL: Why don't you make it two votes as originally planned, first asking whether there ought to be an approval? Then you could ask separately whether it ought to be restricted to transformed, which is sort of the way that the questions were designed. That might be a cleaner way to get the sense of the committee.

DR. NERENSTONE: So, the first is just approval and the second is whether the smaller population should be taken out.

DR. SIEGEL: I think that will give us a clearer -- because otherwise if people vote no on the defined population, it won't be clear whether they think it should be --

DR. NERENSTONE: Okay. So, the first vote is approval on the big group, all three subsets as described in the phase III trial, for accelerated approval.

DR. TAYLOR: And what's the second one?

DR. NERENSTONE: The second question is going to be should the smaller populations be taken out. So, you can vote yes on one and yes, they should be taken out on the second, or you can vote yes on one, but no, they shouldn't be taken out, or you can vote no on both. And there's probably another permutation there too.

DR. SIEGEL: If you vote no on one, we'll skip the second.

DR. NERENSTONE: Right. If we vote no on one, the second is immaterial.

Is that clear as mud? It's the total population. The question is does this data support accelerated approval.

DR. SIEGEL: For patients who have not failed Rituxan.


Dr. Sledge.

DR. SLEDGE: I'm going to vote a strong no here. First, I sincerely dislike the idea of using accelerated approval as a back door for approval when we've turned down a drug for what we consider good reasons. My very strong feeling here is -- the reason I voted no at least -- that I didn't have a clue if there was real clinical benefit here, and nothing I've heard in the last few minutes has changed my mind on that.

DR. NERENSTONE: A point of clarification, Dr. Lippman.

DR. LIPPMAN: The longer I'm on this committee, the more confused I get about accelerated approval. I thought accelerated approval is when you strongly believe. I mean, we saw those curves. We saw the data and everything was going in the right direction. But you're right, it hasn't been absolutely proven. In that case I would vote for accelerated approval and that the subsequent studies would be done to prove that.

DR. SIEGEL: One of the problems with accelerated approval -- that's correct. You need to firmly believe that there is a benefit on a surrogate endpoint or a clinical endpoint other than the critical clinical endpoint. "Firmly believe" isn't the legal standard, but that's a fair characterization.

One of the problems with accelerated approval that I think people grapple with in these settings, though, is that there is still a risk/benefit. You still have to assess whether the benefit is adequate to find the toxicity profile acceptable to determine a drug to be safe and effective. Of course, when you're dealing with a surrogate that hasn't been validated and you don't know how much clinical benefit there is with that surrogate, it implies a significant amount of guesswork when there are significant toxicities.

DR. NERENSTONE: Mr. Ohye, just a point of clarification.

MR. OHYE: Although I'm not a voting member, I have the regulation in front of me, and I think Dr. Siegel would agree that the test is that it's reasonably likely to predict a clinical benefit. So, you don't need this overwhelming data that you would need for full approval.

DR. SIEGEL: You need the same quality of data to prove an effect on the endpoint.

MR. OHYE: Absolutely, yes.

DR. SIEGEL: Reasonably likely is the level for presuming that it will translate to a clinical --

MR. OHYE: But once you've passed that hurdle and you obtain these data, are these data reasonably likely to predict a clinical benefit.

DR. SIEGEL: That is correct.

DR. NERENSTONE: Okay. I'm going to take the vote and everybody can have a 1-minute comment if they want. To read it again, do the data presented support accelerated approval for the total population of the large phase III study, that is, the patients who have not failed Rituxan?

Dr. Sledge.


DR. BRIDGES: Yes. I think we'll unlikely ever show a survival advantage. I say yes.

DR. REDMAN: I still think yes for full approval is a better answer, but if that's not the question, then yes for accelerated approval.

DR. TAYLOR: I agree with him. Yes.


MS. KRIVACIC: I would say yes as well, but making sure that commitments are kept to continue with the trials in a timely fashion.

DR. GEORGE: I would abstain for a reason that hasn't been expressed, because I can't judge this, since it wasn't presented this way. I don't know what those promises will be for future studies, and that would affect my vote.

DR. BLAYNEY: I vote yes.




DR. LIPPMAN: I vote yes, but in my 1 minute, I'd like to say that I do agree with Dr. Sledge completely, that often this is used for a back door. In this case I'm voting yes because I firmly believe that it will be, seeing all those curves in the data in this well-designed trial, but we haven't seen it yet.

DR. LEVINE: I vote yes.

DR. PRZEPIORKA: I vote yes, and I hope to see the current trial extended to the right number of patients.



DR. NERENSTONE: Then the second question is should --

DR. SIEGEL: Was that comment you'd like to see it approved but to continue a placebo-controlled trial?

DR. PRZEPIORKA: No. To extend the numbers.

DR. NERENSTONE: And the vote is 13 yes, 2 no, and 1 abstention.

Then the last questions, guys -- last question -- is whether you think those subpopulations should be taken out of the larger group. Do you want it to be broken down into transformed and the other low-grade?

DR. SIEGEL: I think transformed is -- based on the discussion, I think that's the only one.

DR. NERENSTONE: Okay. So, the question is, do you think the transformed patients should be taken out of the indication even for accelerated approval? Dr. Sledge, we'll start with you.



DR. REDMAN: I really don't know how to answer that question from a clinical trialist standpoint.

DR. NERENSTONE: Is that abstain?

DR. REDMAN: Abstain.

DR. TAYLOR: I'm going to abstain too.


MS. KRIVACIC: I say yes.




DR. NERENSTONE: I say yes. I think this is a separate patient population and we need a phase II of those patients to see what the response rate is, not folded into a larger group or we're not going to be able to see the effects.

DR. ALBAIN: Yes, for the same reasons.

DR. LIPPMAN: No. I think the drug should be made available, but as part of the accelerated approval, more data should be accrued, and that may revisit this question. But at this point, no.

DR. LEVINE: I vote no, but with the provision that in the new trials, there would need to be sufficient power in that group alone to see what you're doing.




DR. NERENSTONE: 2 abstentions, 3 yes, 11 no.

I'd like to thank everybody for their --

DR. SIEGEL: Wait. I'm sorry. I know everyone is eager to go, but I had requested twice. If we are to do an accelerated approval, we need to determine what's an appropriate commitment. I guess we heard one suggestion for a significant, I guess, substantial continuation and enlargement of the head-to-head trial as a way -- and presumably that would be with changing the primary endpoint to a clinical endpoint, time to progression. Are there any other comments about how one would confirm a clinical benefit here?

DR. NERENSTONE: Dr. Lippman.

DR. LIPPMAN: Well, I just bring this up. I think we should certainly treat more patients with transformed disease, as Dr. Levine mentioned. But someone mentioned this during the discussion, that those curves look pretty different. If they continue at that level with further follow-up, they're going to be significant. And the question is if we revisited this in a year or two, with just very close follow-up of those patients, would that be enough to answer the clinical benefit question in addition to a smaller phase II study in transformed patients.

DR. SIEGEL: My guess is if we looked just at follicular subset, that's likely to happen. If we look at the total population, it seems less likely.


DR. KELSEN: If the drug receives accelerated approval, how would you expect to accrue adequate numbers of patients? Would you anticipate being able to accrue adequate numbers of patients?

DR. SIEGEL: Well, in this particular case, rituximab also has an approval for that usage and obviously has a different safety profile.

DR. KELSEN: Let me understand the design. The suggestion was to continue the current design of randomizing patients to rituximab alone or rituximab plus the antibody, with the antibodies now commercial available.

DR. SIEGEL: No further comment.

DR. NERENSTONE: We have some announcements.

DR. TEMPLETON-SOMERS: I'd like to thank you all for coming and staying. I admire your dedication and your concentration.

I have a few announcements. Members, if you're going to stay here, you must recheck into the hotel as soon as possible. Be prepared. It's going to be above government rate. We'll deal with that later. And the same with your travel plans. We'll figure that out later. Just do what you've got to do.

Also, for the general public, the Anti-Infective Drugs Advisory Committee previously scheduled for tomorrow and Thursday has been canceled, as has the Anesthetic and Life Support Drugs Advisory Committee, which was previously scheduled for the 13th and 14th. That was a huge public interest. That was the opiates meeting. And again, the Gliadel session this afternoon is canceled. Thank you very much.

DR. SIEGEL: Oh, wait. One more announcement. There is a severe shortage of blood in New York and Washington, D.C. I think this is an audience that should be aware of that situation.

(Whereupon, at 1:50 p.m., the committee was adjourned.)