00001

1
2
3
4 VOLUME I
5 (Morning Session - November 15, 1999)
6
7
8
9
10 HUMAN TUMOR ASSAY SYSTEMS

12 HEALTH CARE FINANCING ADMINISTRATION

13 Medicare Coverage Advisory Committee

14 Laboratory & Diagnostic Services Panel

20 November 15 and 16, 1999

22 Sheraton Inner Harbor Hotel

23 Baltimore, Maryland

00002

1 Panelists

2 Chairperson

John H. Ferguson, M.D.

Vice-Chairperson

4 Robert L. Murray, M.D.

5 Voting Members

David N. Sundwall, M.D.

6 George G. Klee, M.D., Ph.D.

Paul D. Mintz, M.D.

7 Richard J. Hausner, M.D.

Mary E. Kass, M.D.

8 Cheryl J. Kraft, M.S.

Neysa R. Simmers, M.B.A.

9 John J.S. Brooks, M.D.

Paul M. Fischer, M.D.

Temporary Voting Member

11 Kathy Helzlsouer, M.D.

12 Consumer Representative

Kathryn A. Snow, M.H.A.

Industry Representative

14 James (Rod) Barnes, M.B.A.

15 Carrier Medical Director

Bryan Loy, M.D., M.B.A.

Director of Coverage, HCFA

17 Grant Bagley, M.D.

18 Executive Secretary

Katherine Tillman, R.N., M.S.

00003

1 TABLE OF CONTENTS

Page

2 Welcome and Conflict of Interest Statement

Katherine Tillman, R.N., M.A. 5

Opening Remarks & Overview

4 Grant Bagley, M.D. 10

5 Chairman's Remarks

John H. Ferguson, M.D. 28

Brian E. Harvey, M.D., Ph.D. 30

Open Public Comments & Scheduled Commentaries

8 Frank J. Kiesner, J.D. 48

Larry Weisenthal, M.D. 57

9 Randy Stein 92

Richard H. Nalick, M.D. 99

10 William R. Grace, M.D. 108

John P. Fruehauf, M.D., Ph.D. 110

11 James Orr, M.D. 127

Robert M. Hoffman, Ph.D. 131

12 Andrew G. Bosanquet, Ph.D. 136

David Alberts, M.D. 142

13 Robert Nagourney, M.D. 147

David Kern, M.D. 159

14 Daniel F. Hayes, M.D. 168

Bryan Loy, M.D. 178

LUNCH 196

VOLUME II

Open Public Comments & Scheduled Commentaries

18 Edward Sausville, M.D. 201

Harry Handelsman, D.O. 227

19 Harry Burke, M.D., Ph.D. 234

Mitchell I. Burken, M.D. 262

Open Committee Discussion 304

Day One Adjournment 330

00004

1 TABLE OF CONTENTS (Continued)

2 VOLUME III

3 Opening Remarks - Introduction 336

4 Open Committee Discussion 337

5 Motions, Discussions and

Recommendations 425

Adjournment 487

00005

1 PANEL PROCEEDINGS

2 (The meeting was called to order at

3 8:00 a.m., Monday, November 15, 1999.)

4 MS. TILLMAN: Good Morning, and

5 welcome. Dr. Ferguson, Dr. Bagley, members and

6 guests, I'm Kate Tillman, Executive Secretary of

7 the Laboratory and Diagnostic Services Panel of

8 the Medicare Coverage Advisory Committee. The

9 committee is here today to provide advice and

10 recommendations to the Agency regarding formal

11 requests pertaining to human tumor assay

12 systems. This is the first meeting of the

13 laboratory panel.

14 We are happy to have such a

15 distinguished panel. Thank you all for coming.

16 Today I would like to welcome Dr. Bryan

17 Loy, carrier medical director, from Administar,

18 who is our guest.

19 We have one member of the panel who has

20 received an appointment to temporary voting

21 status, and that is Dr. Kathy Helzlsouer.

22 We have a couple of pieces of business

23 to take care of here. The appointment to

24 temporary voting status. This is signed by

25 Michael Hash, Deputy Administrator for Health

00006

1 Care Financing Administration. Pursuant to the

2 authority granted under the Medicare Coverage

3 Advisory Committee charter, dated November 24th,

4 1998, I appoint the following person as voting

5 member of the laboratory and diagnostic services

6 panel for the duration of this panel meeting on

7 November 15th and 16th, 1999: Kathy Helzlsouer,

8 M.D. For the record, this individual is a

9 special government employee and is a voting

10 member of the panel under Medicare Coverage

11 Advisory Committee. We have undergone the

12 customary conflict of interest review and have

13 reviewed the material to be considered in this

14 meeting. Signed, Michael M. Hash, Deputy

15 Administrator.

16 The conflict of interest statement:

17 Conflict of interest for the laboratory and

18 diagnostic services panel meeting, November 15th

19 and 16th, 1999. The following announcement

20 addresses conflict of interest issues associated

21 with this meeting and is made part of the record

22 to preclude even the appearance of impropriety.

23 To determine if any conflict existed, the Agency

24 reviewed the submitted agenda and all financial

25 interests reported by the committee

00007

1 participants. The conflict of interest statutes

2 prohibit special government employees from

3 participating in matters that could affect their

4 or their employers' financial interests. The

5 Agency has determined that all members and

6 consultants may participate in the matters before

7 the committee today.

8 With respect to all other participants,

9 we ask in the interest of fairness that all

10 persons making statements or presentations

11 disclose any current or previous financial

12 involvement in any firm whose products or

13 services they may wish to comment on.

14 Now I am going to turn the meeting over

15 to our chairman, Dr. John Ferguson, who will

16 introduce the panel.

17 DR. FERGUSON: Good morning, and

18 welcome to everybody here. I would like to have

19 the panel members introduce themselves, starting

20 from over here on my far left.

21 MS. SIMMERS: I'm Neysa Simmers.

22 DR. FERGUSON: Could you also say where

23 you're from and what you're doing in life?

24 MS. SIMMERS: I am Lisa Simmers. I'm

25 from Bridgewater, Virginia. I am currently a

00008

1 health care administrator, and am here in the

2 interest of the laboratory community, I guess.

3 DR. SUNDWALL: I'm David Sundwall. I'm

4 a physician and I'm president of the American

5 Clinical Laboratory Association, in Washington,

6 D.C.

7 DR. FERGUSON: You have to speak into

8 these microphones like a rock singer, I think.

9 DR. KLEE: I am George Klee. I am from

10 Rochester, Minnesota, and I'm a clinical

11 pathologist.

12 DR. FISCHER: Paul Fischer. I'm a

13 family physician from Augusta, Georgia.

14 DR. BROOKS: John Brooks. I am

15 chairman of pathology and laboratory medicine at

16 Roswell Park Cancer Institute.

17 MR. BARNES: Rod Barnes. I am the

18 industry rep on the panel. I work for AlCon Labs

19 in Fort Worth, Texas.

20 DR. BAGLEY: I'm Grant Bagley. I'm the

21 Federal representative on the panel, and director

22 of coverage in HCFA.

23 DR. FERGUSON: I am John Ferguson. I

24 am a practicing neurologist, and I have just

25 retired from the NIH, where I directed the

00009

1 consensus development program for the last 11

2 years.

3 DR. MURRAY: I'm Robert Murray, a

4 clinical biochemist in practice in Chicago,

5 Illinois.

6 DR. LOY: I'm Bryan Loy. I am with the

7 Kentucky Medicare carrier. I represent the

8 Medicare system at the state carrier level.

9 MS. SNOW: I am Kate Snow. I am the

10 consumer rep on this panel, and I am the director

11 of senior services for Northern Michigan Regional

12 Health Service, and I am an advanced practice

13 nurse in gerontology.

14 DR. KASS: I am Mary Kass. I am

15 chairman of pathology at Washington Hospital

16 Center, and director of integrated laboratory

17 services for MedStar Health.

18 DR. HAUSNER: I am Richard Hausner. I

19 am a pathologist practicing in Houston, Texas.

20 MS. KRAFT: I am Cheryl Kraft,

21 administrative director of laboratory services,

22 Minneapolis.

23 DR. HELZLSOUER: I'm Kathy Helzlsouer,

24 medical oncologist and professor of epidemiology

25 at Johns Hopkins School of Public Health.

00010

1 DR. MINTZ: I am Paul Mintz. I direct

2 the clinical laboratories and blood bank at the

3 University of Virginia Health System, where I'm a

4 professor of pathology and medicine.

5 DR. FERGUSON: I would like to now turn

6 this over to Grant Bagley. Grant?

7 DR. BAGLEY: I'll just make a couple

8 introductory remarks and sort of bring everyone

9 up to speed about what we're doing and how the

10 process works.

11 The coverage process for Medicare is

12 one which from the very inception of the Medicare

13 program has been marked by local diversity and at

14 the same time, the ability to have national

15 conformity when the science and practice so

16 dictates. It has always been that way and it

17 continues to be that way today.

18 What we're about here is considering

19 issues for national coverage decisions. Very

20 much like the federalism model for everything

21 else, states or in this case Medicare carriers,

22 can have variable policies, but when the science,

23 when the issue is sufficiently justified, we can

24 develop a national coverage policy. That

25 national coverage policy then takes precedence,

00011

1 and all Medicare carriers in every state and

2 every area follow that same process.

3 So we are going to talk a little bit

4 about how Medicare coverage works and how we are

5 going to deal with it specifically in this issue.

6 What we're talking about is the Medicare

7 statute, and the Medicare statute has one

8 overarching principle, which is in the terms of a

9 bureaucrat, 1862.A.1(a) of the Social Security

10 Act, and this is what it says: That no payment

11 shall be made under Medicare for a service which

12 is not reasonable and necessary for the diagnosis

13 or treatment of an illness or injury. Those are

14 very important words. Reasonable and necessary,

15 diagnosis or treatment, and a disease or

16 illness.

17 Now, reasonable and necessary has never

18 been defined. We've never defined it explicitly

19 and said, this is what it takes to prove

20 reasonable and necessary. But over the years we

21 have articulated principles by which we say,

22 reasonable and necessary means the following

23 things: It doesn't mean safe and effective. It

24 has to be safe and effective to be reasonable and

25 necessary, to be sure, but it has to be a bit

00012

1 more.

2 So in terms of what we have required to

3 show something is reasonable and necessary, first

4 of all, if it requires a safe and effective

5 determination, and clearance for marketing by the

6 FDA, we've always considered that to be a first

7 step. And second, if it doesn't require

8 clearance for marketing by the FDA, we still make

9 an inquiry that it must be safe and effective.

10 But demonstrated effectiveness is one in which we

11 have said the benefits have to outweigh the

12 anticipated risks. It has to be FDA approved, if

13 required. And there has to be authoritative

14 evidence that it improves outcomes, because after

15 all, that's really what we're talking about.

16 So really, the difference between a

17 threshold issue of is it safe and effective and

18 can it be marketed is somewhat different, you

19 know, and we have to look at it a little bit

20 more. So it has to be safe, to be sure. Any

21 product, even a diagnostic test, has to be safe.

22 It has to be effective, that's clear. But it

23 also has to have benefit which is outweighed, or

24 at least outweighs the risk involved in even the

25 procedure or even a diagnostic test, because it's

00013

1 going to guide therapy.

2 But not only do the benefits have to

3 outweigh the risks, but there have to be some

4 kind of outcomes, there has to be some

5 improvement in clinical care. Is it an improved

6 outcome, does it give better treatment, does it

7 give better results, or in terms of the

8 diagnostic tests, does it give information which

9 can guide or improve therapy.

10 And finally, does it have value? Is

11 there any value to this procedure? For

12 diagnostic tests, it's an issue; for anything

13 else it can be an issue of does it improve

14 therapy, do we get not necessarily better

15 survival, but do we get improved quality of

16 life.

17 Well, how do we determine this? And

18 again, we've articulated these over the years and

19 said, you know, we have to look at clinical

20 studies and from these clinical studies, we have

21 to be able to make determinations. And so we

22 have to be able to look to available evidence and

23 say, are there fundamental safety questions?

24 Does a product live up to its claims? Does it

25 provide the clinical utility that we can use in

00014

1 practice, because after all it has to be,

2 remember the statute, reasonable and necessary.

3 And we look at the outcomes and do the clinical

4 studies to provide evidence that there is an

5 improved value from the service.

6 Of course we can look at it in a number

7 of ways. We can look at outcome measures in

8 terms of simply survival, that certainly is the

9 crudest measure we can use for an outcome. But

10 we can look at process changes, which may be

11 indirect, and we can say, how does it influence

12 the disease process, and can we make inferences

13 about value from that. And we can observe just

14 simply effects in terms of does it change a

15 measured process, does it change a physiologic

16 process. Is blood pressure improved? Do we have

17 a metabolic process change? Is cholesterol

18 lowered? And then can we relate those to an

19 outcome.

20 So even when we look at secondary end

21 points, when we are looking at a physiologic

22 measurement or metabolic change, can we make the

23 direct link to an improved outcome. And I think

24 it's going to be important to keep that in mind

25 as we look at intermediate end points.

00015

1 And in terms of looking at the science,

2 this has always been what we've used to determine

3 what's reasonable and necessary. You know, if we

4 look at information, we look at collections of

5 data, and we look at studies, is to keep in mind

6 that we have to consider the bias that can be

7 introduced, are patients selected in less than a

8 random fashion so that the outcomes might be, you

9 know, influenced by the way patients are

10 selected. Do we select patients for one group

11 and then do they become evaluated by another

12 method in terms of trials with more than one

13 arm. Do patients disappear after being entered

14 into the study, and if so, for what reason? Is

15 this going to affect it? Do we have some way of

16 having people evaluate the results of the study

17 without knowing what the outcomes should be or

18 are going to be? And is there an adequate way to

19 control for the information?

20 These are all things to keep in mind

21 when we're considering clinical data and clinical

22 study. Are they big enough? Are we measuring

23 something which is large enough that we can make

24 a determination? Whatever we've measure, have we

25 measured enough to say this is truly an effect?

00016

1 Do we have enough subjects in here? Do we have

2 enough patients to make that determination?

3 And within Medicare, we always consider

4 what we're dealing with. I mean certainly, some

5 diseases have a very high prevalence, they have a

6 large impact on the Medicare population, and in

7 those situations we need to have a great deal of

8 evidence to make a change. On the other hand,

9 some diseases are not highly prevalent, they deal

10 with just a smaller population of people, and in

11 those cases we have to look at the degree of

12 precision in the clinical studies in a somewhat

13 different way.

14 To consider the natural history of

15 diseases and the issues we're talking about

16 today, we have to consider what the uninfluenced

17 outcome would be in terms of what kind of a

18 difference does it make when we start to alter

19 things. And in looking at clinical studies, we

20 have to consider both the issue as presented and

21 we have to look at the source they came from.

22 I think I was on a panel with some

23 folks from Australia where they do a much -- they

24 have a much different process than we do in terms

25 of deciding their coverage in terms of looking at

00017

1 not only the science, but once they've looked at

2 the science, they then look second, and they say

3 now that we've decided we're going to cover

4 something based on the science, let's decide if

5 it's worth it, let's look at what it costs and

6 make that determination. And in doing that we

7 made an interesting point, which I think is

8 worthwhile to relate here, and that is that if

9 you're going to look at a survey, you know, if

10 you're out to buy a car and you're looking at a

11 survey, and you want to look at the report of all

12 the new options that are available in new cars,

13 that you're probably going to say it makes a

14 difference to me whether this is from an

15 independent consumer agency or whether this is a

16 report produced by the auto manufacturer. And

17 the same thing ought to be true when we look at

18 studies, when we look at clinical information.

19 We need to look at the source and say, not

20 necessarily that there is bias introduced, but we

21 need to look at full disclosure of the source of

22 information. We need to look at whether there is

23 real bias or whether there's apparent bias, and

24 that it's not really there, so we need to look at

25 the source of information.

00018

1 We also need to look at the credibility

2 of information. There is a hierarchy of evidence

3 which we need to consider, and that published

4 peer review literature is considered evaluated by

5 the larger community. It's considered to be of a

6 higher validity and we need to consider that. We

7 aren't always going to have that and we will have

8 to look at things which are not peer reviewed but

9 have been presented and published without peer

10 review. We're going to look at things that are

11 not even published in full form, so-called

12 abstract form, where they have been presented at

13 a meeting and an abstract is simply printed, and

14 we are going to look at unpublished data which

15 has been subjected to considerably less scrutiny.

16 So those are the kinds of things we're

17 going to look at in our evaluation of evidence

18 presentations today. You need to look at what

19 the information is showing you. You need to look

20 at what's happening, where it came from, and then

21 evaluate it based on its quality, the hierarchy

22 of the evidence and the source.

23 Now the process we're using, and it's

24 the process Medicare is now using for national

25 decisions, is very different from in the past.

00019

1 This is a part of that, this open public

2 committee meeting, and we put together an entire

3 new process which is now open. It's open because

4 this is a public meeting, and we're having a full

5 and frank public discussion on an issue. It's

6 quite defined in terms of its process. We bring

7 together technology issues and you'll find that

8 we aren't bringing together specific products or

9 specific tests, but we are bringing an area of

10 technology together and we're looking at a whole

11 area of technology, because we don't cover

12 products, we cover areas of technology. It's one

13 in which there's going to be full public

14 participation here, and we're going to make an

15 explicit decision based on the recommendations of

16 the panel. And I want to make it very clear:

17 This panel does not make coverage decisions.

18 This panel is for the purpose of giving us

19 technical advice so that HCFA itself can then

20 make an explicit decision. And not only will we

21 make a decision in a fair and proper fashion

22 after final panel recommendations, but that that

23 decision is subject to challenge, if we

24 significantly misinterpret the evidence or if we

25 fail to consider all the evidence. Of course

00020

1 this is a public process, so we expect all the

2 evidence to be here.

3 So in looking at the clinical trials

4 we're going to look at, we want you to focus on

5 looking for definitive answers, clinical utility,

6 does it improve outcome, is it appropriate, and

7 can we determine for which patients it should be

8 appropriate, so we can administer the Medicare

9 benefit. Is the source of the information

10 unbiased, is it free of conflict, and can we make

11 noncontroversial decisions? Because remember,

12 when I described the process, the very last step

13 of this whole process is subject to challenge

14 when we make a decision. And by this open

15 committee, this advisory process in which there

16 is full participation with the public, we would

17 expect to address all of these issues so that

18 there would be little basis for challenging any

19 decision.

20 Now, getting to the issue at hand, of

21 looking at sensitivity and resistance tests in

22 terms of oncology, just a little bit of history.

23 Medicare has looked at these in the past. We've

24 looked at them for quite a while. As long ago as

25 1980, the technology was around, we looked at

00021

1 these technologies, and at that time what HCFA

2 used to get internal advice was the physician

3 panel. The physician panel was a group of

4 physicians who worked for the Health Care

5 Financing Administration, and also physicians who

6 were in the Public Health Service, with other

7 agencies, that were brought together to look at

8 scientific issues and give the Agency scientific

9 advice. It was an internal predeliverative

10 advisory panel, something which I should say

11 also, is perfectly acceptable, even under the

12 Advisory Committee Act, because it predelivered

13 consideration by internal government employees.

14 This physician panel got together, looked at the

15 issue and requested a technology assessment at

16 that time. And that has been a number of years

17 ago.

18 It was then, as late as 1987, that same

19 physician panel met again. Based on the

20 technology assessment that was available at that

21 time that they reviewed, and they considered the

22 issue and felt that the use of tumor cells for

23 sensitivity determinations was still experimental

24 and that there was not enough information to

25 provide coverage at that time.

00022

1 In 1991, the issues were looked at

2 again, and at that time we were using, it was the

3 physician panel but it was now called the

4 technology advisory group, still composed of

5 internal government physicians. They were from a

6 little bit wider scope in terms of bringing folks

7 from the FDA, from AHCPR, from NIH, and they

8 discussed this at the same time, and they

9 discussed the issues around assays at the same

10 time, and agreed that the existing language which

11 we put in the coverage issues manual, which said

12 that this technology was that this technology was

13 at that time experimental, should be retained.

14 It was then in '97 that the technology

15 advisory committee, which was an outgrowth of the

16 same internal deliberative body, looked

17 specifically at extreme drug resistance testing

18 and considered whether or not extreme drug

19 resistance testing was in fact the same kind of

20 technology that was looked at before in terms of

21 sensitivity. Was it really the same thing, was

22 the technology the same, and was the utility the

23 same. And at that time the technology advisory

24 committee came to the conclusion that perhaps

25 extreme drug resistance testing was enough

00023

1 different of a methodology from sensitivity

2 testing, and that the clinical utility was enough

3 different that the coverage issues manual

4 exclusion of this technology, saying it was

5 experimental, that was put in over ten years

6 previously, perhaps didn't apply and that drug

7 resistance testing was enough of a different

8 technology that it could be left to carriers to

9 have discretion to cover that technology.

10 The current policy, then, is that human

11 tumor drug sensitivity assays are considered

12 experimental and therefore, not covered under

13 Medicare. That is a statement which leaves no

14 discretion for Medicare contractors, and that

15 statement is in force today.

16 Now it's interesting in that we made

17 the interpretation through the technology

18 advisory committee that drug resistance testing

19 was enough different from sensitivity assays that

20 it was not covered by this prohibition, and there

21 has continued to be confusion over that issue

22 over the past, you know, several years since this

23 was done. So that's what's currently in place.

24 In order to reevaluate our position in

25 that coverage issues manual statement, we have

00024

1 convened this panel and we are going to present

2 the following questions, and I will quickly go

3 over them, because these questions are going to

4 be focused on tomorrow and are going to be ones

5 that we are going to ask the panel to answer.

6 First, is the scientific evidence that

7 is amassed thus far, presented to the panel and

8 that's going to be discussed here today and

9 tomorrow morning sufficient that we can make the

10 appropriateness determinations about a coverage

11 utility and about what should happen in terms of

12 clinical care in using these tests?

13 Are the assay techniques described in

14 the literature for single drugs sufficiently

15 transportable to multidrug therapy, and there's

16 going to be a question presented about the

17 appropriateness of using single drug information

18 in terms of testing in multidrug regimens in

19 terms of treatment.

20 Does the scientific evidence

21 demonstrate the clinical benefit? This can be

22 very important, because there's going to be

23 information presented about different kinds of

24 tumors, hematologic tumors, solid tumors, and

25 being somewhat different in character, and should

00025

1 we be able to make determinations of clinical

2 care based on testing which are going to guide

3 therapy, because after all, remember, we talked

4 about clinical utility and value as being very

5 important things that we can draw from these

6 tests. So if we can't make the clinical utility

7 argument or if the value isn't there to be able

8 to directly influence therapy, then that's going

9 to be important to consider in terms of is it

10 reasonable and necessary.

11 If test results in terms of sensitivity

12 or resistance give us predictions about a tumor

13 response, should in fact those predictions guide

14 what happens in terms of direct therapy? Because

15 after all, one of the things we need to know,

16 which as I stated, we need to know not only

17 clinical utility but appropriateness, and we need

18 to be able to say when is it appropriate to do

19 these tests and what is the appropriate clinical

20 action once the test is done, because that's the

21 kind of information we need to be able to put

22 together a coverage policy.

23 Is there sufficient scientific evidence

24 to demonstrate the clinical utility in selecting

25 appropriate chemotherapy?

00026

1 And finally, the committee will be

2 given the opportunity to raise any additional

3 concerns.

4 So basically, these are the questions

5 that we're going to present to the panel. What

6 should we be looking at in terms of measuring

7 this technology? Should we be looking at

8 survival or should we be looking at intermediate

9 responses? Are there appropriate measures that

10 we can look at in terms of response to the tumor,

11 in terms of quality of life, in terms of other

12 intermediate outcomes, or should we be looking at

13 survival, and is one an appropriate surrogate

14 measure for the other. Is information on single

15 versus combination drug regimens relevant in

16 terms of using the results of the test in

17 clinical care?

18 Does the evidence that's presented that

19 we consider here over the next day and a half

20 demonstrate that there is in fact a clinical

21 benefit, not just interesting information, but is

22 there a clinical benefit which we can derive from

23 the use of this methodology? And if there is a

24 clinical benefit we can derive from this

25 methodology, should it in fact determine what the

00027

1 treatment should be in terms of particular

2 patient care.

3 And then finally, are there additional

4 concerns that the panel after a day and a half of

5 considering this technology wishes to bring to

6 the forefront?

7 So those are the issues we're going to

8 be talking about and that's a general overview of

9 the process HCFA uses and the kinds of

10 information and the level and hierarchy of

11 evidence that we wish to have considered over the

12 next few days, or day and a half. We're going to

13 present those questions tomorrow. There will be

14 a discussion of those questions, and we will be

15 asking the panel to vote specifically on those

16 answers and give us determinations which we can

17 then use in the form of recommendations to either

18 clarify, to ratify or to change the existing

19 policy which we have, which is noncoverage of

20 human tumor assay systems, and I think a somewhat

21 confused approach to drug resistance testing in

22 terms of that coverage issues manual

23 application.

24 So that's the charge to the committee.

25 There will be public presentations. There will

00028

1 be presentations from HCFA and from other

2 sources. I think we will all hear different

3 interpretations of information. There will be

4 full discussion, and we look forward to this

5 process giving us recommendations which we can

6 then use to modify or ratify our existing

7 policy.

8 MS. TILLMAN: Now Dr. Ferguson has a

9 few remarks to make.

10 DR. FERGUSON: Thank you. It's the job

11 of this panel to review the evidence and its

12 quality for this group of in vitro drug assays,

13 and arrive at some conclusions regarding the

14 appropriateness of these tests in treating cancer

15 patients. In an ideal world, the evidence would

16 dictate yes or no. Unfortunately in the real

17 world, we are likely to find something less,

18 certain conditions for select patients,

19 et cetera. Asking the research community to

20 consider patient outcomes in evaluating new

21 diagnostic tests seems to be setting the bar

22 higher for the quality of evidence than

23 previously. However, I believe that all of us

24 want the best possible outcomes for all the

25 patients we see, no matter where we sit.

00029

1 As we spend proportionally more money

2 on health care, we should try to achieve the best

3 possible outcomes for our patients, and this may

4 require setting the quality of evidence bar

5 higher than 10 or 20 years ago. The job of this

6 panel is to evaluate the data we are presented

7 with for these assays and to use this evidence to

8 answer the questions that HCFA has posed. It's a

9 bit of a conundrum for society, I think, only

10 paying for what works and yet not stifling

11 innovation in the process. Our job, this panel's

12 job is not easy, and we recognize that the

13 presenters don't have an easy task either,

14 especially given the short time to present their

15 work which has occurred over a number of years.

16 In the interest of time, I would like

17 to try to encourage all of the presenters to

18 stick as closely as possible to the outlines we

19 have, and I would like to get started with our

20 FDA. Kate, do you want to introduce Dr. Harvey?

21 MS. TILLMAN: Sure. Our first speaker

22 is Dr. Brian Harvey, who is the associate

23 director of the Division of Clinical Laboratory

24 Devices for the Food and Drug Administration.

25 Dr. Harvey?

00030

1 DR. HARVEY: Good morning. First of

2 all, I would like to thank the Health Care

3 Finance Administration for the invitation for us

4 to speak this morning, and I would like to

5 commend HCFA for moving towards an advisory panel

6 process, and we are glad at FDA to be a

7 participant in that process. What I would like

8 to do this morning, I am Brian Harvey, a senior

9 medical officer at Center for Devices and Office

10 of Device Evaluation, and currently acting

11 associate division director in clinical labs.

12 And what I wanted to do this morning is actually

13 talk about the FDA process.

14 Often when we hear about HCFA's role in

15 the evaluation of new technologies, we hear well,

16 if the advice is FDA approved then it can go on

17 to the HCFA process. And what I -- the major

18 points I really want to get across today is that

19 there are many roads to the U.S. market that

20 medical devices can go through. One size does

21 not fit all. And by actually going over the

22 various methods that medical devices can get to

23 the U.S. market, give a better understanding of

24 sort of the terms approve versus clear, exempt,

25 et cetera.

00031

1 As most of you know, the regulation of

2 medical devices in the United States really

3 didn't start until May 28th, 1976. There were

4 some medical devices that were regulated under

5 the drug law before that time, but the vast

6 majority of medical devices began to be regulated

7 with the medical device amendments to the Pure

8 Food and Drug Act, May 28th, 1976. The law

9 itself was sort of a hodgepodge of many different

10 concerns, which sort of reflect the great variety

11 which are medical devices, and as we go through

12 some of the aspects of the law, you'll see how

13 the fact that the majority of devices were not

14 regulated has really fed into the whole construct

15 of medical device regulation. I will touch upon

16 the Safe Medical Device Act in 1990 as well as

17 the more recent FDA Modernization Act of 1997,

18 which we all call FDAMA.

19 So once again, medical device

20 amendments, 1976, it was the outline which we

21 still use today stratifying medical devices in

22 classes. It's a risk based classification, class

23 one being those devices which are very low risk,

24 class two devices an intermediate or moderate

25 risk, and class three devices being the highest

00032

1 risk devices. There is actually a very long

2 definition of what a medical device is, I'm not a

3 lawyer, but I won't even spend the time reading

4 that. It's a full page long and in the interest

5 of time, the point being it is defined in law as

6 well as defined in terms such as safe and

7 effective.

8 The vast majority of medical devices in

9 the U.S. do go through something that's called a

10 510(k), which I will explain in a minute. That

11 aspect of the law was strengthened with the 1990

12 SMDA law. It required an indication for use

13 statement, so therefore in a specific part of the

14 application, the specific indication for use for

15 which the company, the sponsor wishes to get FDA

16 clearance was clearly stated. There was a

17 summary of safety and effectiveness in each

18 application, and FDA through the Freedom of

19 Information Act, was able to make that available

20 to the public to give an insight into what led to

21 different medical device decisions.

22 With the FDA Modernization Act, there

23 were actually several other aspects that were

24 clarified. The Center for Devices, a few years

25 before this act, actually instituted a

00033

1 reengineering effort and many of the aspects of

2 the FDA Modernization Act became codified in the

3 law through the FDA Modernization Act, trying to

4 increase the emphasis on post-market evaluation

5 of devices, but keeping an adequate premarket

6 evaluation, the whole concept of interactive

7 reviews, trying to increase communication between

8 the industry and the FDA. Greater inclusion, not

9 only in the public advisory panel but internal

10 meetings. Greater outreach to academic

11 societies. And there is also a section of the

12 FDA Modernization Act, Section 205, which many of

13 you have been hearing about in the news, which is

14 the least burdensome method to get to the U.S.

15 market.

16 And actually, I recommend that you all

17 go to the FDA website, which I will give later

18 on, to look at the draft guidance document on

19 least burdensome, because we still are in a

20 public comment period and we welcome your

21 comments. One of the things you will note is

22 that the current document says it does not apply

23 to IVDs, in vitro diagnostic devices, and one of

24 the clear aspects that we are getting in the

25 public comments is the importance of including

00034

1 IVDs in the process. And as part of my efforts

2 in the clinical laboratory division is to

3 incorporate a section for in vitro diagnostics in

4 the least burdensome framework since it's a very

5 important aspect of medical devices.

6 So what are the different roads to the

7 U.S. market? Well, starting in the beginning,

8 under the IDE, or investigational device

9 exemption, if something is considered a

10 significant risk through the local IRB, it then

11 comes to FDA for a review of the protocol. In

12 1995 an agreement was worked out between Health

13 Care Finance Administration and the FDA to try to

14 designate what was a truly experimental

15 investigational device and what was a more run of

16 the mill or traditional device that just was a

17 newer version. One of the aspects of medical

18 devices for those who are involved with drugs,

19 sometimes catches people off guard, is how

20 medical devices really are just sort of a

21 technology creep.

22 And let's say in pacemakers, the older

23 version through the newer version, very, very

24 minor changes require a new application. So you

25 may have a device that's very similar to the

00035

1 traditional device that's now in an

2 investigational device exemption study, and it

3 might not have gotten covered because due to a

4 new bell or whistle, it was not yet on the

5 market. So through the wisdom of an agreement

6 between HCFA and FDA, the decision was made,

7 there really should be a designation where FDA

8 says this is a way out very experimental device,

9 or this really is just a minor modification, in

10 order to meet the FDA requirements they are going

11 through an investigational device exemption.

12 But our recommendation is based on our

13 evaluation and it is a nonbinding recommendation,

14 that this should be considered for HCFA

15 reimbursement. So the A versus B designation, B

16 being that FDA feels that it should be considered

17 for reimbursement, and 80 to 90 percent of IDEs

18 actually have that B designation. So if

19 something is an established IDE, it gets this B

20 designation as something that could be considered

21 for reimbursement. So, the original idea in this

22 1976 law, the whole concept being is that there

23 was a number of medical devices that were on the

24 market; through the use in the market, they were

25 found to be safe and effective, and if that

00036

1 device was on the market before May 28th, 1976,

2 and a sponsor or company could come in and show

3 that their new device was substantially

4 equivalent to that old device, they could submit

5 a premarket notification, designate it 510(k)

6 based upon the law, that line of the law, and

7 they were able to get to market.

8 So they did not to establish de novo

9 safety and effectiveness, but through a

10 substantial equivalence flow chart, they are able

11 to show by direct comparisons, both clinically,

12 engineering, bench testing, et cetera, that these

13 devices are substantially equivalent. And what

14 we have actually found in a very positive way is

15 that there has been a technology creep and

16 improving of devices, although the older devices,

17 the newer devices were found to be substantially

18 equivalent to the older devices, when you

19 actually look over time, there is a gradual

20 improvement.

21 So it's been a way for the companies to

22 innovate. It is actually in the spirit of least

23 burdensome, long before that provision was

24 written, a way to get to market in a smaller

25 package not requiring advisory panel review, but

00037

1 with internal review for these devices to get to

2 market. And as part of the broader market,

3 traditionally class one 510(k)s were very small;

4 class two 510(k)s, depending on the type of

5 device, were either smaller or larger, depending

6 on whether or not there was needs for clinical

7 data. And then there were some class three

8 devices that were deemed to be 510(k)s. As part

9 of the more recent reengineering efforts and

10 recent laws, those have actually either been

11 converted to class three PMAs, which I'll talk

12 about in a minute, or have been down classified

13 to class two.

14 So in the broad scheme of things, the

15 way to think of it is, the vast majority of

16 devices on the U.S. market are class two 510(k)s,

17 and if you look at the numbers from fiscal year

18 1998, there are about 4600 class two 510(k)s

19 cleared for market, and the term is cleared for

20 510(k)s, versus approved for PMAs. 4600 510(k)s

21 compared to about 50 PMA applications that were

22 approved, and about 250 to 300 PMA supplements.

23 So you can see, the vast majority of medical

24 devices in the United States have actually been

25 cleared through the 510(k) process.

00038

1 So the PMA is the premarket approval

2 application. That's the one that people

3 traditionally think about when they think of an

4 FDA approval. It's based on valid scientific

5 evidence. Often an original PMA has to go to the

6 public advisory panel for their recommendation.

7 The valid scientific evidence is actually defined

8 in the law as well controlled investigation,

9 partially controlled studies, studies without

10 matched controls, well documented case histories,

11 reports of significant human experience. So you

12 can see some parallels between the FDA law and

13 the HCFA law that Dr. Bagley alluded to earlier.

14 So you can see, it's the whole gamut of

15 different sorts of both clinical and evidence.

16 Now in the original 1976 amendments there was

17 another route for class three devices to come to

18 market, and that was the PDP, or product

19 development protocol. And the thought was that

20 if there was something that required a clinical

21 trial, that the companies may want to have public

22 advisory input long before they got to the final

23 presentation that normally happens in the PMAs.

24 So what happens is that a company submits a PDP

25 protocol, in the PDP it must have the animal

00039

1 testing, the bench testing, as well as a proposal

2 for a clinical trial. That is then reviewed by

3 the FDA and taken to a closed session of an

4 advisory panel, since it's still proprietary

5 information, confidential information. The

6 advisory panel comments on the protocol design

7 and has input into that. As part of the

8 protocol, there are actually set end points that

9 have been designated for success criteria. So

10 obviously, it's to be used with those medical

11 devices that are well know as far as what to look

12 for as far as a success criteria. Then if it's

13 deemed approved by the panel and the FDA, the

14 sponsor or the company goes out and does the

15 protocol, and if they actually meet those success

16 criteria, the PDP is deemed approved and you do

17 not need to go back to the advisory panel for a

18 final approval. So once again, another path to

19 market, it's not a PMA, not a 510(k), but it's

20 equivalent to a PMA for a class three device.

21 Another aspect, another way to market

22 is the HUD or humanitarian use device. This is

23 the, HUD is analogous to the orphan drug part of

24 the drug law and actually, the initial

25 application to FDA for designation goes through

00040

1 orphan drugs at FDA in the Center for Drugs. And

2 the concept is that if there is a disease which

3 affects less than 4,000 people per year in the

4 United States and is not being adequately being

5 treated by any current medical device, then a

6 company can come in, and if they have been given

7 that designation by the orphan drug people at

8 FDA, then they can submit an HDE, as opposed to a

9 PMA or a PDP, for their class three device.

10 Safety definition in the law and in practicality

11 is the same as a PMA or PDP. However, instead of

12 establishing effectiveness, they only have to

13 show probable benefit. And the concept being, is

14 that there are fewer patients to study, the

15 benefit to this patient group far outweighs the

16 risks based upon the safety analysis, and the

17 review time is shorter, and these devices are

18 able to get out to the public.

19 So one of the things that the HCFA

20 advisory panel may be asked to comment on, not

21 only this panel but all of the panels, is this

22 whole area of HDE. So it is an FDA approval just

23 like a PMA or PDP for a class three device, but

24 the criteria are different. And once again, the

25 point being that there are these many ways to get

00041

1 to the U.S. market. And perhaps the best way to

2 describe it is not so much an FDA approval, but

3 has a certain medical device met the FDA

4 threshold?

5 So it gets us into specifically in

6 vitro diagnostics and there are sections of the

7 law and the regulations that deal specifically

8 with in vitro diagnostics, and I can see I'm

9 running late on time. Many of you are familiar

10 with all these labeling requirements, the whole

11 concept of reagents and instruments, how these

12 are all integral parts of in vitro diagnostics.

13 Laboratory tests, if something is done at a

14 specific laboratory, it's not exported anywhere

15 else, it's sort of the concept of a home brew,

16 the FDA has chosen not to regulate at this time

17 home brew assays. These are considered class one

18 exempt medical devices. Now if you had a home

19 brew which then was being exported, then there

20 may be parts of that which would be subject to

21 some of the various aspects of FDA regulation.

22 Just on a side note, the whole CLIA

23 effort, which is currently being run by CDC, the

24 decision has been made to transfer that to the

25 FDA, so this will therefore be the same

00042

1 regulation, and the FDA will also, for in vitro

2 devices, will also be doing a parallel clear

3 review. At this time there are no planned

4 changes in the criteria that CDC has been using,

5 but you will be hearing more of a clarification

6 on that aspect of the law.

7 So now the issue that was in the news

8 this past week, the analyte specific reagent, it

9 was an area that was sort of an internal SOP, a

10 standard operating procedure in the in vitro

11 diagnostic group, but the rule was formalized on

12 November 24th, 1998. The concept being is,

13 although you have these home brew assays, they

14 are only being done at one site, you wanted to

15 make sure that the various components of those

16 home brews met certain FDA criteria, the whole

17 concept being is that if you had an analyte

18 specific reagent, you wanted to make sure it met

19 certain good manufacturing levels. And as you

20 can see in the actual regulation, they talk about

21 antibodies, specific receptor proteins, nucleic

22 acid sequences. So you see a heavy emphasis here

23 on biological agents which for other, in other

24 contexts may actually be regulated at FDA through

25 the Center for Biologics. And there are various

00043

1 impacts on manufacturers on labeling through the

2 analyte specific reagent.

3 So to get to today's issue, initially

4 FDA was sent a letter from HCFA, and the inquiry

5 was, are there any medical devices that had been

6 FDA approved that fell under the scope of the

7 types of medical devices that we're going to be

8 discussing at today's meeting. And at a branch

9 level, the reviewers who were involved in this

10 area went through the database, and their initial

11 review was that there was nothing in the FDA

12 database. Because of that review at the branch

13 level, a letter was issued, from which many of

14 you have seen the letter, which actually

15 generated quite an industry response and actually

16 is part of the whole concept of least burdensome

17 and FDAMA and interactive process, this has

18 actually turned out to be a good thing.

19 From my point of view actually, this is

20 when I was brought into the process. I was not

21 directly involved with that initial review. But

22 what we did, based upon the overwhelming input

23 from the industry, it triggered an internal FDA

24 review of the issue, and it actually went up to

25 the level of the new center director, Dr. David

00044

1 Feigel, who took over for Bruce Burlington, and

2 it actually turned out to be a good thing,

3 because Dr. Feigel previously was at the Center

4 for Biologics, was very familiar with the various

5 aspects of what is covered under the analyte

6 specific reagent through his work in biologics,

7 and before that he was a division director in the

8 Center for Drugs. So we were very, very lucky to

9 have sort of a broad perspective of Dr. David

10 Feigel.

11 In addition, Linda Kahn, who was one of

12 her deputies at the center level, was involved,

13 and she was a lawyer by training, had spent a lot

14 of time up in chief counsel's office at FDA, and

15 her input in reviewing the actual regulation and

16 the spirit of the regulation came into play.

17 And I was also actively involved, and

18 my role was, I am board certified in internal

19 medicine, I still practice on evenings and

20 weekends, and before that I was a research

21 biochemist, so I sort of brought both a practical

22 clinical approach to the problem as well as a

23 traditional Ph.D. biochemistry approach. And

24 that with Dr. Gutman, who is the division

25 director, in looking at what was the spirit of

00045

1 the analyte specific reagent statute, and the

2 spirit was that although with home brews, they

3 are used at one site, we want to make sure that

4 good manufacturing practices have been used for

5 all the various components.

6 And in those home brews that use FDA

7 approved drugs, that really is not an issue. If

8 there is a drug which is a chemotherapeutic agent

9 that has been through the FDA approval process,

10 although not at the Center of Devices but the

11 Center for Drugs, they have met all the strict

12 criteria in manufacturing that are really

13 necessary. So really, when you look at the

14 spirit of the regulation, anything that contains

15 an FDA approved drug really does meet that spirit

16 of that.

17 So the official -- the follow-up letter

18 dated November 9th, did go through to say that

19 based upon further evaluation, the FDA not

20 believes that the drugs being used in these

21 assays fall outside the scope of the analyte

22 specific reagent rule and because these products

23 have been approved and are regulated by the

24 Center for Drug Evaluation and Research, there is

25 assurance that they have been produced in

00046

1 compliance with good manufacturing practices. We

2 have concluded that in-house home brew assays

3 prepared using these reagents do not need to meet

4 the requirement of the rule. And then it goes on

5 to say, however, we recommend that certain

6 labeling requirements be considered when these

7 are done.

8 Now as a caveat to that though,

9 however, if these are ever used in kit form, or

10 that kit could be sold and exported to may

11 different laboratories, then they may fall inside

12 the scope of a class three PMA or PDP, or

13 depending on the claims, a 510(k). But if it's

14 at a specific site and falls under the home brew

15 concept, then that's not something that requires

16 FDA direct review. So that's -- I just wanted to

17 go into those details.

18 So to summarize, and to get additional

19 information on all the different areas I talked

20 about today, there is a group called the small

21 manufacturers assistance, and you can be a large

22 or a small, you don't have to be small by

23 definition. They are a group of people who have

24 access to all sorts of information at FDA, and

25 now with the worldwide web, there are various

00047

1 parts of the FDA web site for the Center for

2 Devices that have guidance documents in all the

3 various aspects of which we talked about today.

4 We encourage you to go to that. If you have

5 specific questions, you can talk to the small

6 manufacturing people, and you're certainly always

7 welcome to call us at the clinical labs.

8 But to summarize, I think the best way

9 to consider the FDA process is that there are

10 various ways for devices to get to market. Just

11 to review, there's class one exempt, so

12 therefore, we never see them, but when we say

13 exempt from 510(k), we don't mean exempt from

14 good manufacturing processes. There are those

15 class one devices that have been reserved, and

16 they still do have to come to the FDA. Class two

17 510(k)s, we spent time talking about. And

18 finally, for class three devices, PMAs, PDPs,

19 HDEs.

20 So therefore, perhaps the best way to

21 talk about it is has the FDA threshold been met?

22 And then it's ultimately up to you all to look at

23 the evidence from there. Thank you again for

24 your invitation.

25 DR. FERGUSON: Thanks, Dr. Harvey. I'd

00048

1 like to go right ahead now with Mr. Kiesner, from

2 Oncotech. Are you ready?

3 MR. KIESNER: Yeah.

4 DR. FERGUSON: Since we are 15 minutes

5 later than the schedule says, I'm just going to

6 put things 15 minutes ahead, and take it out of

7 the lunch period at this point.

8 MR. KIESNER: Thank you very much. My

9 name is Frank Kiesner. I am president and CEO of

10 Oncotech, one of the companies that are in this

11 industry. I am here today to give more of an

12 overview of the industry and set the stage for

13 subsequent discussions which will focus on the

14 clinical utility and the clinical application of

15 these technologies.

16 Before I begin, I think it's important

17 to recognize that we are sharing in an historic

18 moment here. That this type of panel, this type

19 of open discussion of medical and patient issues

20 is just starting and that from the outside, we in

21 the industry have been able to witness the

22 gestation of this process, and I can honestly see

23 that what we are involved with today is a major

24 step forward and I think that the coverage and

25 analysis group should take credit for that.

00049

1 As Dr. Bagley was talking, I recalled a

2 town hall meeting that I attended about a year

3 and a half ago, where Dick Coyne and Dr. Bagley

4 proposed some structures. There were about 600

5 of us in the audience, and based on that meeting

6 if there is one thing I am absolutely certain of,

7 is the HCFA staff went through the legal,

8 political, the administrative issues relating to

9 this process. They definitely were not short of

10 free advice.

11 Secondly, I would like to comment just

12 briefly about the FDA issue. And you have all

13 read the letters going back and forth. We are

14 very pleased that this issue was resolved, and

15 with Oncotech, we live in a glass house. By that

16 I mean we every day have to deal with our own

17 issues and our own problems, and I would only

18 hope as we deal with these, that we have

19 ourselves the same sense of urgency, the same

20 decisiveness, and the same unfiltered honesty

21 that we have witnessed within the FDA over the

22 last three weeks, and I think it's a real credit

23 to their organization and to their management.

24 We are very pleased that the issue was resolved.

25 I have tremendous respect for the

00050

1 people that participate in this industry. They

2 are motivated by doing what is good for cancer

3 patients. I am going to share some numbers in

4 relation to the industry to try to get things

5 into a setting. The problem is while everybody

6 is willing to contribute their numbers to

7 industry numbers, there are antitrust issues and

8 problems with duplication of numbers, so what

9 we've chosen to do is just look at the Oncotech

10 numbers, but recognize that the work of others in

11 the industry would probably increase the numbers

12 I am going to show about 25 or 30 percent.

13 In terms of drug resistance testing

14 over the last several years, over 55,000 cancer

15 patients have been tested. If you look just

16 during the last year, or year and a half, we have

17 received tissue samples for testing from over a

18 thousand hospitals throughout the United States,

19 we have reported results to over 2600 physicians,

20 and we have tested 60 different tumor types. The

21 technology is being used in the medical

22 community.

23 Where are we in terms of payor

24 acceptance? The story really began in 1994 when

25 Blue Shield of California had a panel meeting

00051

1 just very similar to this, open discussion,

2 presentations from those in the industry, and a

3 good solid dialog of the science. What they

4 concluded in 1994 was that drug resistance

5 testing in oncology is accurate and reliable and

6 there is sufficient data to determine their

7 safety, clinical utility and impact on clinical

8 decision making.

9 Where have we gone from there? If you

10 look at current payor acceptance, in terms of

11 payor contracts, we have with different managed

12 care entities, 31 million lives under contract as

13 far as payment for drug resistance testing. We

14 have a contract relating to the pricing that

15 involves 2300 hospitals around the country. In

16 terms of not the contract, but in terms of what

17 our payment experience has been for this type of

18 service, in terms of non-Medicare carriers, the

19 managed care and the third party or the

20 indemnities, in the last year and a half, we have

21 probably billed about 17 to 1800 different

22 entities. 1600 of those have paid for the EDR.

23 And I don't want to imply that they've paid

24 everything that we've billed, but they have paid

25 for, they have paid some amount for EDR.

00052

1 The second thing is that in relation to

2 the question of medical necessity or

3 investigational denials, less than one percent

4 have been written off for this reason. Now what

5 I mean by written off is very important, and it's

6 not that questions haven't been raised. At any

7 given point in time, our finance department would

8 be dealing with 25 to 50 different carriers, and

9 we would have to deal with the question of

10 medical necessity or investigational status.

11 What that number indicates is that after we go

12 through that process, that less than one percent

13 are actually written off on that basis. So I

14 want to be very clear on that.

15 How does that contrast with the

16 Medicare experience? Basically, all of our

17 claims have been denied on a local coverage

18 basis, as the technology being investigational.

19 But that's the purpose of this meeting; we are

20 looking at developing a national policy that will

21 be able to integrate all of the information that

22 is current, into a rational approach to this

23 group of technology.

24 Dr. Bagley alluded to the carrier

25 issues manual, the national coverage policy,

00053

1 5041. It was originally enacted in relation to a

2 human tumor stem cell assay. It was 1970s

3 technology. There were technical problems with

4 it. Basically, it was used only in a research

5 setting and for the last 15 years has not been

6 used clinically. It was a very important

7 technology though, and it was important because

8 it highlighted some of the issues involved with

9 the testing of cancer on an in vitro basis. And

10 it was a major step forward because the people

11 that were involved in that technology learned

12 from it and went into a second generation

13 technology, and ultimately to the technology

14 we're using today, which is a third generation

15 technology. So the point is that there has been

16 an evolution in technology, there's been a

17 learning process and a growth, and I would just

18 urge that we look at what is available today and

19 what is the science to support what's available

20 today. It's not something that was in existence

21 in 1982.

22 A recommendation is that this provision

23 5041 should be removed. It was the right thing

24 to do at that time, there is no doubt about that,

25 but it's outdated, it doesn't apply to what's

00054

1 being done today, and in the kindest terms, we

2 feel that it may be confusing to local carriers.

3 In terms of standards, the point that

4 we would like to make is that drug resistance, in

5 vitro drug response testing is a laboratory

6 test. We are not marketing a product like a drug

7 that goes into the human body and affects both

8 normal cells and malignant cells. We are dealing

9 with information, and the criteria by which in

10 vitro drug response tests should be measured are

11 the same criteria against which other diagnostic

12 tests should be measured.

13 There is a question four, which relates

14 to, should payment be dictated by the results of

15 drug resistance testing? How we answer this is

16 to look at what happens in the real world. And

17 we're dealing with information with a diagnostic

18 test. The fact is that this information is only

19 one of many pieces of information which a

20 physician at the bedside has to integrate

21 together to determine what is in the best

22 interest of this patient. And it's laboratory

23 information, it's clinical information, and it's

24 a multitude of human factors, all determined at

25 the site, that should determine the applicability

00055

1 of this technology. In that case, we don't think

2 that drug resistance information should replace

3 clinical trials, it should only supplement it.

4 We don't think that it should dictate treatment,

5 but it should be one of several factors that are

6 integrated into the treatment decision.

7 And finally, we feel that it should not

8 be used to dictate payment. I can't think of any

9 single issue that would arouse or marshal

10 together the opposition of the oncology community

11 than the thought that a test is going to

12 determine what they have to do at the bedside,

13 singly and in and of itself.

14 So that brings us to the main focus of

15 the meeting today, and that is the fundamental

16 question: Can you take malignant cells from a

17 patient into an in vitro environment, test them

18 in a controlled laboratory assay, identify either

19 resistance or sensitivity, and then translate

20 that into usable information that can be helpful

21 to the clinician when he is at the bedside.

22 That's the fundamental question. And in order to

23 help answer that question, you will find today

24 that there are a number of leading physicians and

25 scientists here to give you their thoughts, their

00056

1 views and their interpretation. They will focus

2 on evidence, clinical application and they will

3 focus on the patient benefit. When you're

4 listening to these individuals, recognize that

5 without exception, they have spent 20 to 25 years

6 of their lives dealing with these technologies.

7 They bring a unique perspective.

8 They just don't know a technology; they

9 know an evolution of multiple technologies. In

10 terms of the literature, they don't know an

11 article; they have read and studied and been able

12 to integrate all of the articles together and

13 created a body of knowledge. And finally, the

14 one thing that should be evident is that the

15 people that are involved in this industry are not

16 just scientists developing a laboratory test;

17 they are clinicians. And they have a perspective

18 to see how you can take laboratory data, input

19 clinical decisions, and over a long period of

20 time they have witnessed the patient benefit.

21 Thank you very much.

22 DR. FERGUSON: Thank you. I guess you

23 have organized this session, so the next

24 speaker?

25 MR. KIESNER: Dr. Weisenthal.

00057

1 DR. WEISENTHAL: Before I get

2 started --

3 MS. TILLMAN: Dr. Weisenthal, excuse me

4 just a moment. We request that all the speakers

5 that are going to come up just make a statement

6 as to whether you're here on your own behalf or

7 who is sponsoring your trip.

8 DR. WEISENTHAL: I am here on my own

9 behalf, I bought my own plane ticket and am

10 paying for my own plane ticket. Can I ask, is

11 Mr. Randy Stein here? Mr. Randy Stein? I didn't

12 see Randy. He was a patient who was going to

13 follow me. Is Dr. William Grace here?

14 Dr. Grace, hi.

15 DR. GRACE: Good to see you.

16 DR. WEISENTHAL: Frank mentioned some

17 of us having 25 years experience in this field.

18 My experience began in the year 1969 when I

19 started doing cell culture drug resistance

20 testing on human tumor specimens while a graduate

21 student at the University of Michigan. My career

22 really began in earnest when I started doing this

23 in the fall of 1978 while I was a clinical

24 associate in the medicine branch at the National

25 Cancer Institute. And ever since July 1st, 1979,

00058

1 this has really been my full-time job.

2 For the first eight years, between 1979

3 and 1987, I did this on a research basis as an

4 associate professor at the University of

5 California, Irvine. Since 1987 100 percent of my

6 time, full time has been spent providing this as

7 a service to patients and physicians in the

8 community. I have about 25 minutes to discuss my

9 life's work. That's not a lot of time, and

10 there's so much, you know, that could be said,

11 and should be said. I will just have to try to

12 do the best I can, I guess.

13 In the beginning, though, I wanted to

14 put everything in context, and you're going to

15 hear from the following speakers admonitions

16 about scientific rigor and levels of evidence and

17 things like this. I think that you have to put

18 this in context. Mr. Kiesner mentioned that we

19 are talking about a laboratory test. We're not

20 talking about a treatment, we're talking about a

21 laboratory test. If you look at analogous

22 laboratory tests such as bacterial culture and

23 sensitivity testing, there is much less direct

24 data indicating correlations between the

25 laboratory tests and the clinical response, and

00059

1 there certainly is a lack of data indicating that

2 it makes an impact on patient care, whether you

3 use the test or not.

4 After 20 years in medical oncology,

5 there's still a debate, should you treat with

6 empiric antibiotic therapy or should you really

7 go to great lengths to try to identify the

8 organism and do sensitivity studies. More than

9 half of the chemotherapy that's given in this

10 country is given for non-FDA approved

11 indications. These are off label indications.

12 In many cases of situations in which Medicare

13 routinely pays for therapy, all that can be

14 pointed to is one or two small pilot studies. In

15 many cases, many oncologists choose drugs on the

16 basis of an abstract that they heard at the

17 American Society of Clinical Oncology.

18 Two weeks ago I talked to Dr. Robert

19 Livingston, who is a professor at the University

20 of Washington, and probably one of the top five

21 experts in the world on chemotherapy and lung

22 cancer. He's very active in the Southwest

23 Oncology Group. We have an explosion of cancer

24 drugs that have been approved in the last five to

25 ten years. We've got Docetaxel, vinorelbine,

00060

1 Gemcitabine, Irinotecan, et cetera. I dare say

2 that the most common regimens used to treat

3 patients today are regimens made up of newer

4 drugs. Carboplatin plus Taxol; Docetaxel plus

5 Carboplatin; Gemcitabine; vinorelbine platin, and

6 so forth. According to Dr. Livingston, firstly,

7 there's no data that none of these are any better

8 than platinum Etoposide, two drugs which are both

9 off patent, much cheaper, outpatient therapy, and

10 personally in his own opinion, there isn't. You

11 know, he doesn't believe that regimens like

12 platinum Taxol are superior to platinum

13 etoposide.

14 And yet, this is the reality today, and

15 that is that the vast majority of individual

16 patients are being treated with treatments that

17 have never been approved by the FDA and are based

18 on levels of evidence that are very preliminary,

19 and that is a fact. And I would like you to keep

20 that in mind when you are looking at the levels

21 of evidence that I am going to be presenting here

22 in the following speakers. I am going to turn

23 this on; okay. I suppose, if I talk loudly, can

24 I go off? I've got to talk on the mike? Too

25 bad.

00061

1 I want to tell you a little bit about

2 the technologies. What we have done here is,

3 this is a Petri dish with some liquid media and

4 this was a patient's stomach cancer. And this

5 has been chopped with scissors into small little

6 pieces about a half a millimeter to a

7 millimeter. Now there are lots of different

8 technologies, but I'm going to try to show you

9 that the technologies have a lot more in common

10 than they have that separate them.

11 One of the differences between the

12 technologies is that some investigators will stop

13 at this point. They will cute the specimen into

14 pieces a half millimeter or so, and they will

15 plate that in plastic dishes with liquid media,

16 and expose them to drugs and then determine drug

17 effect. In other cases, patients will take --

18 investigators or laboratories will take this and

19 pass it through wire mesh screens to give you

20 smaller pieces. And finally, what most

21 laboratories do, is they take the fine pieces and

22 they further digest them with collagenase to

23 break down the tissue matrix to liberate small

24 clusters of tumors.

25 Now when you do that, and what we've

00062

1 done here is that we've taken the stomach cancer,

2 the same one I showed you on the slide

3 previously, and it's been ingested with

4 collagenase, and we've spun this down on a

5 cytospin slide, and we've stained it with a stain

6 that stains dead tissue and dead cells green, and

7 living tissue pink. And you can see here,

8 clusters of pink tumor cells amid a background

9 debris of dead tissue, and there will be single

10 inflammatory cells such as macrophages. Well,

11 applying various methods, you can get very nice

12 enrichment of, you can get rid of all the chaff

13 and get down to the wheat, and what you're left

14 with is microclusters. So one difference between

15 technologies is that some use what I call

16 macroclusters, that is, visible tumor pieces,

17 others digest them down to smaller quantities to

18 give you microclusters.

19 This explains why sometimes the drug

20 concentrations used in the assays are a little

21 bit different. If you've got a large piece of

22 tumor, the drug doesn't penetrate into it very

23 well. And assays that use little pieces of

24 tumors tend to use higher drug concentrations

25 than if you break it down to the smaller cluster

00063

1 level. So -- but in both cases, you're dealing

2 with a similar situation; you're dealing with a

3 tumor and you're testing it in a three

4 dimensional form. And this is very important.

5 So we're testing three dimensional microclusters

6 of cells, other laboratories might test three

7 dimensional macroclusters.

8 This is the same tumor now, the stomach

9 cancer, and it has been cultured for four days in

10 the absence of any drugs. This would be a normal

11 saline control. And this is a drug which was

12 only partially effective, so you've got some

13 reduction in the number of cells. Again, some of

14 the dead cells stained green rather than staining

15 pink.

16 A somewhat more effective drug is this

17 one, and now it has mostly been killed, and you

18 only have a few small clusters of viable tumors

19 left. And a drug that killed everything would

20 give you this, so you'd get the absence of the

21 pink clusters. And the way that this particular

22 end point is scored is manually. Yesterday -- I

23 had to take the red eye last night, because

24 yesterday I spent ten hours counting laboratory

25 assays. It takes me about three hours of my own

00064

1 time to do each and every one of our assays.

2 It's -- I consider the morphologic end

3 point that I just showed you really the gold

4 standard. I like the standard. I like to see

5 the tumor cells on the slide. I like to know

6 whether the drug has worked or not. This

7 technique, though, has some drawbacks. First of

8 all, it takes a lot of time. There aren't very

9 many board certified medical oncologists that are

10 willing to spend three hours looking through a

11 microscope on an individual assay. It's also

12 subjective.

13 Now, that led investigators to try to

14 come up with easier end points, end points that

15 were not subjective and were automated. So what

16 we're talking about here, if you go back two

17 slides, here we are looking at living cell and

18 then with drugs, either cell death assays, and

19 the main assays that I'm going to be talking

20 about here are cell death assays. Later on,

21 Dr. Fruehauf and Dr. Kern are going to talk about

22 cell proliferation assays. But I think you can

23 take all assays and kind of divide them down the

24 middle, it's kind of like the animal kingdom and

25 the plant kingdom, but you've got assays based on

00065

1 the cell proliferation end point, assays based on

2 the cell death end point, and I am going to talk

3 about the cell death assays.

4 Now there's many ways of detecting the

5 death of a tumor cell. This should not be of

6 concern to you. As a clinician, there are many

7 ways of detecting the death of a patient. You

8 can go up and feel for the carotid pulse or the

9 radial pulse. You can put your stethoscope on

10 the chest and osculate for heart sounds, you can

11 observe for spontaneous respirations. You can

12 see if the pupils fixed and dilated. You can

13 take an electroencephalogram, you can measure

14 core body temperature. All of these are methods

15 for determining, is the patient living or dead.

16 Likewise, at the cellular level, there

17 are many ways of determining is the cell living

18 or dead. There is more than one way to skin a

19 cat. So for example, you can look at the

20 morphology of the cell and say has it been

21 killed, has it undergone apoptotic death. You

22 can say, has it lost its ATP. When cells lose

23 their viability, they lose their ATP very

24 rapidly. When cells die, they lose their Krebs

25 cycle reductase activity. So there's one of

00066

1 these assays, the MTT assay, that measures the

2 Krebs cycle enzyme, so when the cell dies, it

3 loses that enzyme activity. And then there's

4 another assay called the fluorescein microculture

5 assay or the fluorescent cytoprin assay, both are

6 really the same thing, and what they're doing

7 there is measuring the membrane integrity with a

8 dye called fluorescein, which is cleaved by

9 membrane esterase and gets trapped in the cell if

10 it has an intact membrane. But the point is

11 here, there are many different ways of

12 determining cell death, just -- there are other

13 ways of determining other things too.

14 Estrogen receptor. Most of the

15 literature which validates the estrogen receptor

16 was based on wet lab assay procedures, but that's

17 been replaced as you know with

18 immunohistochemistry, and at the beginning there

19 really weren't any clinical correlations, but

20 they showed that basically the

21 immunohistochemistry correlated with the wet lab

22 procedures. But these are all methods for

23 detecting cell death.

24 Now, these assays -- that's important,

25 because the assays are very difficult to do in

00067

1 the sense that, I mean, actually generating the

2 data that's going to be presented is an enormous

3 amount of work, and I would love it if I had

4 myself done large numbers of prospective

5 randomized trials in huge numbers of patients, to

6 show beyond the shadow of a doubt that patients

7 did better when treated on assay results.

8 Goodness knows, I tried, and I and several other

9 people made major efforts. I won't give you the

10 anecdotes of the various trials that never got

11 underway, or got underway and were well funded

12 but didn't accrue patients and so forth.

13 But suffice it to say, this is

14 difficult work; if it wasn't difficult work,

15 after 20 years of full time in it with people

16 like me and Dr. Bosanquet and Dr. Kern, and

17 others, Dr. Salmon, Dr. Von Hoff, who is

18 certainly one of the most energetic organizers of

19 clinical trials, even he was unable to

20 successfully complete a single study. So this is

21 very difficult, so it's important to look at all

22 of the evidence. So that's why I am going to try

23 to make a point that you need to lump together

24 these various cell death end points.

25 Basically the assays are done in the

00068

1 same fashion. You take the tumor, you culture it

2 for four to five days; you expose it to drug, and

3 then you determine, are the cells living or

4 dead. And the fact that there is different ways

5 of determining is the cell living or dead is not

6 of importance.

7 This is another assay here. This is

8 the MTT assay, and this is based on mitochondrial

9 succinate dehydrogenase activity. Living tumor

10 cells will produce a lot of pink reagent and if

11 they have been killed they don't produce that

12 reagent, so this is a positive control. These

13 are ineffective drugs, this is a single effective

14 drug.

15 And what we do is since there's

16 advantages -- the advantages of the DiSC assay,

17 which is the microscope assay is that to me, it's

18 the gold standard. You're actually looking at

19 the tumor, you're seeing whether the drugs really

20 work. The disadvantage is that it's a subjective

21 test and it's labor intensive.

22 The advantage to the MTT assay is that

23 it's objective, you get a nice machine readout,

24 but it's not specific for tumor cells. If you

25 have some normal cells in there, it can skew the

00069

1 results, so it's very important that you take a

2 lot of efforts to make sure that you've got a

3 population of cells.

4 In practice, we do both end points.

5 These cells here, we've had some fast green dye

6 added to them, they're going to be spun down on

7 cytospin slides. These are the same drugs tested

8 in the MTT assay. So we run all these assays in

9 parallel; we always do an MTT and a DiSC,

10 microscope assay, and by doing that, I think I've

11 got a good handle on what's happening.

12 Now, these end points correlate very

13 well together. This allows us to lump together

14 the results for analysis. This shows 775 solid

15 tumor specimens tested to Cisplatin, and on the Y

16 axis is the MTT assay result, on the X axis is

17 the DiSC assay result, and you can see that in

18 cases where we've got pure tumor preparations,

19 there's a very good correlation between the two

20 end points.

21 There have been many papers published

22 in the literature. These are just -- I know it's

23 difficult to read, but these are papers comparing

24 the two end points, DiSC and MTT, fluorescein

25 diacetate and DiSC, MTT and fluorescein

00070

1 diacetate, DiSC and ATP, and all of these end

2 points for cell death, not surprisingly,

3 correlate with each other very well.

4 How is this information used in the

5 real world? Well, what is done is this, and that

6 is that in the beginning, 20 years ago people had

7 the idea that what they were trying to create was

8 a scale model of chemotherapy in the laboratory.

9 And so they tried to use what are known as

10 clinically achievable drug concentrations. And

11 Dr. Alberts, who's a speaker here, is a real

12 pioneer there, and Dave did a lot of work in the

13 late '70s figuring out exactly what the

14 clinically achievable levels of different drugs

15 were, and he created some tables that I and other

16 investigators used initially.

17 I will tell you, though, that if you

18 read the literature today, that's not what people

19 do. Here's what they do, and that is that you

20 get a drug and you do some training set studies,

21 but you try to find the concentration that gives

22 you the widest scatter of results. So on this

23 slide here, what I'm showing is a thousand

24 randomly selected fresh tumor MTT assays for

25 Cisplatin. And this is percent of control cell

00071

1 survival. 100 percent cell survival means the

2 drug didn't work, the cells are all alive; zero

3 percent cell survival means the drug did work,

4 the cells are all dead. And what you can see is

5 that in a thousand randomly selected solid tumor

6 assays, there is a widespread scatter of

7 results.

8 So in fact, you try to choose the drug

9 concentration which gives you the greatest

10 standard deviation. You choose a concentration

11 with an index concentration which gives you the

12 greatest scatter. You can then draw the line

13 down the middle for analysis. And operationally

14 you say if the cells are killed in the culture

15 dish, that's resist -- they're sensitive to the

16 drug. If they are not killed, they are resistant

17 to the drug.

18 Now in practice, you can see that

19 there's a lot of grouping around the middle, so

20 obviously what we do if they are around the

21 middle, that is, if they are plus or minus a half

22 standard deviation from the median, we just say

23 it's in the median and we really can't tell you

24 anything about it, about it. But if it's down

25 here, it's clearly sensitive; if it's up here,

00072

1 it's clearly resistant.

2 Now 20 years ago when we first started

3 doing this, we formulated a hypothesis, and our

4 hypothesis was that if you used this method and

5 you obtained a broad scatter of results, that on

6 average, patients with resistant assays would do

7 worse than patients with sensitive assays. That

8 was the hypothesis. And that is really what I

9 call the central hypothesis to all of this

10 testing, and that is, the central hypothesis is

11 the drugs testing in the sensitive range will be

12 more likely to work than drugs testing in the

13 resistant range. 20 years ago, just a

14 hypothesis. What did the data show?

15 Well, in the 20 years since then, there

16 have been many papers published, now in excess of

17 40 papers showing correlations with cell death

18 assays and results of chemotherapy in the

19 patient. For purposes of this slide, I have

20 arranged them in order of increasing response

21 rates in the overall patient population, so this

22 white dashed line shows the response rates in a

23 given study for all the patients in the study.

24 Each of the vertical lines represents a different

25 study. So in this slide, I think I'm showing, if

00073

1 I can read it, 36 studies, or 35 studies,

2 totaling 1603 patients. But what you can see is

3 that rating from low response rate tumors to high

4 response rate tumors, and this would be something

5 like previously treated phalangeal carcinoma, and

6 this would be acute lymphoblastic leukemia, but

7 in 35 out of 35 studies, in every single case the

8 hypothesis has been confirmed. In fact, patients

9 who are sensitive in the assay do better than the

10 group as a whole. Patients that are resistant in

11 the assay do worse than the group as a whole.

12 And patients that are sensitive in the assay do

13 dramatically better than patients that are

14 resistant in the assay.

15 So in other words, this assay is an

16 excellent prognostic factor for prognosis if

17 treated with chemotherapy. If you are treated

18 with chemotherapy and the test is in the

19 sensitive range, you do better than average. If

20 you're treated with the drugs in the resistant

21 range, you're worse than average. In solid

22 tumors, the advantage to getting an assay

23 sensitive drug over an assay resistant drug is a

24 nine to one advantage, patients are nine times

25 more likely to benefit if they're sensitive in

00074

1 the assay than if they're resistant.

2 People say that this field is

3 controversial. This is not controversial. These

4 data are unchallenged. There has never been a

5 single study of these technologies in modern

6 history which has failed to show this. If you

7 break it down by tumor types, and here I'm sorry,

8 I can't read it, my contacts -- I took the red

9 eye last night and my contacts are a little bit

10 not clean, but what I have done here is broken

11 this down by disease type and it includes things,

12 stomach cancer, breast cancer, ovarian cancer,

13 non-small cell lung cancer, multiple myeloma,

14 chronic lymphocytic leukemia, acute

15 nonlymphocytic leukemia, acute lymphoblastic

16 leukemia, and so forth and so on. But again, if

17 you break this down by tumor types, patients that

18 have sensitive assays do better, patients that

19 have resistant assays do worse.

20 So hypothesis I would put to you is not

21 an extraordinary hypothesis, it's a very ordinary

22 hypothesis, and yet, this is an extraordinary

23 level of proof. The hypothesis holds. These

24 data are unchallenged. No one has ever shown

25 anything to the contrary.

00075

1 You can use the technique that was

2 described in the New England Journal of Medicine

3 a few years back, of cumulative Meta analysis,

4 and I don't have time to explain this, it's in my

5 handout, but basically when you do this, these

6 are 95 percent confidence limits, and what you

7 see is that if you had a P of 10 to the minus

8 eighth, patients that are sensitive in the assay

9 do better that the group as a whole. At P 10 to

10 the minus eighth patients that are resistant in

11 the assay do worse than the group as a whole, and

12 these are again, thoroughly consistent.

13 Now receiver operator curve plots and

14 Bayes' Theorem. Receiver operator curve plots

15 are, receiver operator plots are used as an

16 assessment of laboratory tests. To generate

17 receiver operator plots, one needs to know how

18 changing the cutoff lines affects sensitivity and

19 specificity. However, with the literature

20 validating cell culture drug resistance testing,

21 what are available instead is the sensitivity and

22 specificity of a single cutoff line, which is

23 around the median. And also, the test accuracy

24 at different pretest response probabilities. The

25 broad applicability of test results to different

00076

1 disease states may be evaluated by comparing

2 calculated base predictions to actual

3 observations. This is described in detail in my

4 handout, if I go a little rapidly.

5 Most studies do not show this type of

6 data. This is a single study by Wilbur in

7 non-small cell lung cancers published some years

8 back, but basically it was showing that when you

9 change the cutoff of the assay from 90 percent

10 survival to 80 percent, 70 percent, 60 percent,

11 what happens is the actual sensitivity and

12 specificity of the assay changes, as you would

13 expect, but in all cases, people with sensitive

14 assays are more likely to respond than patients

15 with resistant assays. So this is -- these are

16 not an artifact of just drawing, you know,

17 picking a cutoff.

18 By applying Bayes' Theorem, you can

19 generate the following theoretical curve. These

20 tests in aggregate, if you add up all the 2,000

21 or so clinical correlations that have been

22 published, they have an overall specificity for

23 drug resistance of .92, an overall sensitivity

24 for drug resistance of .72. And if you do that,

25 you get these sorts of predictions, and this

00077

1 shows the relationship between pretest response

2 probability, expected response probability, and

3 then response probability given a different test

4 result. So the blue line shows what it would be

5 predicted for patients with a sensitive assay.

6 So in other words, let's take colon

7 cancer as an example. Untreated colon cancer's

8 got a 20 percent chance of responding to 5 FU.

9 If it's assay sensitive, the prediction says that

10 the patient has a 40 percent chance of

11 responding. If it's resistant, it goes down to

12 about 2 percent. Contrary-wise, if you're

13 dealing with untreated ovarian cancer, which has

14 a 75 percent response rate, if you're sensitive

15 in the assay, it goes up close to 90 percent and

16 if you're resistant, it falls down to about 15 to

17 18 percent. So those are just the theoretical

18 predictions.

19 How about if you break this down by

20 individual types of tumors? And I think that

21 these data compelling showed that these assays

22 are broadly applicable for really all types of

23 tumors in which they've been studied, both solid

24 tumors and hematologics, ranging from stomach

25 cancer, colon cancer, non-small cell lung cancer,

00078

1 ovarian cancer, breast cancer, chronic

2 lymphocytic leukemia, acute lymphoblastic

3 leukemia, acute nonlymphocytic leukemia, in all

4 cases it holds exactly according to base

5 predictions.

6 There are many correlations published

7 in the literature about patient survival, and

8 these are the patients that, survival of patients

9 sensitive in the assay, survival resistant in the

10 assay. I give references in my handout, and

11 these will be discussed by other speakers.

12 Now I have to, I'm already bumping up

13 against my time limit, but I have to discuss a

14 group of papers which are very, very important,

15 because you as panelists have probably spent the

16 most attention to these, because these were

17 studies done at the National Cancer Institute,

18 published in prestigious journals, and so

19 naturally you think that these are really quite

20 important papers. There was a review by Cortazar

21 and Johnson in The Journal of Clinical Oncology.

22 What this review showed was that there were three

23 non-randomized small studies which showed

24 nonsignificant inferior survival with assay

25 directed therapy, compared to control therapy.

00079

1 Again, these were non-randomized studies and the

2 results were nonsignificant but still, three of

3 them showed, suggested slightly inferior survival

4 with assay directed therapy compared to control

5 therapy. What's important for you as panelists

6 to realize is that one none of these studies, not

7 a single one, evaluated the fresh tumor assays

8 which are used in the real world and have been

9 used for the past 12 years, and which are now

10 being considered for reimbursement. This whole

11 paper is utterly irrelevant because it does not

12 review the technologies that are under

13 consideration here.

14 Specifically, I want to take you

15 through the three NCI studies. The NCI did a

16 study in non-small cell lung cancer, they did a

17 study in extensive disease small cell lung cancer

18 and limited disease small cell lung cancer. In

19 general, the non-small cell study was highly

20 negative, highly negative. There's no one that

21 could read that paper that could possibly

22 conclude that this particular assay was of any

23 utility whatsoever. It's a totally negative

24 study. The extensive disease small cell study

25 was modestly positive. The limited disease study

00080

1 was highly positive, and I'd like to tell you why

2 that is.

3 First of all, the limited non-small

4 cell study. This is probably the most important

5 one for you to consider; this gets quoted the

6 most. In 1994, when I presented this at

7 California Blue Shield, I had to spend half of my

8 time debunking this one paper, because somebody

9 at the University of California San Francisco

10 brought it up. For the past seven years I've had

11 people come up to me over and over and they say,

12 well, they tried that at the NCI, it didn't work,

13 they're the mecca of meccas, if they couldn't get

14 it to work, what makes you think you can get it

15 to work? Well, I've been doing this full time

16 for 20 years and if you work at something very

17 hard, you can actually get it to work.

18 But let's talk about this. Non-small

19 cell lung cancer study from the NCI. Firstly,

20 they used passage cells. These were not fresh

21 tumor assays. It said in the study methods that

22 they were fresh tumor assays, they were not.

23 That's an incorrect statement. This paper was

24 not written by an investigator associated with

25 the study. This was written by an investigator

00081

1 named Gail Shaw, who at the time was an oncology

2 fellow. She rotated through the NCI Navy branch.

3 This was after the investigators, after Dr.

4 Meadows' group already left. She went and did

5 these chart reviews, she wrote the paper, and she

6 incorrectly stated that these were done on fresh

7 tumors.

8 In fact, I called Audie Gasner at the

9 University of Texas, and he confirmed that every

10 single one of these studies were done on passage

11 cells. That's what they were trying to do, they

12 were trying to see, could they use cell lines to

13 do assays. So these are not fresh tumor assays,

14 these are on cell lines.

15 Why is that important? Well, because

16 papers have shown if you generate cell lines,

17 that with subsequent passages, that the drug

18 resistance changes. And this has been well shown

19 in the literature.

20 Secondly, these were monolayer

21 cultures. These were not -- they were not

22 testing three dimensional cultures of clumps of

23 cells, clusters of cells, they were testing

24 monolayers. And in a study, seminal study in

25 PNAS, 1993, Tyker and Kerbil, they showed that if

00082

1 you do monolayer cultures, that that doesn't

2 correlate, but that when you do three dimensional

3 cultures, it does. These were monolayer

4 cultures. No one does monolayer cultures. In

5 this study they did.

6 They had a 22 percent overall

7 evaluability rate, and 7 percent with lung

8 primaries. In 1985 I did a study in conjunction

9 with the Loma Linda VA, Dr. Dave Wilbur, in which

10 they just sent us by regular mail specimens of

11 non-small cell lung cancer. So this would take

12 two to three days to arrive in the mail, and this

13 was with technology in 1985. We had an overall

14 evaluability rate of 75 percent. Today -- I

15 reviewed my data last night before coming here,

16 and in the past five years, we have received 347

17 non-small lung cancer specimens, and 326 of those

18 assays were evaluable, which is a 93 percent

19 overall evaluability rate. Twenty of those had a

20 negative histology, and that's a reason for

21 inevaluability, so if you only look at cancers

22 that actually had cancer when it made it to our

23 lab, we had a 97 percent evaluability rate,

24 including a 96 evaluability rate with 124 primary

25 lung tumors.

00083

1 So these guys are testing a subset of

2 patients, 22 percent, and 7 percent with lung

3 primaries. So what do we know about that

4 subset? Well, it turns out that they had

5 previously shown in Annals of Internal Medicine

6 that when they got a tumor that they were able to

7 subculture, that just the fact that the cells

8 could be subcultured was a powerful negative

9 prognostic factor. And they said that this is a

10 marker, in the Annals of Internal Medicine paper,

11 for biologic aggressiveness. So think about it.

12 The only people getting assay directed therapy

13 are the people with the worse prognostic group,

14 with the biologic aggressive group, and they're

15 being compared with a group of patients that you

16 can't subculture, and they have the biologically

17 indolent group.

18 And finally, and last but not least,

19 they did not give assay directed therapy until

20 the fifth treatment cycle. They biopsied the

21 patient, it took them four treatment cycles to

22 get these cell lines going to test them, and so

23 they didn't actually get the assay directed

24 treatment until five treatment cycles.

25 This paper has been thrown up in my

00084

1 face again and again and again at the best

2 universities in the country, and this paper is a

3 bunch of rubbish. It should never have been

4 published. It is misleading, and it's terrible

5 that it keeps resurfacing. And I hope that the

6 previous speakers, or speakers that follow me

7 just don't -- this paper is irrelevant, let's not

8 waste any more of our time about it.

9 Now, the small cell lung cancer study

10 in extensive disease, this was modestly

11 positive. Why was that? Well, they still used

12 passage cells; that was bad. This paper in the

13 International Journal of Cancer again showed that

14 if you used passage cells in small cell lung

15 cancer, that doesn't correlate. However, small

16 cell growth is three dimensional spheroid

17 cultures, unlike non-small. That's good. They

18 had a 55 percent assay evaluability rate. That's

19 good too. They're not dealing with a selected

20 population. Assay directed patients were a

21 similar prognostic group relative to control

22 patients. So all these were good. And the only

23 thing that was bad is that they weren't giving

24 assay directed therapy until the fifth cycle.

25 Now I mentioned that the results of

00085

1 this study in limited disease were positive. In

2 fact, the patients who got assay directed therapy

3 lived a median of 38 months and those who got

4 standard therapy lived a median of 16 months.

5 This was statistically significant. It's a small

6 study, but it was statistically significant.

7 So why was this study positive and the

8 other one wasn't positive? For perfectly

9 explainable reasons. They're not treating a bad

10 prognostic group. They're using three

11 dimensional cultures. I would say they would

12 have had even better results if they had used the

13 assay chosen drug up front, but they didn't.

14 And in the extensive stage non-small

15 cell study, this was less positive than in the

16 limited stage study; why was that? Well, it

17 turns out that the assay directed were also a

18 worse prognostic group. In the extensive disease

19 study, patients could only get assayed if they

20 had peripheral lesions for biopsy under local

21 anesthesia. They did not do general anesthesia

22 for this. And they showed that just, when they

23 analyzed the patients that had biopsiable tumors

24 versus patients that didn't have biopsiable

25 tumors, there was a significantly shorter

00086

1 survival if you had a biopsiable tumor. That

2 makes sense. They've got more extensive disease;

3 of course they're going to die faster. So if the

4 only people getting assay directed therapy are

5 people that have a higher tumor burden, that's

6 really biasing it against it. And also, they're

7 not getting treated until the fifth cycle.

8 In summary, these NCI studies have

9 nothing to do with the real world. They don't

10 apply to the technologies that you're evaluating.

11 They are utterly irrelevant.

12 There have been many studies showing

13 correlations with patient survival. I wish that

14 I had time to take you through these and show you

15 the survival curves. In particular, one group of

16 studies is not going to get presented here, and I

17 just want to tell you very briefly about it. And

18 that is studies by Vierman's group in acute

19 lymphoblastic leukemia. There is some

20 controversy. Should you include pediatric

21 leukemia in a discussion about assays applicable

22 to Medicare patients. I think that you have to,

23 because it makes a consistent story.

24 If you look at the data that validate

25 these technologies, just to pick one type of

00087

1 disease, human lymphatic neoplasms, AOL and COL.

2 In 1962 an investigator named Schreck showed in

3 Annals of Clinical -- Journal of Clinical

4 Investigation, that if you did an apoptosis assay

5 on fresh cultures of COL, that radiation response

6 in that assay correlated with patient survival.

7 That work was lost. Nobody knew about apoptosis

8 in the '60s, nobody cared about it. In the early

9 '60s -- in the early '70s, people that were

10 working on assays and leukemia had the idea that

11 you had to do clonogenic assays. In fact, those

12 of you who are familiar with the literature, they

13 would look at clonogenic assays, actually do

14 clones of clones. They would plate single cells,

15 let them grow two to three weeks until you had a

16 clone, remove the clone, desegregate it and

17 reclone it, so you had this very cumbersome assay

18 that would take about six weeks, and people

19 thought that you had to do that, because there

20 was this phenomenon of the stem cell, and the

21 only thing that's relevant for chemosensitivity

22 is the stem cell.

23 I came up with this really radical

24 idea, based on Dr. Schreck's work, which I was

25 familiar with, that if you just expose the cells

00088

1 to the drug and divide them into groups, and

2 one's above average and one's below average, that

3 that will be a strong correlation with clinical

4 response. I published several papers on these in

5 the early 1980s. And what I showed in one of the

6 papers, for example, is that if you looked at

7 assays on both COL and AOL, pediatric AOL, adult

8 COL, that there was strong correlations with

9 clinical response. Furthermore, if you looked at

10 previously treated patients, they were much more

11 resistant, significantly more resistant than

12 untreated patients. And finally, if you I

13 followed individual cases of patients over time,

14 if they were assayed multiple times with no

15 intervening chemotherapy, there was no change in

16 the assay results, but if they had intervening

17 chemotherapy, they became demonstrably more

18 resistant in the assay. As a result of these

19 papers, there were other investigators that got

20 into the field.

21 DR. FERGUSON: Dr. Weisenthal, there

22 are three other people apparently that are

23 supposed to talk, if I extend the time to 10:15,

24 which I said I would, and so perhaps you could --

25 DR. WEISENTHAL: Well, Mr. Stein is not

00089

1 here today.

2 MR. STEIN: I'm here.

3 DR. FERGUSON: So that's four other

4 people. So maybe, if you could wind up?

5 DR. WEISENTHAL: Okay. It's very

6 frustrating. You know, I've got some really good

7 stuff to tell you.

8 DR. GRACE: I'll donate my time to Dr.

9 Weisenthal. Will that help?

10 DR. FERGUSON: That will help some.

11 DR. WEISENTHAL: I think I can finish

12 up in five minutes, can I -- okay.

13 I was going to summarize the data in

14 human lymphatic neoplasms. So basically, you

15 know, that's what I showed. Dr. Bosanquet has

16 for the last 18 years been studying these assay

17 systems in chronic lymphocytic leukemia, and I'll

18 let his work speak for itself.

19 The work that can't speak for itself is

20 a parallel work that was done in acute

21 lymphoblastic leukemia, and this work with the

22 MTTS, and what these investigators did at the

23 Free University of Amsterdam, first they started

24 using the DiSC assays, just as I described. They

25 said it's a lot of work, you've got to count with

00090

1 the microscope, so they preferred using the MTT

2 assay, but they used exactly the same culture

3 conditions, 96 hour culture, same exact identical

4 conditions. What they showed in a series of very

5 rigorous trials, published in excellent journals,

6 several publications in Blood, publications in

7 the Lancet, these are superb studies, and I am --

8 they should not be excluded from this

9 consideration. They showed in very rigorous

10 studies strong correlations between the assay

11 result and patient survival.

12 In fact, the assay results were the

13 strongest predictive factor, and it turns out

14 they were the only independent predictive factor.

15 And all these other cell marker studies that

16 people do on pediatric AOL were not significant

17 once you consider the cell culture results.

18 So if you take and look at

19 historically, the correlations with response, the

20 correlations with treatment status, and then if

21 you look at Dr. Bosanquet's excellent studies in

22 chronic lymphocytic leukemia, and you also

23 consider very identical studies in acute

24 lymphoblastic leukemia, it is a continuous

25 consistent whole.

00091

1 Now the last thing I want to briefly

2 address is the issue of drug synergy. Should you

3 test single agents, combinations? What this data

4 are showing is that most drug combinations in

5 human solid tumors are not synergistic. There is

6 very little, if any, evidence of clinical synergy

7 in clinical data treating human solid tumors.

8 Combinations, citoxin plus adreomyecin is never

9 synergistic. Taxol and platinum is not

10 synergistic, it's additive. These data show, it

11 says platin and etoposide -- these are results

12 where it says platin alone, etoposide alone, this

13 is what you would expect if they were additive,

14 and in fact they are additive. Now there are

15 occasional combinations which are uniquely

16 synergistic. One of them is gemcitabine plus

17 cisplatin, which is one of the most exciting new

18 combinations to come along in a long time. In

19 contrast, this is a highly synergistic

20 combination, and so when you've got a synergistic

21 combination, it make sense to test the drugs in

22 combination.

23 I am going to show -- this is a patient

24 that Dr. Nalick is going to be presenting. This

25 is an ovarian cancer patient, control culture.

00092

1 Carboplatin alone had a minimal effect.

2 Gemcitabine alone had a minimal effect.

3 Combination of the two wiped everything out.

4 Highly synergistic.

5 The next speaker you're going to hear

6 is Mr. Stein, and he was the subject of this

7 paper in Scientific America last February, and

8 he's going to tell you his story, but I'm going

9 to show you his assay. These are his control

10 cultures, this is pancreatic cancer. Control

11 culture. Platinum alone, modest effect;

12 Gemcitabine alone, modest effect; platinum

13 Gemcitabine, wiped out.

14 I'll stop here. Anyway, and other

15 speakers I'm sure will amplify the remarks that I

16 said.

17 I'd now like to introduce Mr. Randy

18 Stein.

19 MR. STEIN: I want to thank all the

20 distinguished members of this advisory committee

21 for listening to my testimony. I also want to

22 publicly state that I have no financial interests

23 or involvement with any manufacturers of any

24 products being discussed or with their

25 competitors. I would also like to inform you

00093

1 that I feel the importance of my being here to

2 testify is of such magnitude that I flew here

3 from Southern California and postponed a trip to

4 Acapulco. I plan on joining my wife, who went as

5 previously scheduled and is awaiting my arrival

6 directly after my testimony.

7 I truly feel I had no choice but to

8 testify though, because I realize that without

9 the cell culture drug resistance testing, I would

10 be dead. And to think that other people will die

11 if drug resistance testing is not approved, while

12 I sit and bask in the sun is totally unacceptable

13 to me. You see, I was blessed with an incredible

14 gift, the gift of life, and now everything I do

15 is about giving back, touching as many lives as

16 possible and trying to make a difference in the

17 cancer community.

18 I was diagnosed with four stage

19 non-operable pancreatic cancer that had

20 metastasized to my spleen and kidneys on January

21 22nd of 1997. My CA-19 tumor markers, a blood

22 test used to determine the severity of the

23 disease, were at 12,930, with normal being 0 to

24 37. My gastroendocrinologist sent me to a local

25 private practice oncologist named Dr. Stuart

00094

1 Nagasawa. Dr. Nagasawa explained to my wife and

2 I what I had, the grim statistics associated with

3 four stage pancreatic cancer, and then told us if

4 he were to treat me, he would like to get

5 aggressive. He went on to explain to us that

6 getting aggressive meant having a laparoscopy

7 done, taking a tissue sample from one of my many

8 tumors, and sending it to the Weisenthal Cancer

9 Group.

10 He then told us the principles behind

11 sensitivity testing and why that was so

12 aggressive. He also explained that with such a

13 fast growing cancer, my chances for recovery

14 would be better by knowing which chemotherapy was

15 the most effective, and even more importantly,

16 which was the least effective on my tumor. He

17 also explained that with conventional treatment,

18 we would not know the effectiveness of any

19 chemotherapy on my tumor for three months, and at

20 this time that was my expected life span. This

21 seemed like a no brainer to me, no now or maybe

22 blow my chances on a chemotherapy that wasn't

23 effective.

24 Being a little on the snobbish side and

25 realizing the probable outcome of my disease, we

00095

1 wanted to talk to some of the top doctors in

2 California. Second, third, fourth and fifth

3 opinions. To tell the truth, I was sorry we

4 didn't stop at the first. We went to the most

5 famous and most prestigious facilities available,

6 UCLA, USC, City of Hope, and the John Wayne

7 Cancer Center. We were told go fishing if that's

8 what you like. We were told, maybe three

9 months. We were told, I don't know why you're

10 still alive. We were kept waiting by one of the

11 grand gurus of cancer for over an hour, told yep,

12 it's pancreatic cancer, you have three months to

13 live, and I have to leave; I'm late for a root

14 canal appointment.

15 We discussed sensitivity testing with

16 each and every one of them and their collective

17 reactions were all the same. They were against

18 relying on the cell culture drug resistance test.

19 Their reasoning was simple, the test tube

20 doesn't exactly duplicate conditions in the body,

21 and the testing may prejudice the doctor's

22 choices. These same doctors were also sure I

23 would be dead two and a half years ago.

24 After much deliberation with friends

25 and family, we decided that although testing in

00096

1 the test tube may be different than the human

2 body, it did give us a better idea of what the

3 tumor was resistant to and what might work. And

4 let me tell you, when you have three months to

5 live, that sounds a lot better than just guessing

6 what chemotherapy to use, or doing what everyone

7 else is doing, with little or no help.

8 Conventional for four stage inoperable

9 pancreatic cancer patients has a 3 percent chance

10 of prolonging life for more than three months,

11 and 0 percent past one year. And as far as the

12 doctors being prejudiced, I'm a child of the '60s

13 and I hate prejudice, but this type of prejudice

14 seems very reasonable. We made the decision to

15 use Dr. Nagasawa. We felt getting aggressive

16 made more sense than waiting to die.

17 The current FDA approved treatment for

18 pancreatic cancer with metastases is Gemzar, and

19 had I treated within the FDA guideline, I would

20 be dead. When we received the results of the

21 drug resistance testing, Gemzar alone scored very

22 poorly. But the combination of Gemzar, when

23 combined with Cisplatin reacted very favorably on

24 my tumor sample. Dr. Nagasawa explained to us

25 that although by themselves, the Gemzar and

00097

1 Cisplatin scored poorly, together there was an

2 incredible synergy, meaning one plus one did not

3 equal two, it equaled ten. And without the

4 testing, we would have never known it.

5 He then told us I would be on this

6 combination of chemotherapies until further

7 notice. Three months later he took another CA-19

8 tumor marker test. The results came back at

9 8,970, down by approximately 30 percent. That

10 gave us hope, and hope is crucial to anyone's

11 survival. After nine months, in September of

12 1997, my tumor markers came down to 6,300, and my

13 chemotherapy was changed to every two weeks. The

14 following year showed a continuous decline in my

15 CA-19 tumor markers. I started to gain back the

16 50 pounds I had lost, and I was able to

17 discontinue the use of all pain medications.

18 Eighteen months after diagnosis, on

19 June 17th, 1998, my doctor called and said I've

20 got good news. I just received your latest CA-19

21 tumor markers and they came back at 31.4, well

22 within the normal range, congratulations. I

23 looked at my wife and she looked at me. The

24 tears started rolling off both of our faces. I

25 then replied, congratulations to you, Doctor.

00098

1 The rest of that night was spent celebrating. We

2 picked up chili cheese dogs and Dom Perignon,

3 called all the wonderful friends and family that

4 were so incredibly supportive during this time,

5 and had one incredible evening. The joy of that

6 night will live with me forever.

7 August of 1998, my chemo was reduced to

8 once every three weeks, giving my wife and I back

9 a life and allowing us the ability to travel, as

10 well as spend time with our loved ones. July of

11 1999, my doctor arranges a PET scan, and the

12 results come back: No caner anywhere in my

13 body. August of 1999, with our healing

14 professionals, the decision is made: No more

15 chemo.

16 The NCI pamphlet on pancreatic cancer

17 on page 10, states that cancer of the pancreas is

18 very hard to control, that the disease can only

19 be cured when it is found in its early stage,

20 before it has spread. Last week we had a huge

21 fund raising function for the pancreatic cancer

22 action network. I had the opportunity to tell my

23 story to the 800 people in attendance. There

24 were doctors in the audience from the top

25 facilities in the country, Johns Hopkins, M.D.

00099

1 Anderson, Sloan Kettering, and many others.

2 Needless to say, they were all in awe of my

3 recovery and were very anxious to speak

4 personally to my doctor and myself regarding my

5 recovery.

6 I was lucky. Although my insurance now

7 covers this test, at the time it didn't.

8 Dr. Weisenthal allowed me to pay him over time,

9 and by the grace of God we could afford to do

10 that. Most people on Medicare can't, and they

11 don't need or deserve to die. It is up to us who

12 survive to become activists and to do our best to

13 see that proper actions are taken and that the

14 effective treatment and diagnostic aids are

15 researched and made available in the future. All

16 of this is why I am standing here today, cancer

17 free and begging you to approve this type of

18 testing.

19 Thank you for your consideration of

20 this very important coverage.

21 DR. FERGUSON: Thank you very much, Mr.

22 Stein.

23 Richard Nalick, or are you --

24 DR. NALICK: I am.

25 Good morning. My name is Richard

00100

1 Nalick. I'm a gynecologic oncologist at USC

2 School of Medicine, professor there, clinical

3 professor, and for about the last 15 years in

4 private practice in gynecologic oncology in Los

5 Angeles. I have no involvement with any company

6 or any individual or manufacturer of any of these

7 products being discussed today.

8 I am here today because of my interest

9 and passion in this particular form of testing of

10 chemotherapeutic agents. I finished my training

11 in gynecologic oncology in about 1974, went to

12 Texas for three years at Parkland Hospital, and

13 then back to USC as a professor. And at the very

14 beginning, I was interested in this form of

15 testing. It made the same sense to me as testing

16 a urine for culture insensitivities. If you can

17 do that for bacteria, why not do it for cancer

18 cells? Of course we had less drugs at that time

19 but it was still interesting.

20 We tried to set up an assay at USC and

21 it was a clonogenic type of assay, and it was

22 fairly good, but we didn't have the finances to

23 really carry that through. I then started

24 sending the assays of tumor to Dr. Von Hoff in

25 San Antonio. The problem there was that most

00101

1 often, because they had to be sent in dry ice,

2 the tumor tissue usually didn't make it to San

3 Antonio, at least not in good shape, and I had

4 several communications with Dr. Von Hoff and we

5 had some data, but it was somewhat difficult to

6 interpret, and I stopped using that assay.

7 About that same time a man by the name

8 of John Daniels, who's a medical oncologist and

9 Ph.D. at USC School of Medicine developed his own

10 clonogenic type assay, only this he did at USC,

11 and I started sending tissue to him. I

12 eventually collected at least data on 200

13 patients of my own, that I obtained initially for

14 use later, if the patient did not respond to

15 primary treatment. However, as time went by I

16 saw more often and more often that the findings

17 in the test correlated with my findings in the

18 patients, so I started using the assay, certainly

19 in patients who failed the treatment, and towards

20 the end started using the assay up front, because

21 I had such confidence in those findings.

22 However, after accumulating about 200

23 patients of experience, Dr. Daniels went on to

24 other things and started another company, and his

25 physicians were referred to Oncotech, which had

00102

1 just started. So I have experience with about

2 200 patients with the clonogenic assay, and then

3 at least a hundred if not more patients

4 experience with proliferative assays at Oncotech,

5 and that test proved very effective in my

6 practice.

7 And at this point now, I started using

8 this assay up front. Some people thought that

9 wasn't ethical, it wasn't correct, but with my

10 experience I had at that point was such that I

11 knew that if the test showed extreme drug

12 resistance for that drug, that drug did not work

13 in my experience, so I stopped using it. And I

14 would then pick the drugs left that looked the

15 best, keeping in mind toxicity, and their track

16 record in oncology.

17 Well, after a period of time, I sent

18 assays to Dr. Weisenthal's group, and I actually

19 compared them with Oncotech and Weisenthal, same

20 patient, two assays, and they correlated fairly

21 well. But in my opinion, the Oncotech assay was

22 certainly excellent for finding drugs that the

23 patient would not respond to very actively. The

24 Weisenthal assay, however, allowed me to test

25 more drugs.

00103

1 And I also felt relative to what you

2 have seen already and will hear later on, that

3 testing combinations is important also, and

4 actually with Oncotech we tested a few

5 combinations. But since with Dr. Weisenthal's

6 group being utilized, I've tested all patients

7 with single agents and combinations. So I always

8 test for carboplatin, cisplatin, one of the

9 platinums in Taxol, one of the platinums in

10 Topotecan, one with Gemcitabine. Other

11 combinations that aren't usually used, but that

12 have shown benefits in other studies, such as

13 Navelbine and Thiotepa. I will test Doxorubicin

14 and Doxil.

15 And I can tell you that up front, long

16 before the papers came out, we knew per that

17 assay that there was synergy between Gemcitabine

18 and platinum, that we didn't know before. And

19 that it had become, as far as I'm concerned, that

20 should be the gold standard today, not Taxol and

21 platinum, in my opinion. We also saw that

22 combinations worked in other situations too; very

23 often, carboplatin and Taxol would be equal to

24 Gemcitabine and Taxol.

25 But my feeling was that if they were

00104

1 equal on the assay, and knowing that really

2 platinum was probably the most important drug,

3 and now knowing that there is synergy between

4 Gemcitabine and platinum, and knowing that there

5 is no hair loss with platinum and Gemcitabine,

6 but there is with Taxol, and knowing that there

7 is there is no significant neurotoxicity with

8 carboplatin and Gemcitabine, there is with

9 Cisplatin, but when you combine platinum and

10 Taxol, you have very significant neurotoxicity,

11 and this leads to very important toxicity for the

12 patient in terms of quality of life.

13 So since the onset of my work in

14 gynecological oncology, probably the most common

15 cancer I deal with is ovarian cancer, it is

16 sensitive to these drugs, but the overall

17 prognosis is still poor, and the five-year

18 survival is till around 38 percent for all

19 stages. And since 70 percent of the cases we see

20 are advanced disease to start with, it's

21 extremely important to be very aggressive

22 surgically and to be very aggressive with

23 chemotherapy up front, at the beginning.

24 I think it's totally wrong to treat a

25 patient and hold chemotherapy until the patient

00105

1 metastasizes. I mean, I feel just like

2 Shakespeare said in Hamlet: Diseases desperate

3 grown are by desperate appliance relieved, or not

4 at all. And I think you have to be aggressive

5 from the beginning. So my aggressiveness is

6 based on extensive radical surgery, tumor

7 reductive surgery down to an optimal level if

8 possible, and then treating the patient with the

9 drugs that have been found on the assay to be the

10 ideal combination, keeping in mind the goal of

11 curing the patient when possible, palliating them

12 always, and minimizing toxicity, which is

13 extremely important, because that's basically the

14 quality of life problems with hair loss,

15 neurotoxicity where they can't walk or pick

16 anything up, and not to mention bone marrow

17 toxicity and so on.

18 So I pick the safest combination that

19 looks most effective on the assay. I will

20 continue to do that until I quit practice. And I

21 now have, I think probably the largest series

22 that any one physician has in the country. It's

23 between 450 and 500 patients, and I am trying to

24 write that data up. And I know what it's going

25 to show, because I've already seen it in my

00106

1 patients. So, this is what I feel, and I think

2 this will be proven by further speakers.

3 Thank you very much.

4 DR. FERGUSON: Thank you very much.

5 MS. TILLMAN: Dr. Nalick, did you state

6 whether you were here on your own behalf?

7 DR. NALICK: Yes. I am. May I show

8 three slides?

9 DR. FERGUSON: Well, we have William

10 Grace and John Fruehauf in the next five minutes.

11 DR. NALICK: Okay.

12 DR. FERGUSON: Or we can just forego

13 the break.

14 DR. NALICK: I'll just tell you the

15 case without the slides. It will take me two

16 minutes. As an example, I had a patient who was

17 a gynecologic nurse, oncology nurse. She had

18 early ovarian cancer, treated at UCLA, had a

19 hysterectomy, had fairly decent tumor reductive

20 surgery. She received platinum and Taxol. She

21 had persistent disease. She had a second look

22 that showed persistent disease. After a long

23 period of time, she was finally accepted for a

24 bone marrow transplant. She received that at

25 UCLA, was in the hospital for almost two months

00107

1 with a bill of over $200,000. Her disease still

2 recurred.

3 I eventually saw her in 1997, opened

4 her abdomen. She has unresectable disease. She

5 was tested on the assay. She had a bowel

6 obstruction, had a colon resection. She was

7 found to be resistant to every single drug of 27

8 drugs, except synergy with Gemcitabine and

9 platinum. Dr. Weisenthal showed her slides. She

10 was treated with that combination, later had a

11 third operation, had only microscopic disease.

12 Had radioactive P-32 placed in the abdomen, and

13 then followed with six more courses of

14 Gemcitabine and platinum. She's now totally free

15 of disease. This is a woman who had stage four

16 disease, positive pleural effusions and

17 unresectable abdominal carcinomatosis. She's

18 free of disease, off chemotherapy for a year and

19 a half, and she's still working full time as an

20 oncology nurse.

21 DR. FERGUSON: Thank you. Ladies and

22 gentlemen, we have approximately 50 minutes for

23 six speakers. And I'm going to take the chair's

24 prerogative at this point to say okay, I would

25 like to just do them serially, and you can take a

00108

1 break as you need it. But then I'm going to

2 limit everybody to seven minutes apiece, unless

3 this group of people decides otherwise. So, I

4 would like to call William Grace.

5 DR. GRACE: I'm William Grace. For 24

6 years I was chief of cancer research and chief of

7 medical oncology at St. Vincent's. I am now full

8 time private practice. I have no financial

9 interest in any of the companies that are

10 represented here, but I have an enormous conflict

11 of interest, because Larry Weisenthal makes me

12 look good. As you know, I am in New York City,

13 where we have the world's greatest cancer

14 centers. And of course as you know, most people

15 don't fit into clinical trials, and in New York

16 City, probably only one in 20 gets into them. So

17 I see a lot of patients who have a lot of very

18 advanced disease, and one of the things that I

19 have done with my patients, and indeed in my

20 family, who have had cancer, I have used Larry

21 Weisenthal and other members here in order to

22 help manage their care.

23 What I can tell you is I could have

24 paraded in here a roomful of patients with

25 stories such as we've heard today from patients.

00109

1 And I am a strong believer in this technology,

2 and it's amazing. I have found that whenever I

3 present this technology to my patient, almost

4 none of them ever resent the cost involved; they

5 find the money somehow. It would be good for

6 some of my Medicaid patients and some of my

7 Medicare patients on fixed incomes if they could

8 have this technology available to them. And I

9 will tell you one thing; it likely wouldn't

10 reduce the cost, because I know many of my

11 colleagues who don't believe in this technology,

12 essentially chemo patients to death with one

13 chemotherapeutic combination after another, not

14 using this, but just going from one ASCO Journal

15 to another ASCO abstract to another, and boy,

16 that adds up to an awful lot of money. And if

17 you could predict that these patients would not

18 respond, you'd save patients a lot of grief and

19 you'd save patients a lot of money.

20 So, the only thing I'm going to do is

21 turn my remaining four minutes to somebody else,

22 because I am here on my own, I'm here because the

23 assay, one, makes me look good in a very

24 competitive market, and I believe that this

25 technology should be brought, and indeed the

00110

1 combinations that Larry has given us have been

2 applied to my pancreatic patients. And I have

3 become now a guru, along with Howard Bruckner, in

4 New York for pancreatic cancer, and we are all

5 using Larry's stuff to tell us what combinations

6 to use, and we're doing some exciting work in

7 pancreatic cancer. You just will not believe the

8 results when you finally see them.

9 Thank you.

10 DR. FERGUSON: Thank you very much.

11 Let's see. John Fruehauf?

12 DR. FRUEHAUF: It's a pleasure to be

13 here today. I am John Fruehauf. I'm a medical

14 oncologist. I have a conflict of interest. I am

15 the medical director for Oncotech.

16 And I began my career in oncology

17 really as a an M.D. Ph.D. student, and my Ph.D.

18 was in pharmacology, in Chicago. And I studied

19 the effects of BCNU on leukemia, using tritiated

20 thymidine as a measure for measuring cell

21 proliferation. I had learned that technique as a

22 co-step commissioned officer in a program at the

23 NCI where I spent three months in Dr. Herberman's

24 laboratory. And then went on to do my fellowship

25 in medical oncology after residency at the

00111

1 University of California, and my fellowship was

2 at the NCI. And as a first year fellow, I

3 participated in some of the small cell lung

4 cancer studies that were presented, where I was

5 treating patients based on these assays. So for

6 about the last 19 years, I have been involved in

7 this field. And as a practitioner at UC Irvine,

8 where I treat patients, I can see the value of

9 these technologies.

10 And I wanted to talk a little bit,

11 briefly go over the historical perspective of

12 where we came from to get to where we are today.

13 And I want to talk a little bit about the

14 clinical guidelines. The speaker for the FDA

15 talked a little bit about how we judge laboratory

16 testing, and I wanted to talk about the levels of

17 evidence that we use in decision making, and

18 summarize clinical data, and then present an

19 algorithm for how we can use this technology

20 clinically.

21 Now the principles that we employ in

22 cancer treatment are basically that most patient

23 are not likely to be cured. And so, we focus our

24 efforts on palliation, to prolong life rather

25 than cure patients. And we know there is

00112

1 morbidity related to our chemotherapy treatment.

2 People have talked about alopecia,

3 neurotoxicities, gastroenteritis, all sorts of

4 risks, and we have to balance these risks of

5 therapy with the benefits of modest life

6 prolongation for most patients.

7 So when we look at outcomes, we look at

8 survival, we look at improvement, in disease free

9 survival, complete response rates, we look at

10 cost effectiveness, but quality of life is really

11 one of the critical end points as an oncologist

12 who sees the patient sitting down in front of me,

13 how are you feeling today? Do you have

14 neuropathy? Should we change your chemotherapy?

15 Because if you aren't going to cure somebody,

16 your goal is to help them live a quality life

17 until they die. So this testing can be used to

18 do that, by eliminating ineffective therapy.

19 So, how successful are we in treating

20 medical oncology patients today? There's about a

21 million patients diagnosed a year. 64 percent of

22 these patients present with localized disease.

23 So there's a large bulk of patients, 44 percent

24 can be cured with surgery, 18 percent can be

25 cured with radiation therapy, but a dismal 2.4

00113

1 percent are cured with chemotherapy.

2 Chemotherapy, unfortunately, has not reached the

3 level of great success at this point in time.

4 And so we're struggling in research to find new

5 drugs and to do a better job.

6 If people present with metastatic

7 disease, again, 3 percent are cured. 5 to 6

8 percent will have remissions for two years; 15

9 percent can have a remission for a year; 76

10 percent have no, or minimal life prolongation.

11 So the vast bulk of patients we're treating with

12 chemotherapy are not particularly benefitting

13 from this. And so with that backdrop, we can

14 look at the value of tests to help avoid

15 ineffective therapy, where this is so common.

16 Now we all take an oath as physicians,

17 the Hippocratic oath, to keep patients from harm

18 and injustice. And I think it is an injustice to

19 treat people with drugs if you can figure out

20 ahead of time, they're really not going to

21 benefit the patient. And I tell my patients,

22 each patient I treat, what are the ground rules

23 that we want to do, use as a team in approaching

24 your disease? And we all agree, we never want to

25 make the treatment worse than the disease. And

00114

1 of course, giving ineffective therapy that's

2 toxic breaks this rule.

3 And Dr. DeVita, in the third edition in

4 his textbook, Principles and Practices of

5 Oncology, has stated that the most important

6 reason that people do fail in treatment is drug

7 resistance. So this is why in our laboratory, we

8 focused on drug resistance as an end point.

9 Now, this all started back in the '70s,

10 the 1870s, when Pasteur looked at chemotherapy,

11 and bacterial cultures. So this has a long

12 history in terms of in vitro testing. But more

13 -- as the technology evolved, the first studies

14 on cancer cultures were published in '54. In '56

15 it was found, very importantly, that agar could

16 selectively allow you to measure drug effects on

17 cancer. And this was really a breakthrough in

18 terms of then developing the clonogenic assay. I

19 think other speakers have pointed out, and will

20 point out that these basic clonogenic stem cell

21 assays were not effective. There were many

22 problems. Clump artifacts; it would take three

23 weeks to get an answer; you only got an answer

24 half of the time. This was not a good

25 technology.

00115

1 Advances were made, introduced, in

2 second generation, and then finally, third

3 generation technologies. And as Dr. Weisenthal

4 pointed out now today, we get answers 85 to 90

5 percent of the time within seven days, which is

6 relevant to having an impact in clinical

7 utility. So I think how long it takes, how often

8 you get an answer, was a great advance over the

9 original clonogenic assays.

10 So we want to talk about a little bit

11 today about how assays can predict response,

12 which is one end point, but also importantly, do

13 they relate to patient survival? Now, what are

14 the statistical requirements? And I'm just going

15 to say we will show about, talk about sensitivity

16 and specificity, predictive accuracy,

17 reproducibility, cost effectiveness, and that

18 these are the kinds of means you use to validate

19 a home brew test in a laboratory following CLIA

20 guidelines.

21 Now there are levels of evidence we

22 follow. Level one is metanalysis for prospective

23 studies, and Dr. Weisenthal has shown you a

24 metanalysis of sorts. Although most of the

25 studies were not necessarily randomized or

00116

1 controlled, it doesn't meet level two and level

2 three requirements, which is evidence obtained

3 from at least one well designed experimental

4 study. And I think critically, that these

5 studies are internally consistent. And level

6 three is evidence obtained from well designed

7 quasi-experimental studies such as non-randomized

8 controlled single group studies.

9 Because tests are not drugs, you don't

10 really compare Test A to Test B in a prospective

11 randomized study. You compare a test to

12 outcomes. Does the test predict an outcome. So

13 I think that actually, this kind of experimental

14 model fits testing, whether or not a laboratory

15 procedure can predict an outcome in the clinic.

16 And so, Grade B evidence is what we

17 want to look at here, because that's taking

18 levels of evidence type two and three, or four,

19 and showing that the findings are generally

20 consistent. And I think what Dr. Weisenthal has

21 reinforced is that if you look at all the studies

22 that have been done in the last 20 years, these

23 studies show very consistently that you can

24 identify ineffective agents.

25 Now, it's also been -- this is very

00117

1 difficult to read, I apologize, but over here

2 what this says is that class two and three

3 evidence is used to make decisions in the

4 treatment of breast cancer.

5 DR. FERGUSON: I'm giving you four more

6 minutes, because you were donated.

7 DR. FRUEHAUF: We've divided our time

8 up with the other people in our group.

9 DR. FERGUSON: Oh, you have?

10 DR. FRUEHAUF: Yes.

11 DR. FERGUSON: So you will be finished

12 by 11?

13 DR. FRUEHAUF: I was going to take

14 about 20 minutes; so what time did I start?

15 DR. FERGUSON: Okay. And so that Orr

16 and Hoffman, and Bosanquet, and David Alberts

17 will finish in --

18 DR. FRUEHAUF: They will take about ten

19 minutes each.

20 DR. FERGUSON: Well, it was about eight

21 the way I calculated before, so I mean, there are

22 four after you.

23 DR. FRUEHAUF: Well, I'm going pretty

24 fast here.

25 So, I want to talk about the data now,

00118

1 with that background. The cancer is a really big

2 problem. We can't really cure people very often,

3 and if we can identify drugs that don't work,

4 that can have great value. Did the data show

5 this?

6 Now here are correlations that we

7 published in the Principles and Practice of

8 Oncology. Dr. DeVita asked Dr. Bosanquet and

9 myself to write a review, you have that in your

10 packet, this table comes from that review, from

11 the textbook. And we looked at 4,263 cases

12 tested with a variety of technologies, and what

13 we found was that the predictive accuracy of

14 these tests was 90 percent or better for

15 predicting drug resistance, and about 72 percent

16 for predicting response. And the sensitivity and

17 specificity for this technology regardless of the

18 end points is very comparable, 85 and 80

19 percent.

20 Now how would that compare to tests we

21 use every day? In fact, the predictive accuracy

22 of drug response assays is comparable to hormone

23 receptor assays, and slightly better than

24 bacterial culture and sensitivity assays. And I

25 would submit, who would want to treat a breast

00119

1 cancer patient with Tamoxifen before getting an

2 ER or PR result? Or if you have a patient with a

3 refractory bacterial infection who's neutropenic

4 and they're not responding to primary empirical

5 therapy, almost every one of us will get cultures

6 to direct our therapy. So these are commonly

7 used tests that have comparable predictive

8 reliability to in vitro drug response.

9 Now we did a study of 450 cases.

10 Actually UCLA did a study, which is the

11 validation study for the technology we used in

12 our laboratory at Oncotech. And there were 332

13 colony end point assays, which at that time was a

14 gold standard, and the newly developed technology

15 was thymidine incorporation, and so there were

16 118 of these assays that were then compared to

17 the gold standard to make a determination of, can

18 the new technology give us similar results.

19 And without belaboring the point,

20 tumors are cultured in three dimensions, which is

21 important, as Dr. Weisenthal indicated. They are

22 cultured in drugs for five days, which means that

23 their exposure, in vitro drug exposure or

24 concentration times time, is going to be about

25 five to 20 times higher than you can achieve in a

00120

1 patient. And at the end of a three-day exposure

2 period, for the last two days to make five days,

3 treated thymidine is added and if the cells are

4 dividing they will incorporate thymidine and you

5 can measure that with a scintillation counter.

6 Now these tumors were cut out of

7 patients, shipped across town, put into a

8 culture, chopped into pieces, exposed to five

9 times higher exposures than you can give in

10 patients, and if they grew through that drug

11 exposure, what was the outcome clinically? What

12 we could see is that the colony end point was

13 very comparable to the thymidine end point. In

14 your handouts you can see and probably read more

15 legibly, that if you looked at the overall assay

16 predicted response probabilities, zero people

17 were responding in the assay if they had a below,

18 one standard deviation below the median result.

19 And this zero response rate was true whether they

20 got drug combinations containing different single

21 agents that were tested in the assay, and it was

22 true across different tumor types as well. So

23 there was a robust quality to showing people

24 didn't respond, zero response basically, except

25 for one responder with this cut point of one

00121

1 standard deviation below the median. So this was

2 chosen then, to evaluate patients further.

3 And this is showing that overall, it

4 didn't matter what drug was chosen or what tumor

5 type was evaluated, that this end point was

6 true. So here if we're looking at different

7 drugs, basically very few people responded. This

8 is the same patient here in the extreme

9 resistance group. Responding didn't matter what

10 the drugs were or the tumor type.

11 And this is a summary overall, and the

12 black dots are the thymidine end point, the open

13 circles are clonogenic assays, and we can see

14 that they're all clustered together. They were

15 found to correlate directly. So this was a way

16 of validating thymidine compared to what had been

17 an important end point, the colony forming

18 assay. And this is the threshold that was chosen

19 as the resistance end point that shows that only

20 one out of 127 patients responded clinically if

21 their tumor fell in that category.

22 So Bayes' Theorem, I won't go over this

23 in detail because Dr. Weisenthal did, but you

24 have a pretest result, you do a test, and you can

25 assign a post-test result now to the patient's

00122

1 probability of response. And this is the

2 prediction based theorem showing that in this

3 study at UCLA, that the patient results fell on

4 those predicted Bayesian lines, and the pretest

5 probability was then altered by the test to give

6 a post-test probability.

7 And here we can see, if you're an

8 extreme resistance category, your post-test

9 probability is going to be significantly lower

10 than your pretest probability, whereas if you're

11 in the low resistance category, it's the

12 opposite.

13 So we looked at a number of patients

14 who for Taxol, to compare because of the

15 reasoning that if you're over 65, are you going

16 to be different than you're under 65. And I

17 think Medicare would be concerned that there

18 might be differences in age groups for results in

19 the assay. We found no difference if they were

20 less than 65 or more than 65 in terms of the

21 frequency of extreme resistance to Taxol, which

22 is commonly used in breast cancer.

23 Now, another way of validating a test

24 is to determine if the end point your test

25 measures can be confirmed by a second end point.

00123

1 So we looked at peglotical protein expression, we

2 published this in Clinical Cancer Research, where

3 the degree of peglotical protein increasing on

4 the tumor correlated directly with decreased

5 response to Taxol in the assay. So we took a

6 known mechanism and correlated it with the assay

7 result, and found a direct correlation.

8 Adreomyecin did not correlate as well, because

9 there were multiple mechanisms of resistance for

10 adreomyecin.

11 And I think this is a very important

12 point to emphasize, that we can't go out and

13 measure specific mechanisms, because cells are

14 very complicated. So what an in vitro test does

15 is it takes all the mechanisms working in situant

16 cells, grown as little clumps, to recapitulate

17 the in vivo growth. And it shows that multiple

18 mechanisms can be integrated into a net effect

19 result that correlates with clinical response.

20 And I don't know of any patients who if they

21 don't respond to chemotherapy, are going to do

22 well. The only patients to do well are the ones

23 that do respond.

24 So this is just briefly then, to close,

25 a study that we have done that was peer reviewed

00124

1 by the ASCO committee and presented at the ASCO

2 meeting last year, where we looked at breast

3 cancer survival as another end point, and

4 compared survival of 96 patients who were tested

5 in the assay, who received chemotherapy with

6 their EDR scores, nodal status, and clinical

7 stage. And if they were resistant to two drugs,

8 they were given a score of 0. If they had low

9 drug resistance to both drugs, they were given a

10 score of 4, and so forth in intermediate

11 categories. And just briefly, I'll say there

12 were no statistical differences between the

13 groups who were resistant and sensitive in the

14 assay in terms of stage, lymph node status, tumor

15 size and so on.

16 They were also treated evenly in terms

17 of hormonal therapy, mastectomy versus

18 lumpectomy, and chemotherapy agents that were

19 chosen to treat them. This was a blinded study.

20 The patients were treated with chemotherapy

21 empirically, and the outcome and survival was

22 compared to the assay result. We found in

23 multivaried analysis, progression free survival

24 was significantly worse. The relative risk of

25 progression was 2.9 fold higher for patients who

00125

1 were resistant versus low resistance, and that

2 was similar to the poor prognosis conferred by

3 high nodal status greater than 10, or stage four

4 versus stage one. And similarly in overall

5 survival, there was a significantly worse

6 survival if you had any resistance in the assay.

7 And this shows the progression free survival

8 curves, which were significantly different, and

9 the overall survival difference. For patients

10 with low resistance in the assay versus

11 whatsoever to the drugs they received.

12 Now there are many other survival

13 studies that have been done, smaller studies, but

14 again, the majority are internally consistent for

15 showing significant differences in survival where

16 test resistant people survived for significantly

17 less time than patients who got drugs to which

18 they were sensitive in vitro. This is a more

19 recent summary of survival correlations, where in

20 these studies with over 300 cases, 400 cases,

21 where different people looked at different tumor

22 types, it shows significant survival advantages

23 to receiving drugs that were not resistant

24 compared to drugs that were resistant.

25 So I think that the end point of

00126

1 survival has been addressed, correlations between

2 the assay technology and other validating methods

3 have been addressed, so I believe that if these

4 tests are used by normal incorporation into the

5 routine practice, you get a biopsy, you look at

6 the diagnostic information from pathology, you

7 look at prognostic markers, you look at drug

8 resistance information, staging information, and

9 then planning goes on between the physician and

10 the patient, based on integrating this

11 information together.

12 So in vitro assays do correspond to

13 response with specificity and sensitivity that

14 are adequate and comparable to other clinical

15 tests. Survival is significantly associated with

16 in vitro response. Assay directed therapy has

17 improved outcomes. And other people talked about

18 cost. You've heard about survival. So I believe

19 that levels of evidence two and three have been

20 met, the standard criteria for covering these

21 kind of things, and this is how we approach tests

22 in medical oncology.

23 So, because there isn't time for

24 questions, I will conclude at this point. Thank

25 you.

00127

1 DR. FERGUSON: Thank you. We have

2 actually 25 minutes now, for four presentations,

3 Dr. Orr, Dr. Hoffman, Dr. Bosanquet, and David

4 Alberts. And that's, you know, seven minutes or

5 so apiece, and I guess since the presenters -- I

6 guess I'm going to have to crack the whip more

7 than I have. So, Dr. Orr? I am going to hold

8 people to seven minutes this time.

9 DR. ORR: I am Jimmy Orr. I am a

10 gynecologic oncologist in private practice in GYN

11 oncology in Fort Meyers, Florida. I currently am

12 a clinical professor at the University of South

13 Florida. And I do have a conflict of interest,

14 in that some of the data that I will present

15 today was supported by Oncotech, some of the peer

16 review data.

17 If one talks about therapeutics and the

18 roles of therapeutics, we've already alluded to

19 Robert Lowell's rules, who was a barred professor

20 of medicine at Columbia. And all of these apply,

21 I think, to the treatment of patients with

22 cancer. If what you're doing is good, keep doing

23 it. If what you're doing is not good, stop doing

24 it. If you don't know what to do, do nothing.

25 And finally, never make the treatment worse than

00128

1 the disease. And each of these apply, I think,

2 to the treatment of women with gynecologic

3 cancer.

4 As far as peer review evidence as it

5 relates to the use of drug resistance assays in

6 women with ovarian cancer, and I will remind you

7 that they comprise about 25,000 cases a year in

8 this country, and the median age is about 64.

9 Two years ago Roswell Park presented some data

10 looking at drug resistance assays, and I think

11 two very important aspects of that paper need to

12 be emphasized.

13 Number one, in looking at patients with

14 extreme drug resistance assay, if you compare

15 those patients who had no extreme drug resistance

16 to platinum and Taxol, or extreme drug resistance

17 to platinum and Taxol, it took two very important

18 end points, complete surgical response and

19 progressive disease, one can see that the

20 presence of EDR to platinum and Taxol halved the

21 complete surgical response. And that becomes

22 extremely important, because if you look at

23 complete surgical responders, survival is clearly

24 different from those who were found at second

25 look to have persistent disease.

00129

1 And the incidence of progressive

2 disease during initial treatment was almost

3 doubled. In a current abstract that has been

4 submitted, when one looks at extreme platinum

5 resistance, and looks at progression free

6 survival, one can see that patients who are

7 extreme platinum resistant have roughly half of

8 the progression free survival, and roughly half

9 of the estimated five-year survival, both being

10 statistically and certainly clinically

11 significant.

12 I would like to address for the

13 remaining three to four moments, is this test

14 cost effective. In a recent peer review journal

15 article we submitted, the Cancer Journal, we

16 evaluated the cost effective treatment of women

17 with advanced ovarian cancer by cytoreductive

18 surgery and chemotherapy directed by an in vitro

19 assay for drug resistance. All patients received

20 cisplatin but the second drug in combination was

21 guided by the results of the extreme drug

22 resistance assay. That is, platinum with Taxol,

23 or platinum with cyclophosphamide, as guided by

24 the assay results. As one knows and understands,

25 there are significant incidents of extreme drug

00130

1 resistance across the board of patients who have

2 up front treatment with ovarian cancer, in the

3 neighborhood of 35 percent for platinum, 20

4 percent for Citoxan, 15 to 20 percent for

5 carboplatin and cisplatin. So the incidents of

6 drug resistance is very common in patients who

7 have not received previous treatment.

8 Our overall survival was 66 percent at

9 three years. There was no significant difference

10 between patients treated with Taxol carboplatin

11 and platinum in Citoxan. If one looks at the

12 average cost to the bottom line, the average

13 total drug cost to the average drug cost per

14 patient, and then adds in the average

15 chemotherapy cost per patient with or without the

16 assay, and then sorts out treatment related

17 results as far as drug cost per patient, assay

18 cost per patient and treatment cost per patient,

19 we can see that assay directed therapy appeared

20 very favorably in comparison to the standard as

21 most would say today, of Taxol and platinum. And

22 more importantly, the cost effectiveness per

23 patient, that is, taking the cost divided by the

24 overall survival, also appeared very important.

25 Assay drug related treatment can be

00131

1 used cost effectively in a significant number of

2 women with gynecological malignancy, and

3 particularly in those women with ovarian cancer.

4 Assay guided therapy can improve their survival

5 and is clearly prognostic.

6 Thank you very much.

7 DR. FERGUSON: Thank you very much, Dr.

8 Orr. Next is Robert Hoffman.

9 DR. HOFFMAN: My name is Robert

10 Hoffman. I am the founder and president of

11 AntiCancer, Inc., who sent me here. I am a

12 graduate of Harvard University, where I did my

13 Ph.D. in cell biology. I trained in tissue

14 culture at Massachusetts General Hospital with

15 John Littlefield, and I have been in this field

16 for approximately 30 years.

17 We have developed what we call the

18 histoculture drug response assay, which is based

19 on three dimensional culture of cancer tissues,

20 and I'd like to introduce you to this assay, if I

21 may. The emphasis is on three dimensional

22 culture that preserves the tissue structure of

23 the cancer during the culture. The evaluability

24 rate is approximately 95 percent. Mono and

25 combination chemotherapy can be evaluated.

00132

1 There's correlation with sensitivity, correlation

2 with resistance, correlation with survival, and

3 increased survival has been found for assay

4 directed therapy.

5 This is just an example of a stomach

6 tumor. This is what it looked like coming out of

7 the patient, looked like after two weeks in

8 histoculture. So, the key point here is

9 preservation of tumor structure and tumor

10 physiology.

11 To give you an idea of how the test

12 works, this is just one example. This is human

13 breast cancer. Sensitivity to Doxorubicin

14 measured with the MTT end point. And here we

15 have tissue culture plates; these are sponge gels

16 on which pieces of tumor are cultured, and we

17 have increasing concentrations of Doxorubicin,

18 and you can see just from the no treatment, the

19 dark staining MTT, to treatment at 29 micrograms

20 per ml, where there is no staining at all. And

21 you see a gradation over the increasing

22 concentration. So this gives you an idea of how

23 the assay works.

24 Here is an example of breast cancer

25 with a front line therapy, adreomyecin or

00133

1 Doxorubicin, and we were measuring the percent of

2 dividing cells, in this case by a thymidine

3 assay. And the point of this is that there is

4 approximately two to three orders of magnitude in

5 sensitivity over a series of patient tumors. So

6 empirical therapy, I believe, does not give us

7 very valuable information.

8 This is a study we published a few

9 years ago on the clinical applications of the

10 HDRA or histoculture drug response assay,

11 published in Clinical Cancer Research, where we

12 emphasized correlation with survival.

13 And these are studies done on gastric

14 cancer patients treated with mitomycin C and five

15 fluorouracil, using the MTT end point. And we

16 just, this is survival and this is disease free

17 survival, and these are patients that were HDRA

18 or histoculture drug response assay sensitive,

19 surviving a considerable period of time. These

20 were patients that were resistant in the assay to

21 these drugs, mitomycin C and 5-FU. And the

22 recurrence free survival has a very similar

23 curve. In other words, the HDRA sensitivity

24 correlated with increased survival.

25 We did further studies on the survival

00134

1 with the HDRA in a 46 center study, also

2 published in Clinical Cancer Research, and I'd

3 like to share these data with you, some examples

4 of survival. These are gastric cancer patients,

5 high stage, three and four, and the patients that

6 were sensitive in the assay are, have a

7 considerably long survival. The patients that

8 were resistant in the assay, in this case to

9 mitomycin C and UFT, which is a 5 fluorouracil

10 derivative. So again, correlating with survival.

11 In a subsequent study that's not yet

12 published, we correlated survival in the

13 histoculture drug response assay with response to

14 mitomycin C. And all the patients here were

15 treated with mitomycin C. Survival again, here

16 are the patients that are sensitive in the assay,

17 surviving significantly longer than the patients

18 who were resistant. HDRA resistant patients

19 here, HDRA sensitive patients here. So there is

20 a statistically significant difference in

21 survival between HDRA sensitive and HDRA

22 resistant, as there have been in all the studies

23 I've shown you thus far.

24 Here I would like to show you a

25 correlation between assay directed therapy and

00135

1 clinician's choice, using the histoculture drug

2 response assay for survival. And this is with

3 gastric cancer patients, and here are the

4 patients with, that are directed by the assay,

5 assay directed therapy, and there are a series of

6 12 patients here, and this is their survival for

7 a period of 18 months or so, and here are the

8 patients that were resistant in the histoculture,

9 in the HDRA, and treated by clinician's choice,

10 and you see their survival here. And the average

11 survival for the HDRA or the assay directed

12 patients who were HDRA sensitive was 9.8 months,

13 compared to 4.7 months for those who were HDRA

14 resistant and treated by clinician's choice.

15 So what I -- we've published over 40

16 papers on the HDRA and what I've summarized for

17 you here are studies with survival, which we

18 consider to be the ultimate end point. And we

19 have shown that the HDRA not only correlates with

20 survival, but even in a prospective study, assay

21 directed therapy can seemingly increase

22 survival.

23 Thank you very much.

24 DR. FERGUSON: Thank you, also for

25 staying on time. Next, Andrew Bosanquet?

00136

1 DR. BOSANQUET: Thank you very much. I

2 am Dr. Andrew Bosanquet, from the Bath Cancer

3 Research Unit, in England. I've come here by the

4 invitation of Larry Weisenthal. Because I run a

5 small charity, my fare is being paid for by

6 Oncotech. At Bath Cancer Research, I want to

7 show you some of the work that we published in

8 the last year with the DiSC assay that was

9 proposed by, sort of invented by Larry

10 Weisenthal. In 1991 we published this survival

11 curve in chronic lymphocytic leukemia patients.

12 We have worked very much with chronic lymphocytic

13 leukemia and most of the results that we've got

14 are with this disease, and all the results that I

15 am presenting now are with chronic lymphocytic

16 leukemia.

17 So here you see in the 1991 work,

18 patients who were sensitive to the drugs they

19 received survived longer than those who were

20 resistant to the drugs that they received. In

21 this latest paper that we published in the

22 British Journal of Hematology this summer, we are

23 looking at fludarabine, and we used the DiSC

24 assay to determine drug sensitivity to

25 fludarabine. Patients were treated independently

00137

1 and then we compared the results. The treatment

2 of patients was either labeled with fludarabine,

3 i.e., fludarabine was given within the first year

4 of the test being done, after the first test

5 being done, or with any other chemotherapy. And

6 the point was that no fludarabine should be

7 given.

8 We have 243 patients who came into this

9 study. Those who received fludarabine versus

10 those who received other chemotherapy, these are

11 the numbers, very similar age and stage, sex

12 ratio, relatively similar previous chemotherapy.

13 The response to chemotherapy, which is one of the

14 end points that we've looked at often, in the

15 test sensitive patients was around 18 percent and

16 in the test resistant patients was a zero, or one

17 patient out of 15 responded, but just for a short

18 time. Very significant difference in response.

19 But survival is the important thing.

20 Here, of patients who received fludaribine, here

21 is their survival if they were test sensitive,

22 here is their survival is they were test

23 resistant. Now in chronic lymphocytic leukemia,

24 you would expect a survival of four or five

25 years, and with fludarabine, fludarabine is the

00138

1 modern drug to use for this disease, and so you

2 would expect patients to have this sort of

3 survival. But notice these patients; they have

4 been given the best drug for this disease, they

5 are expected to survive for five years, and they

6 were all dead by 17 months. These patients we

7 looked at in some detail, the 15 of them. They

8 were too sick, having received fludarabine to

9 which they didn't respond, they were too sick

10 then to receive any other chemotherapy. If they

11 had only received some other chemotherapy first,

12 other than the best drug, what is considered the

13 best drug, they could have survived longer.

14 Was it that these patients had a poor

15 stage and so on? The answer is no. These are

16 the same two curves divided into those without

17 any previous treatment, and there is still this

18 same difference in survival in those who had

19 received previous treatment. Stage and sex and

20 so on were very similar between those two.

21 Now this is the same 15 patients who

22 died very soon after receiving fludarabine, but

23 note, these are all fludarabine test resistant

24 patients, and this line is a line of patients who

25 received any other chemotherapy. It didn't

00139

1 matter what they had, as long as it was

2 chemotherapy within one year of the test. It

3 wasn't assay directed even, it was just they did

4 not receive fludarabine, they did not receive the

5 best drug. And they survived longer, because

6 these patients were test resistant to the best

7 drug.

8 So here we see patients surviving

9 longer even though they are not chosen by the

10 test result. And if we look at these two groups

11 of patients, again, there's very similar

12 characteristics, the age, stage, and sex, and

13 previous chemotherapy that constitute parameters

14 that are important in the treatment of CLL. But

15 actually you will see that they are actually very

16 sensitive to other CLL drugs. Both groups of

17 patients, 80 percent of them were sensitive to

18 other drugs, whether it be prednisolone,

19 doxorubicin, pentostatin, vincristine.

20 And so as a result of these and other

21 experiments that we've performed, the DiSC assay

22 is part of the second randomization in the U.K.'s

23 national medical research counsel, CLL 4 trial.

24 And the second randomization is to treatment

25 guided by the DiSC assay, versus treatment guided

00140

1 by protocol, which is essentially physician

2 choice.

3 So in five years time, I hope we can

4 have a result from that for you, which will be,

5 this is going to have 500 patients entering into

6 it. This will be a good robust study using

7 randomized control trial.

8 Very briefly, we did test a very

9 similar drug, calatropin. 34 patients were

10 treated. We did a concurrent DiSC assay, so this

11 was a prospective study looking at -- patient

12 characteristics, I won't go into, and here you

13 say see the test results, raw data on the

14 left-hand side on whether the patients received a

15 complete, partial or no response. And as you

16 see, those who responded had low test results,

17 i.e., high sensitivity, and those who did not

18 respond had a very resistant test, apart from

19 these two who were withdrawn drawn early.

20 Just briefly on the economics of it,

21 here are a set of CLL patients who were either

22 given drugs to which they were resistant to, and

23 every other test, drug that we tested, they were

24 also resistant to, so we couldn't expect to do

25 very much for these patients. Here is a survival

00141

1 of patients who were given drugs to which they

2 were sensitive to, and here is the survival of

3 patients who were given drugs to which they were

4 resistant, but they had drugs in the test to

5 which they were sensitive. They should have

6 gotten drugs they were sensitive to, and survived

7 along this line. And if we work out, if we look

8 at the data on this, this was very significant,

9 this unused sensitivity we called it, where they

10 could have had better treatment.

11 And the cost per life year gained, if

12 we'd used DiSC assay guided treatment there, over

13 all the patients tested, not just that group, but

14 over all the patients tested, was 1,500 pounds,

15 or $2,500, and this value of $2,500 compares with

16 the cost of treating CLL patients per life year

17 gained, which is enormous, compared to the cost

18 of extending the life for a year in patients by

19 using a drug sensitivity test.

20 DR. FERGUSON: Thank you very much, Dr.

21 Bosanquet. Do you have handouts for some of this

22 material?

23 DR. BOSANQUET: We do have copies of

24 the three papers to which I referred, and we can

25 give them to you.

00142

1 DR. FERGUSON: Now, Dr. Alberts, is

2 Dr. Alberts here?

3 DR. ALBERTS: Yes.

4 DR. FERGUSON: Okay.

5 DR. ALBERTS: I am here on my own

6 recognizance. I am a professor of medicine and

7 pharmacology and public health at the University

8 of Arizona, associate dean for research in the

9 College of Medicine. And in terms of my

10 experience, I am the chair of the gynecologic

11 cancer committee for the Southwest Oncology

12 Group, and have been since 1977. I also chair

13 the cancer prevention and control committee in

14 the Gynecologic Oncology Group.

15 I came to the University of Arizona

16 in 1975 to help Dr. Sid Salmon develop an assay

17 that could individualize chemotherapy for

18 patients with a broad variety of tumors. I must

19 say that it's a sad note that I'm here today,

20 because Dr. Salmon died October 6th, of

21 pancreatic cancer. But I think his spirit is

22 here and in fact, he pretty much developed this

23 field.

24 I will point out that since 1975, the

25 options, the possibilities for treating ovarian

00143

1 cancer are tremendously increased, that's my

2 expertise, ovarian cancer, and in fact, it's a

3 very confusing field. People might want to make

4 you think that it's a simple field in terms of

5 selecting agents. There are 22 drugs that are

6 FDA approved that have activity for ovarian

7 cancer, 11 of them are specifically approved for

8 ovarian cancer, and it is absolute chaos

9 certainly in the second line treatment of these

10 patients to determine what drug should be used

11 for any one patient. And I can assure you that

12 physicians are not infallible in this situation.

13 On the other hand, I think what you've heard

14 today is that the tests that we have available to

15 us can lead us out of the wilderness in

16 relationship to these problems.

17 Now I think, I am very impressed with

18 the presentations today. I mean, I know where I

19 would vote on this. There is acceptable quality

20 control and reproducibility, acceptable accuracy,

21 and acceptable clinical utility of these tests,

22 and this has been shown over and over and over

23 again, for a variety of tumor types. We know

24 that drugs that don't work don't help people.

25 And certainly, if we can identify at least those

00144

1 drugs that are not active, and not give them to

2 patients, we're not going to at least harm those

3 patients. Giving inactive drugs to patients is

4 harmful, and it's cost ineffective.

5 I think Mr. Stein very eloquently

6 pointed out that if a patient was given the

7 opportunity to really understand what the options

8 were, there is no question that they would want

9 to be treated according to the best knowledge

10 that existed for them on the basis of their

11 tumor. There are always questions about risk

12 benefit concerns, palliation, quality of life,

13 and when you have assays that are 99 percent

14 accurate in identifying inactive drugs, we've got

15 to be serious about taking these results to

16 heart.

17 In this very same city, just, I think

18 it was just exactly a year ago, actually not even

19 that, nine months ago, at Mercy Hospital, I

20 participated in a symposium, a gynecologic cancer

21 symposium, and I was asked to speak on this:

22 Drug resistance assays, when possible, should be

23 used to guide primary therapy of ovarian cancer,

24 and I was asked to speak on the pro side. It was

25 sort of a randomized way in which speakers were

00145

1 selected.

2 I finished this talk asking the

3 audience, which were, there were about a hundred

4 gynecologic oncologists in the audience, if

5 you're sitting in your office and you have

6 specific information on the tumor of the patient

7 that you're treating that shows that nine out of

8 those ten drugs in the second line treatment are

9 associated with extreme drug resistance, and one

10 of these drugs is associated with sensitivity,

11 and you're going to see that patient in five

12 minutes, would you choose to look at that data,

13 would you be interested in that data, or would

14 you like to avoid that data? A hundred percent

15 of the people said of course, they would take

16 into consideration the data that were presented

17 in the laboratory report from a valid lab. I won

18 the debate, by the way.

19 Well, I just want to sum up. I'm sort

20 of the summary speaker here. Drs. Weisenthal,

21 Hoffman and Bosanquet have all shown survival

22 advantage for test sensitive versus test

23 resistant drugs. Dr. Fruehauf pointed out to the

24 accuracy of the extreme drug resistance assay, 99

25 percent to identify ineffective drugs, and that

00146

1 test results are valid across a whole variety of

2 tumor types and drug classes, and finished with a

3 discussion of poor breast survival with test

4 resistant drugs. Dr. Hoffman talked about GI

5 cancer survival being poor with inactive agents,

6 and I think that's really impressive, with

7 gastric cancer especially. And finally, Dr.

8 Bosanquet's presentation that you've just heard,

9 showing again, poor survival with chemoresistant

10 disease.

11 And I just -- I am not going to belabor

12 this any further. I am going to give you my

13 conclusion slide. In vitro drug response assays

14 for cancer specimens have definitely matured with

15 third generation technologies. I was involved

16 with the first generation and I think it's

17 extremely interesting that Dr. Salmon's own human

18 tumor stem cell lab has been converted completed

19 to using tritiated thymidine end point assays for

20 solid tumors. The accuracy, sensitivity,

21 specificity are excellent, and comparable to

22 conventional testing. Results apply to both

23 first line and salvage settings. Cancer is still

24 primarily incurable but many new agents are

25 available with activity. Their selection must

00147

1 be, not just should be, guided by data, but not

2 gut feelings. And unfortunately in oncology

3 today, and I think you're all aware of it, gut

4 feelings are too pervasive in our selection of

5 treatment. These tests should be covered by

6 HCFA.

7 Thank you very much.

8 DR. FERGUSON: Thank you. I hope the

9 panel and the audience and participants will

10 excuse this sort of marathon session, but I think

11 people need to have their say, to present their

12 information.

13 We will go right on to Dr. Nagourney,

14 and I think these next four presentations are

15 going to be allotted 15 minutes each, and we will

16 be just a few minutes late for lunch.

17 DR. NAGOURNEY: First of all, I'd like

18 to introduce myself. I am Robert Nagourney. I

19 am a hematologist oncologist and also founder of

20 Rational Therapeutics, which is a laboratory

21 which applies cell death end points in the study

22 of human tumor biology. I am here by virtue of

23 United Air Lines frequent flier miles, and I paid

24 for my own hotel room.

25 I'm going to try to put this in a

00148

1 slightly different context, as I present. I

2 would like to look at some of the scientific

3 issues that have led me to certain conclusions

4 and perhaps will bring you some of the insights I

5 have gained over about 18 or 20 years of work in

6 this field.

7 I'd like to start off by saying that

8 good medicine always follows good science. When

9 we think, and you have heard a review of the

10 various techniques that have been applied in

11 primary culture studies going back to 1954

12 through more recent techniques. Various

13 investigators have looked at this area and have

14 tried to cess out mechanisms by which they can

15 assess responsiveness in individual patients

16 based on findings in a laboratory. The

17 underpinnings of this might be described by the

18 equation biomass equals cell growth minus cell

19 death. This equation has been primarily examined

20 when biomass grew in a tumor as a function of

21 cell growth. But I think we are now witnessing a

22 change in that thinking and the focus is now

23 shifted to cell death events.

24 So I'd like to discuss a little bit

25 about advances in our understanding of tumor

00149

1 biology, the concept of cell proliferation end

2 points versus cell death or apoptotic end points,

3 and how they may help us to decipher some of the

4 data. It was said by John Reed in a recent

5 editorial in Journal of Clinical Oncology,

6 October '99, that essentially all traditional

7 anticancer drugs use apoptosis pathways to exert

8 their cytotoxic actions. Thus, drugs that were

9 largely developed for cell growth inhibition and

10 other purposes perhaps really act through

11 different mechanisms. In addition, it is

12 difficult to measure these events using cell

13 growth end points, which was pointed out actually

14 earlier by Shakespeare, who said that the absence

15 of proliferation doth not apoptosis make. That's

16 not actually a true quote, although it sounds

17 like it could be. Don't look for it in the Henry

18 cycle.

19 In any case, what we've really focused

20 on is the concept of an apoptotic event, the

21 induction of cell death in vitro as a predictor

22 of outcome. When Isaac Newton was asked how he

23 discovered gravity, he said by thinking upon it

24 continuously. And between 1990 and 1995, as an

25 in residence faculty for UC Irvine, I really

00150

1 thought on continually what it was that

2 constituted cell death events in a test tube.

3 What could you measure that might allow you to

4 predict a patient's outcome based upon apoptosis

5 rather than growth inhibition?

6 My first stab at this effort was to try

7 to apply the morphologic changes described in

8 1972 by Kirwelli and Curry in their original

9 paper, in this dosed response curve from cells in

10 culture all the way down to the apoptotic and

11 shrunken cells morphologically characteristic of

12 apoptosis. We tried to apply the DNA degradation

13 profiles known as the 180 KBP DNA degradation

14 ladders. We found that to be a relatively

15 difficult method to use, and in fact was not

16 predictive, because ladder profiles are not

17 actually predictive of human tissue in primary

18 culture, but more of a cell line phenomenon.

19 We then examined a different end point,

20 which is the inverted field gel electrophoresis

21 which uses a 50 KBP DNA degradation, and we were

22 able to show that this indeed did correlate with

23 responses in some of these profiles where

24 patients had very excellent responses. This is a

25 gel, inverted field gel of patient's tissues

00151

1 studied following drug exposure, and looking for

2 the 18 hour DNA degradation profile. However,

3 again, this can only be applied in pure

4 cultures.

5 In addition, a variety of papers were

6 showing that some of the end points using DNA

7 labeling, such as the insulin tunnel end points,

8 did not reliably identify apoptotic cell death.

9 We moved from that area then, and during the same

10 time, into membrane perturbations, alterations of

11 mitochondrial function and membrane potentials

12 that might enable us to predict responses based

13 on what was occurring at the metabolic level.

14 And although again, these were very interesting

15 findings, they did not apply broadly to human

16 tissues, because you needed pure cultures, and it

17 was only applicable to, in our studies, the

18 chronic lymphocytic leukemia and leukemias in

19 general.

20 Finally, we moved to some of the DNA

21 markers, some of the mutational events that lead

22 to cancers as modulators of apoptosis, and the

23 morphologic events at the bottom can be

24 characterized by the interplay between positive

25 and negative modulators of apoptosis. We are

00152

1 currently completing and have submitted a study

2 on the correlation between BCL XL overexpression

3 and drug resistance in human primary cultures.

4 However, again, this is a system that can only be

5 applied in very pure cultures, and did not allow

6 us a practical and high throughput end point.

7 All of the assays that we have been

8 interested in are what I would describe apoptosis

9 based or cell death based, and these have been

10 described so far for you as DiSC, or the one that

11 we applied which is a modification thereof, the

12 apoptotic, MTT, ATP, FMCA, and others. My

13 principal work has been with the differential

14 staining technique, which takes cells in culture

15 for three days, and then assess the

16 responsiveness of the tissues based on the

17 ability to induce apoptosis morphologically and

18 metabolically.

19 Now, as a comparator, I thought it

20 might be of use to look at an older technique,

21 the soft agar cloning in a preclinical setting

22 and compare it with a cell death end point for a

23 drug that we now know is used in the treatment of

24 ovarian cancer. In 1992 a study was published

25 using the soft agar assay, where they examined

00153

1 the ability of a drug, Topotecan, to induce cell

2 inhibition or growth inhibition culture, and they

3 showed under these conditions of culture that 83

4 percent of renal cell cancers were sensitive,

5 leading to a clinical trial conducted at Sloan

6 Kettering by Dr. Ilsin, in which 15 patients

7 received Topotecan for renal cell carcinoma,

8 without single response.

9 At about the same time, we were

10 applying the cell death measures in a laboratory

11 setting to the same drug under similar

12 conditions. And we found quite surprisingly, in

13 1994 and presented in 1995, that ovarian cancers

14 appeared to be a particularly good target. Two

15 years later, and three years later, several

16 published studies revealed the same, and the

17 observation led to an FDA indication for that

18 drug. So in a growth cell assay, a very

19 erroneous result was predicted, whereas in a cell

20 death assay, a more robust end point, a very

21 accurate prediction.

22 In point of fact, these observations

23 have continued through the years that I've

24 applied this laboratory test to a variety of

25 observations. Starting with alpha interferon

00154

1 synergy, which was subsequently proven in 1996,

2 12 years later, by Wadler, et al., to show that

3 this was occurring by virtue of up regulation of

4 thymidine phosphorylase. The original

5 observation of chloradioxydenasene's activity in

6 hairy cell leukemia was conducted in my

7 laboratory at Scripp's Clinic, subsequently

8 proven by Larry Purot, published in the New

9 England Journal of Medicine, providing a 90

10 percent complete remission rate in hairy cell

11 leukemia. Our observations of pure synergy using

12 a cell death end point, have led to a point where

13 Howard Hockster has reported with ECOG a 100

14 percent response rate with fludarabine plus

15 Citoxan in patients who received a combination

16 really which was found years earlier in our

17 laboratory. Our observation that

18 chloradioxydenasene in blastic CML was

19 subsequently confirmed in a clinical trial

20 showing a 47 percent response rate in patients

21 with blastic CML.

22 An observation which I'm particularly

23 proud of, and one which you've heard consistently

24 and which I think functions as a perfect example

25 of the discriminating and robust nature of this

00155

1 assay was my observation in the laboratory

2 between 1992 and 1995, reported originally in

3 1995, of the true synergy between Gemcitabine and

4 Cisplatin. When we were first provided this

5 drug, LY 1808, that drug didn't even have a

6 name. We began to study it, and found that there

7 was an enormous amount of synergy. That has now

8 been confirmed, as you have heard repeatedly, in

9 a variety of studies, having been approved for

10 FDA indication in non-small cell lung cancer.

11 However, ovarian, breast, and other diseases are

12 rapidly showing the same data that we have

13 generated back as far as four or five years ago.

14 And finally, as I've mentioned, the

15 Topotecan data. I would like to use the

16 Gemcitabine data which you have seen, and so

17 eloquently provided by Dr. Nalick and also by the

18 gentleman who had presented with pancreatic

19 cancer, Randy Stein, who was the beneficiary

20 through Dr. Weisenthal's laboratory of this

21 observation.

22 When I describe this interaction, I

23 would like to point out that if you use the

24 laboratory test as the only indicator of where

25 the FDA should find use of this, you would find

00156

1 that bladder carcinoma, which has now been

2 published, to provide 60 and 70 percent response

3 rates, with substantial numbers of complete

4 remissions. Bladder cancer is one of the best

5 candidates for this combination. Ovarian cancer

6 is the second best when you use a gradation of

7 IC-50, median IC-50 as the determinant, ovarian

8 cancer would be your second choice.

9 Interestingly, sarcoma, non-small cell

10 lung cancer, for which there is an indication,

11 and interestingly breast cancer, which I will

12 show you some data on, the only data actually in

13 the world on this point. When we worked on this,

14 we tried to come up with laboratory based

15 therapeutics that might mimic the laboratory

16 testing, and so I'd like to show you a couple of

17 things that have grown directly from the lab

18 testing. And this is ovarian cancer data, some

19 of which you've heard. I just presented this

20 last Thursday in New York, and this is a Phase II

21 trial and completion. The interesting point of

22 this is that of 17 patients who had failed up to

23 six prior chemotherapy regimens largely deemed

24 untreatable, our response rate overall has been

25 70.6 percent, with now four complete remissions.

00157

1 Most striking in this has been the observation

2 that two of two complete remissions were obtained

3 in patients who had failed prior bone marrow

4 transplants. In addition, platinum resistant and

5 platinum sensitive patients have been shown to

6 respond.

7 I think the data indicating that the

8 laboratory assay correlated with response is

9 provided here, showing that those patients who

10 responded to the combination were the most

11 sensitive, versus the non-responders, less

12 sensitive, and a statistically significant

13 difference between the two groups. When we

14 extended this on the basis of the laboratory, in

15 a disease that is not used, in which these drugs

16 are not applied -- in fact, the only data

17 available if from our laboratory, and I can't

18 present it in a formal way because it's still in

19 submission for review for publication, but the

20 only data existing on this comes from our

21 observation in the laboratory, which indicated

22 that a disease for which platinum and gemcitabine

23 are not widely used would be an excellent

24 candidate for this treatment.

25 In our experience in this Phase II

00158

1 trial of 30 patients, we have had an overall

2 response rate of 30 percent and again, very

3 interestingly, two of four post-bone marrow

4 transplant patients have shown objective

5 responses. When we examine survival in this

6 group of patients, you can see that the patients

7 who were found assay sensitive in the green line,

8 versus the patients who were found assay

9 resistant in the pink line, statistically

10 significantly differed, and this cut across other

11 statistical considerations, including HER-2

12 positivity and number of prior treatments and

13 performance status. The single strongest

14 predictor of these patients treatment response

15 was in fact their sensitivity in vitro.

16 So when we look toward the hardest

17 evidence that we can provide to a committee like

18 this as to what it is that would provide you

19 evidence to move forward on an approval, I think

20 this is sort of a work in progress, with a lot of

21 very encouraging observations. You've already

22 heard from Dr. Bosanquet some very elegant data

23 that he's generated over the years. This most

24 recent paper, I think the most compelling,

25 published in the British Journal of Hematology in

00159

1 '99. Dr. Hoffman has presented you some very

2 exciting data in a small study with colorectal

3 cancer. I've given you some data on breast

4 cancer and ovarian cancer, much of which is in

5 progress.

6 What I'd like to point out, however, is

7 that in the coming years, there will be trials

8 under GOG auspices, a trial in development right

9 now through a group in New York for an assay

10 directed ovarian cancer first line trial, and a

11 meta-analysis, part of which Dr. Weisenthal has

12 provided to you, as an indication of the merit of

13 this and the developing data to support the merit

14 of this in the years to come.

15 I leave you with a quote from Albert

16 Einstein, who said that in a good mystery story,

17 the most obvious clues often lead to the wrong

18 suspect. In our attempt to understand the basis

19 of nature, we find similarly that the most

20 obvious intuitive explanation is often the wrong

21 one. Thank you.

22 DR. FERGUSON: Thank you. Okay. Dr.

23 Kern?

24 DR. KERN: I appreciate the opportunity

25 to address the committee today. I want to first

00160

1 reveal the financial support received in the

2 early research development of the test, and also

3 mention that I left the academic world in 1988 to

4 join Oncotech as its first director of

5 operations, a commercial firm. I left Oncotech a

6 year ago and now I am a paid consultant for

7 ImPath, who sponsored my trip and who currently

8 markets the test as a drug resistant assay.

9 We consider in our laboratory, there

10 are two aims of predicting response to

11 chemotherapeutic agents. One of course is to

12 improve response rates and perhaps even survival

13 by selecting active agents. These are the

14 so-called chemosensitive assays that most labs

15 were concentrating on in the 1980s. However, in

16 our laboratory we decided to concentrate on the

17 resistance aspects of the assay, that is, trying

18 to reduce the side effect by ruling out agents

19 that would not work clinically.

20 One way we tried to achieve this was to

21 use very high drug concentrations in the

22 laboratory. By doing this we used an average

23 exposure of five to ten times what is achievable

24 in the clinic as a maximum tolerated dose. We

25 reasoned this would not only increase the

00161

1 predictive accuracy for resistance, but also

2 would minimize the likelihood that a clinically

3 active drug would be overlooked. In other words,

4 we wanted to see, look at the aspect of, would

5 the patient be harmed if the test was faulty or

6 gave false results.

7 Now in a study conducted at UCLA and

8 published in 1990, we looked at correlations from

9 450 patients. In this slide what I'm showing is

10 the data where each patient is plotted as a

11 single dot. The patients were tested in the

12 laboratory with the same chemotherapy to which

13 they were treated in the clinic and the results

14 are plotted here, first as responders, as

15 complete or partial responders, and in the column

16 on the left, non-responders. And we looked at

17 the assay results. At the top are those patients

18 that were very sensitive in the laboratory, and

19 in the bottom those patients who had tumors very

20 resistant to the chemotherapy to which they were

21 tested.

22 And in this bottom group, which we

23 called extreme resistance, a term coined by

24 Dr. Weisenthal and myself in 1990, we found there

25 was only one responder in this group of 127

00162

1 patients. In other words, the assay was 99

2 percent accurate in predicting clinical failure.

3 Also, I wanted to point out that the assay did

4 not predict clinical resistant patients, but

5 rather, this data is tumor and drug specific. In

6 other words, the data is for the exact drugs to

7 which the patients were given. They were tested

8 in laboratories with the exact drugs that we used

9 in the clinic, a very important point.

10 Now there were different levels of

11 evidence to judge laboratory tests as well as

12 therapeutics. It has been proposed that the

13 National Cancer Institute standards of levels of

14 evidence might be applied to laboratory tests. I

15 just want to point out that these levels of

16 evidence weren't designed, however, for

17 therapies, new therapies. And I believe more

18 appropriately would be to look at the levels of

19 evidence proposed by the CDC for evaluating

20 clinical tests. One is to look at the accuracy

21 and precision of the test, its clinical

22 effectiveness, the clinical context in which the

23 test is used, including, do the patients have

24 free access to the test, turnaround time and

25 cost, the practical values of the test, and also,

00163

1 what is the impact if the test is wrong? What

2 would be the effect of, if we gave medically

3 misleading information?

4 One of the classical ways of looking at

5 the effectiveness of a laboratory test is to look

6 at its receiver operating characteristics. In

7 this slide I will show the receiver operating

8 characteristics first for the assay's ability to

9 predict sensitivity, that is, predict active

10 drugs. Now, the use of the test can be

11 determined by how far it deviates from this

12 diagonal line. If all the data fell upon this

13 diagonal line, the laboratory test would be

14 totally worthless. By measuring this area under

15 the curve, one can get an estimate of how

16 worthwhile the test is.

17 Also, the worthlessness or usefulness

18 of a test can be estimated by the prevalence of

19 the marker or the facts that the test was

20 designed to measure. In this case we're trying

21 to measure resistance. So in a clinical setting

22 where you're looking at a high prevalence of

23 resistance, that is, in very refractory cancers,

24 the test is not vary good in identifying active

25 agents. The red line shows the ability of the

00164

1 test to detect resistance, clinical resistance.

2 When there is a high -- I'm sorry -- when there's

3 a low prevalence of resistance, that is, for

4 those cancers that are extremely sensitive to

5 chemotherapy and highly curable, the assay is not

6 very good at predicting drug resistance or

7 clinical failure.

8 But in the real world, knowing that

9 most of the tumors fall in a central range here,

10 the test is extremely good for both identifying

11 active drugs and also inactive drugs. In fact,

12 it's almost perfect in this range of identifying

13 drugs that will fail in the clinic.

14 Another way of looking at the test is

15 to look at its negative and positive predictive

16 values. I have plotted across -- the predictive

17 accuracy, the positive predictive accuracy are a

18 number of data sets, first, from one laboratory

19 shown in green, against a number of different

20 tumor types. Second, I have plotted results from

21 a number of different laboratories, mostly

22 representing one tumor type. For example, Dr.

23 Albert's in ovarian cancer, and so on.

24 You see the assay is extremely accurate

25 in predicting clinical resistance. The negative

00165

1 predicted value here I plotted as its converse.

2 That is, 100 minus the negative predicted value.

3 So the test showed about a 90 to 100 percent

4 accuracy in predicting clinical failure.

5 However, the prevalence of resistance had a

6 dramatic effect on the ability of the assay to

7 predict drugs that would work in the clinic.

8 I also want to point out that although

9 there is no single gold standard about which test

10 is better, this data indicates an incredible of

11 interlaboratory reproducibility. Using this kind

12 of data analysis, you can compare results from a

13 number of different laboratories and in the hands

14 of experienced investigators like those

15 represented here, the tests are remarkably

16 similar. Also, the test is extremely accurate.

17 All the tests show extreme accuracy in predicting

18 clinical failure.

19 Looking at standards to judge the

20 medical usefulness of the test, which is of

21 course your job, knowing that the drug resistance

22 assay predicts clinical failure with extremely

23 high accuracy, the test can be used, first of

24 all, to avoid giving worthless treatments to

25 cancer patients, and to help the patients avoid

00166

1 the terrible side effects of useless

2 chemotherapy. The assay is also cost effective

3 because when you eliminate worthless therapies

4 and useless side effects, it obviously makes

5 economic sense.

6 I want to just take a minute to

7 indicate what I think are some of the reasonable

8 indications for the use of this committees.

9 Remember, we talked about the test appears most

10 worthwhile in the real world represented by the

11 breast cancers, the ovarian cancers, the lung

12 cancers. In breast cancer, even if the test was

13 only used to identify adreomyecin resistant

14 patients, and to switch those to, let's say CMF

15 or an alternative therapy, the test would be cost

16 effective if you only identified 10 percent of

17 the adreomyecin resistant patients and switched

18 them to CMF.

19 Here we talked about Gemcitabine.

20 Certainly useful in lung cancer, but if the test

21 was only able to identify 3 percent of the

22 patients that were Gemcitabine resistant and

23 switched them to platinum etoposide or platinum

24 vinorelbine, the test would pay for itself.

25 Also, equally important are where the

00167

1 tests may not be very useful, and two particular

2 examples are in cancers where there is an

3 extremely high prevalence of drug resistance, for

4 example, in renal cell carcinoma, the test would

5 not be very useful. That doesn't mean that it's

6 a bad test. The test still has the same

7 sensitivity and specificity, it just means that

8 there are no worthwhile drugs, no useful drugs

9 for treating, drugs like kidney cancer. On the

10 other end of the spectrum, testicular cancer is

11 clinically very responsive, very low prevalence

12 of resistance, the test is probably not needed in

13 that clinical study.

14 So I want to stop here and just

15 summarize. First, let's talk about how the

16 patients benefit from drug resistance testing.

17 Chemotherapy causes a lot of suffering for cancer

18 patients. The suffering is tolerable if the

19 treatment leads to prolonged survival. But

20 wouldn't it be beneficial to the patients if they

21 could be spared the needless suffering of

22 worthless chemotherapy?

23 Let's consider the cost benefits of the

24 test. The Medicare system could benefit

25 immediately if it didn't have to pay for drugs

00168

1 that are useless in the clinic. So don't you

2 think Medicare would benefit economically by

3 using the drug resistance tests?

4 And finally on a, consider it a

5 personal note. Imagine if a wife, spouse, parent

6 or loved one should be unfortunate enough to be

7 diagnosed with cancer. Would you not want that

8 loved one to have the benefit of drug resistance

9 testing? So with that in mind, I respectfully

10 ask the committee, please do not deny America's

11 seniors the access to the proven benefits of the

12 drug resistance test. Thank you.

13 DR. FERGUSON: Thank you, Dr. Kern. I

14 think our next speaker, rather than Dr. Bailes,

15 will be Dr. Hazie.

16 DR. HAYES: Hayes.

17 DR. FERGUSON: Hayes, I'm sorry, from

18 Georgetown.

19 DR. HAYES: I'll introduce myself. I

20 know who I am. I am Dr. Daniel F. Hayes. I'm a

21 medical oncologist. I'm the clinical director of

22 the breast cancer program at Georgetown. I'm

23 also a member of the American Society of Clinical

24 Oncology.

25 I have a couple of credentials. One of

00169

1 those is that I am the chair of the solid tumor

2 and correlative science committee of the Cancer

3 and Leukemia Group B, one of the major

4 multi-institutional groups funded by the Federal

5 Government. Our committee is essentially the

6 tumor marker committee of the CAGLB. And I'm

7 also a member of the American Society of Clinical

8 Oncology tumor expert guidelines panel that was

9 convened roughly five years ago. I am not the

10 chair of that. And I'm also hoping that I will

11 have some slides here, which is why I'm wasting

12 your time mumbling around here until this thing

13 gets up and running.

14 And finally, I apologize. Dr. Bailes,

15 who is the president of the American Society of

16 Clinical Oncology meant to be here today and was

17 unable to do so. And Dr. Dan Van Hoff also meant

18 to be here and also was unable to be here. And

19 finally, I have no conflicts of interest with any

20 of the companies that have been presented here,

21 not performed any research with any of them. I

22 guess my only conflict of interest is I am a

23 member of the American Society of Clinical

24 Oncology.

25 Finally, while this thing is hopefully

00170

1 waiting to boot up, I will say, because I wrote

2 some notes in case my slides didn't work, that

3 the American Society of Clinical Oncology is

4 neutral on this issue and is not here to serve as

5 either a proponent or opponent of your decision

6 to reimburse for any of these assays in the

7 elderly population. Rather, we are here to make

8 a plea regarding reimbursement for treatment, and

9 in fact this was brought up by the last speaker,

10 and this is one place where I believe we would

11 contend that that would be an error, at least

12 with the currently available data.

13 Thank you. I apologize for the time

14 it's taking to do this. So as I said, we are

15 actually neutral on this issue. We are actually

16 very interested and believe there is a

17 substantial amount of interesting data that are

18 coming out of the various studies, much of what

19 you have seen today. We appreciate the

20 remarkable progress in technology that has

21 occurred over the last 15 years since the

22 original publications by Salmon, et al., but we

23 remain very concerned about the reliability to

24 exclude therapy, especially in relationship to

25 combination chemotherapy.

00171

1 We have been very interested in

2 establishing guidelines for the members of our

3 society. There have been at least two guidelines

4 panels related to tumor markers. One of those

5 which was specifically related to tumor markers

6 for breast and colon cancer, the most recent

7 update of that was published in JCO in 1998. A

8 second related to tumor markers, but not

9 specifically excluding, or specifically related

10 -- the specific focus on it was the follow-up of

11 primary therapy in breast cancer, again,

12 published in 1997.

13 In both of these, particularly in the

14 first, we spent a great deal of time trying to

15 decide when a tumor marker is truly ready for

16 prime time, something I believe you are being

17 asked to contend, and in fact the discussions by

18 the FDA this morning I found very interesting.

19 We like they found there are very few rules about

20 how to use a tumor marker in clinical practice,

21 unlike how to use therapeutics. And our

22 predecessors 30 to 40 years ago, Fry, Holland,

23 Karnoski, established rules and guidelines about

24 how to talk to each other, what is a Phase I,

25 what is a Phase II, so on and so forth, what's a

00172

1 complete response. Those sorts of things have

2 not been established well for tumor markers, and

3 we found it very confusing.

4 Ultimately, one of the things that we

5 suggested was that the results of the tumor

6 marker must be known to influence the decision to

7 result in improvement in overall survival,

8 disease free survival, quality of life or cost,

9 echoing many of the things that have been said

10 here today.

11 Also echoing many of the things that

12 have already been said, we noticed an astounding

13 amount of heterogeneity among tumor marker

14 studies, and these are for many reasons. Patient

15 selection, different assay issues, the use of

16 different drugs; for example in this case, we've

17 been talking about the difference between the use

18 of single agent therapy and multiple combination

19 therapy. And probably importantly, the

20 difference between the settings, whether or not

21 these sorts of assays should be used to direct

22 the therapy in the adjuvant setting where you

23 don't have an identifiable end point immediately

24 but rather, must wait for progression or

25 survival, or the use of metastatic where you have

00173

1 things like immediate quality of life and

2 immediate response rates.

3 These are some conceptual slides that

4 we developed relative to any tumor marker, and

5 that is a pure prognostic factor, for example,

6 separates two groups of patients in the absence

7 of therapy or in the presence of therapy

8 equally. It does not tell you whether or not

9 that therapy will be helpful but rather, it tells

10 you how the patients will do. Often, the

11 difference between prognostic and predicted

12 factors gets mixed in many papers and also in

13 many discussions. We felt it important to point

14 this out.

15 So that for example, these curves are

16 clearly separate, but they are equally separate

17 in the absence or presence of therapy. It does

18 not tell you whether these patients should be, or

19 how they should be treated, it just might tell

20 you that they should be treated because their

21 prognosis is worse.

22 A pure predictive factor on the other

23 hand, the curves are not separated at all in

24 terms of prognosis in the absence of this

25 specific therapy. I put no therapy here, but in

00174

1 fact, we could be talking about the specific

2 therapy one is concerned about, whereas in the

3 presence of that therapy, the curves are

4 separated. In this case, if it's a predictive

5 factor for sensitivity to therapy, then the

6 curves are separated, with those of a factor

7 positive are doing much better than those who are

8 factor negative.

9 Indeed, the real issue that we're

10 discussing, I believe here today, is the

11 separation between these curves, not whether or

12 not they are statistically significantly

13 separated. That is what the P value tells you.

14 What really tells you is the magnitude of the

15 difference, and in fact this has already been

16 brought up earlier at least once today. And that

17 is, for example, many of us would be very willing

18 to use this marker to separate patients into two

19 groups, and treat them differently, especially if

20 the toxicities of the drug are high. Whereas in

21 this group of patients, we would probably all

22 treat this group the same way we treat this

23 group, because the magnitude is not large.

24 Now in order to really assess the

25 utility of a predictive factor, it requires a

00175

1 control group that did not receive the therapy.

2 That is of course best done in the presence of a

3 prospective randomized trial in which biases are

4 minimized because of the randomization.

5 Historical controls are acceptable in some cases,

6 but they are fraught with the usual biases, that

7 is, comparing the outcomes of your patients with

8 those who did not receive the therapy in

9 non-randomized fashion. The use of response rate

10 gets confusing and in fact, it assumes a

11 historical control in which no therapy, frankly,

12 equals no response. We must assume that patients

13 who are not treated will not have their tumors

14 regress. That's not always true, but often it

15 is, and in general, this is a relatively fair

16 assumption.

17 So then for example, if we take what I

18 believe the best predictive factor in oncology,

19 and that is S Receptor and Tamoxifen, it is both

20 a prognostic and a predictive factor, which makes

21 it confusing. In the absence of therapy, ER

22 positive patients do slightly better than the ER

23 negative patients, but in the presence of

24 therapy, ER positive patients do substantially

25 better than ER negative patients. In this case I

00176

1 have used Tamoxifen as the therapy, because it's

2 the one for which we have the most data.

3 And in fact, one could begin to use

4 these sorts of relative differences and develop

5 overview analyses of the relative risk, so that

6 in this case, one means therapy is no better than

7 therapy in the presence of a randomized trial in

8 which the patients are randomized, the therapy

9 are not. And then we have a 10 percent

10 difference, a 20 percent benefit, a 30 percent

11 difference, or a 40 percent reduction in the odds

12 of the event, in this case let's say it's a

13 recurrence. A weak predictive factor may split

14 these patients apart statistically, but may not

15 be important clinically, since both groups of

16 patients benefit. A moderate predictive factor

17 splits them further apart and a strong predictive

18 factor splits them much further apart, so that

19 one might not treat these patients, and one is

20 very likely to treat these patients.

21 And we came up with what we called a

22 relative predictive factor, and that is just the

23 relative odds of response in the positive group

24 divided by the relative odds response -- I'm

25 sorry, I have this running on battery. Okay.

00177

1 Well, I made the point.

2 And then one might say, all right, we

3 will discard those that are very weak, consider

4 those that are intermediate, and accept those

5 that are very strong. And of course, what we do

6 here, this is highly subjective, depending on the

7 toxicities of therapy and the patient

8 perspective; there are some patients who would

9 accept therapy regardless of the toxicity in the

10 presence of any benefit, and others who would be

11 more thoughtful and say I'm willing to give up

12 some benefit in order to avoid some toxicities.

13 Again, if we use ER in the adjuvant

14 setting to predict relative benefits from

15 Tamoxifen, the proportional reduction from the

16 last Oxford overview for adjuvant Tamoxifen

17 versus nil, and those were ER poor, was 0.06. In

18 other words, there was a 6 percent proportional

19 reduction in the odds of recurrence and death,

20 whereas in those that were ER rich, it was nearly

21 50 percent, and the relative predictive factor

22 for ER for Tamoxifen in the adjuvant setting in

23 this case was well over what we would consider

24 acceptable for routine clinical utility, and in

25 fact S receptor is used very much in that

00178

1 setting.

2 So the ASCO view of these proceedings

3 are that we believe in vitro chemosensitivity

4 assays are promising. We believe the current

5 data are insufficient to withhold potentially

6 effective drugs. Perhaps in the metastatic

7 setting these might be very helpful. In the

8 adjuvant setting, I really believe that we need

9 prospective clinical trials to address the issue,

10 because all adjuvant therapies are empiric by

11 definition, since one does not have disease to

12 measure at the time.

13 There is a technology assessment

14 planned by ASCO to be convened in the winter,

15 this winter 2000. We hope that by next summer,

16 the results will be available and published. And

17 we advise against noncoverage for agents found to

18 have drug resistance for individual patients.

19 Thank you for your time and again, I

20 apologize for the inconvenience of the

21 technology.

22 DR. FERGUSON: Dr. Loy?

23 DR. LOY: I'm Dr. Bryan Loy. I'm a

24 carrier medical director. I'm a guest of the

25 Health Care Financing Administration. I don't

00179

1 need my slides to get started. In the interest

2 of time, I will proceed onward.

3 I am a part B carrier medical director

4 or CMD for the State of Kentucky. In the

5 interest of time, I will go ahead and start here

6 without my slides. I just wanted to get

7 started. I am not dependent on my slides,

8 because my presentation is of a different tact

9 and not of a scientific nature.

10 I'm a part B carrier medical director

11 for the State of Kentucky, or CMD. I appreciate

12 being here and I am very interested in these

13 presentations. I am going to go to the end

14 because in my role I will be asked to implement

15 or execute any national coverage policy decision

16 that results from these deliberations. My

17 concerns are the appropriate use of these

18 technologies for the treatment of patients, but I

19 am also concerned about the possible

20 vulnerabilities that a national coverage policy

21 can create that could result in the

22 misapplication of these technologies.

23 So the intent of my presentation is to

24 describe the approach that I take at the carrier

25 level to assess the need for policy development

00180

1 for new technologies and my carrier, and to

2 discuss the impact of implementation of national

3 policy on the carrier medical director, and to

4 discuss some of the dilemmas that may arise as a

5 result of the outcome of these deliberations, and

6 then finally, to present my prospectus regarding

7 human tumor assay systems.

8 And the reason I am interested in

9 presenting it in this way is because I think it's

10 necessary for the panel to understand that this

11 won't be a yes no answer to a question. It won't

12 be a yes, we will have coverage, no, we will not

13 have coverage. Somehow, carriers and carrier

14 medical directors will be asked to somehow

15 implement a policy decision that is both

16 reasonable and necessary in accordance with the

17 law.

18 Let me start off by describing the

19 environment CMDs work in as it pertains to

20 medical policy, whether it be national policy or

21 local medical review policy, and this has already

22 been stated. We're currently in an environment

23 where we're dealing with the local medical review

24 policy in the absence of a national coverage

25 policy when it comes to talking about resistance

00181

1 testing.

2 The practice of medicine is ever

3 changing for many reasons, and one of the reasons

4 is the introduction of new technologies or new

5 applications of existing technologies.

6 Technologies can fill voids in the practice of

7 medicine. For example, they can provide

8 previously unavailable information. Technologies

9 can also replace or supplement existing

10 technologies. They can provide better, faster,

11 more complete information. They can direct

12 patient care. The role that technologies play in

13 the treatment of patients is in part dependent on

14 the acceptance and the implementation of the

15 technology by the practicing providers within our

16 jurisdictions. I am fully aware that acceptance

17 and implementation can also be related to

18 reimbursement and coverage decisions.

19 In my CMD role, routinely I will

20 receive a call asking me to confirm my name and

21 address from a technology company, and usually

22 within a week or so I will receive a packet from

23 a company representative asking me to review the

24 material, and in follow-up I will receive a note

25 or a telephone call asking me if we have coverage

00182

1 policy or local medical review policy addressing

2 the technology in question. Less frequently, I

3 will receive requests from treating physicians

4 asking for coverage for a particular diagnostic

5 or therapeutic application of the technology.

6 If there's a significant number of

7 requests for coverage and we are permitted

8 discretion as the local carrier, we will consider

9 making a coverage decision in compliance with the

10 Medicare carrier's manual. Policy at the local

11 level is important in that it describes

12 appropriate coverage within the Medicare program

13 in accordance with the Social Security Act.

14 National policy, on the other hand, can

15 have different effects on the carrier.

16 Regardless of the number of requests for

17 coverage, local carriers are expected to

18 implement all national policy of the carrier, and

19 therefore, the same national policy can have

20 different effects on different carriers. And

21 here's why: For carriers in states that have

22 providers already utilizing the technology as we

23 have heard described today, when this is a topic

24 of a national policy, well then, the carriers and

25 the providers will have an interest in the

00183

1 coverage and the utilization and the pricing

2 language in the policy, and how the carrier

3 implements the policy. For carriers in states

4 whose providers do not utilize the technology,

5 the carriers do not receive claims and therefore,

6 there is less interest in the coverage decision.

7 For providers who do not utilize the

8 technology, creating policy creates a possible

9 future benefit. The policy becomes relevant only

10 when the providers are utilizing the technology

11 and billing the carrier. Over time, technologies

12 gain acceptance by providers and hopefully gain

13 acceptance of national organizations. Many times

14 national organizations will publish guidelines,

15 position papers, et cetera, describing the

16 appropriate use of the technology for patient

17 care. When these guidelines and positions are

18 supported by scientific evidence, practice

19 patterns usually become a national uniform

20 standard of care. This is not always the case.

21 Sometimes technologies fall out of

22 favor, they do not gain acceptance, or are

23 replaced by better technologies. Unfortunately,

24 sometimes coverage decisions are considered

25 before practice patterns utilizing technologies

00184

1 are firmly established, and the policy is

2 developed whether they be national or local

3 carrier policies, may be premature.

4 Subsequently, indications can be added, off label

5 use established, as additional research results

6 become available. Provider acceptance of the

7 technology can also change, and some providers

8 may utilize the technology different than

9 providers in other states, and the standard of

10 care can change dramatically as a result of these

11 refinements.

12 These variances can quickly render

13 policies obsolete. Creating premature policy

14 that is silent on evolving uses of new

15 technologies can tend to work disparate coverage

16 between the states. If a national policy is

17 silent on an evolving use of the technology, then

18 the local carrier has to make coverage decisions

19 in response to the provider inquiries when they

20 arise. This commonly occurs when off label use

21 of technologies are introduced in the current

22 medical practice. Scientifically based well

23 written national policy can minimize many of the

24 disparities among contractors, and an added

25 benefit is that this process decreases the amount

00185

1 of resources used by contractors to create local

2 medical review policy.

3 In my opinion, the Medicare carrier

4 advisory committee is the appropriate forum for

5 raising and addressing coverage questions

6 regarding human tumor assay systems, as opposed

7 to the local carrier level. I also believe that

8 this is the forum to discuss the scientific

9 validity of the test, and to assess the

10 scientific validity of the clinical applications

11 of these test results for each cancer.

12 Implementing this national policy without

13 operationalizing this national policy at the

14 carrier can create some disparities in coverage

15 if the specific clinical indications are not

16 clearly articulated in the policy. When specific

17 clinical indications are identified, specific

18 codes can be employed in order to create system

19 edits for automating appropriate claims payment.

20 When appropriate frequencies are identified,

21 these parameters can also be implemented for

22 automating appropriate claims payment. Ideally,

23 this results in efficient correct payment of

24 claims.

25 It is my opinion that if the science

00186

1 does not support the clinical utility of the test

2 results, then the policy should specify

3 noncoverage at this time. If the science does

4 support the use of specific clinical applications

5 of the technology, then the policy should clearly

6 specify the indications in the clinical scenario.

7 These indications in the case of a test, could

8 include the clinical setting. For example, in

9 the case of cancer testing, the policy should

10 specify the appropriate cancers and the

11 appropriate times the test should be performed.

12 For example, first line, adjuvant, and/or the

13 metastatic settings. The frequencies of testing

14 and what clinical intervention should have taken

15 place should also be identified in the policy.

16 Again, this should be supported by the science.

17 Implementing narrative for policies can

18 be difficult and, therefore, clinical parameters

19 need to be well defined. If policy language is

20 vague, or subject to multiple interpretations,

21 then this can lead to misapplications of the

22 policy and can give rise to interstate coverage

23 disparities. Moreover, if national policy is

24 silent on an evolving use of the technology, then

25 providers and carriers are left with a policy

00187

1 that won't address these evolving issues. Left

2 unaddressed, these applications must be evaluated

3 by the carrier, and local policy will be created

4 to address these new issues. Again, this can

5 lead to disparate coverage.

6 Polices for technologies in evolution

7 are difficult to write, difficult to implement,

8 and very difficult to maintain. This committee's

9 deliberations will most likely result in one of

10 the following outcomes: A no decision with the

11 possibility of local carrier discretion

12 remaining; a no coverage policy, which can be

13 easily implemented in uniform across carriers; a

14 limited coverage policy which may be supported by

15 current science and lead toward a uniform

16 national coverage, but subject to early

17 obsolescence if applications continue to evolve;

18 and finally, a broad coverage policy, allowing

19 for flexibility for different and evolving

20 practice patterns, but most likely containing

21 vague language.

22 I would like to conclude by expressing

23 my personal prospectus concerning coverage for

24 human tumor assay systems. Human tumor assay

25 systems have been in existence for many decades,

00188

1 but in my community these technologies have not

2 been routinely employed for the treatment of

3 cancer patients. I have reviewed the information

4 supplied by the presenters. These technologies

5 sound quite promising. In my mind, one of the

6 indicators for assessing whether a technology

7 that generates information is needed for treating

8 patients is the number of requests that I receive

9 for coverage of the technology in question from

10 the end user of the information. In this case,

11 it would be the practicing oncologist in my

12 community making the request, not the producer or

13 the laboratory producing this information that

14 would in my mind begin to legitimize the request.

15 I have had no requests for coverage for this

16 type of testing in my state.

17 I have asked some well respected

18 practicing oncologists in my community and have

19 generated little interest in these technologies.

20 Coverage and reimbursement issues have not

21 entered the conversation. The discussions have

22 focused on the controversies relating the in

23 vitro results to the long-term clinical benefit.

24 I then questioned multiple carrier medical

25 directors in my region and throughout the United

00189

1 States. Of those responses, only two states

2 reported claims submission or requests for

3 coverage. I do not see where the clinical use of

4 these types of chemotherapy drug assays have been

5 generally accepted or adopted as a national

6 standard of care. It is not clear to me that the

7 practicing oncologists, or those providing these

8 methodologies are yet in agreement on the

9 clinical applications and the clinical value of

10 these tests.

11 More specifically, questions remain in

12 my mind unanswered as to what point the testing

13 should be used, for what cancers, what clinical

14 scenarios, how frequently these tests should take

15 place. I have not identified in the presenters'

16 packet any position statements, guidelines,

17 et cetera, that would convince me that this

18 technology has matured into a standard of care.

19 Furthermore, even though some evidence supports

20 the use of this testing for specific clinical

21 indications, the evidence supporting a broad

22 national coverage is insufficient, in my

23 opinion. At this time it is clear to me that

24 this technology is not being utilized routinely

25 for medical decision making for most cancer

00190

1 patients. During your deliberations I would

2 encourage all members of the impact panel to

3 envision a finished document that would only

4 incorporate scientifically based or evidence

5 based medicine that is currently applicable for

6 treatment of specific cancer types and their

7 appropriate clinical scenarios. If the science

8 only sufficiently addresses certain aspects of

9 clinical utility, then only allow for that

10 coverage. Allowing flexibility in policy for

11 anticipated future trend allows for possible

12 coverage of misapplications of this technology.

13 In my opinion, the science supporting the

14 clinical applications for these testing

15 methodologies is still evolving, still coming in,

16 and there are many unanswered questions

17 remaining. Thank you.

18 DR. FERGUSON: Thank you, Dr. Loy.

19 It's four minutes to 12. I suppose we

20 could have questions or comments for four

21 minutes. Any of the panel members? Yes, Dr.

22 Sundwall?

23 DR. SUNDWALL: One thing that has been

24 going through my mind this morning as we've heard

25 all these excellent presentations, and that is,

00191

1 how much patient variability is there when they

2 have the same malignancy? Now that may have been

3 addressed, but if it was, it's not clear to me if

4 in fact, once you use this testing, which I think

5 seems to have great value, potential value, but

6 I'm not certain how much variability per tissue

7 type there is from patient to patient, or if in

8 fact once you get help, that there is sensitivity

9 to a particular drug, or combination of drugs,

10 why isn't that applicable across the board?

11 DR. FERGUSON: Did you all hear the

12 question? I guess we can have a comment from one

13 of the presenters. Yes, Dr. Weisenthal?

14 DR. WEISENTHAL: The question that was

15 asked is about disease specific activity versus

16 patient variability. Firstly, clinical

17 heterogeneity with a given disease is an

18 established fact. That's shown by the fact that

19 chemotherapy number one is not universally

20 effective. For example, in the case of a disease

21 like breast cancer, or ovarian cancer, first line

22 chemotherapy will produce a response about 70

23 percent of the time, in the case of colon cancer,

24 only 20 percent of the time. But more telling is

25 the large numbers of patients that fail first

00192

1 line therapy that subsequently respond to second

2 line therapy. And I'm not sure that -- but you

3 heard of five cases today that were presented

4 between me and Dr. Nagourney, five patients who

5 failed high dose chemotherapy with bone marrow

6 transplantation. $200,000 in therapy. These

7 were patients that never responded to anything,

8 including ultahigh dose of chemotherapy. They

9 then got an assay and they went into complete

10 remission.

11 In the case of Dr. Nalick's patient,

12 that was someone that failed first line Taxol

13 platinum, failed tandem stem cell transplants.

14 The amount of money that was spent on ineffective

15 therapy for this patient would pay to run my lab

16 for six months. So the issue is that clinical

17 heterogeneity is an established fact. There are

18 many, many patients who fail first line therapy

19 who respond to second line therapy. They should

20 have gotten the second line therapy the first

21 time around.

22 Dr. Bosanquet talked about his patients

23 with fludaribine resistance all died, because by

24 the time they got the fludarabine which didn't

25 work, they were too sick to get anything else.

00193

1 Had they gotten the correct therapy the first

2 time, they wouldn't have been dead. So the thing

3 is, that all the laboratories that do this know

4 that there is a tremendous heterogeneity with any

5 given disease type. Some tumors with -- some

6 breast cancer tumors are very resistant to

7 chemotherapy, some are not. And the same thing

8 holds for the clinic. That's the whole purpose

9 for doing the testing.

10 DR. FERGUSON: Thank you. Dr. Brooks?

11 DR. BROOKS: I have a question for Dr.

12 Kern, who is now with ImPath. Does ImPath have a

13 different type of test, or is it going to offer a

14 certain type of test based on a different

15 methodology?

16 DR. KERN: No. The methodology is

17 basically the test described also by Dr.

18 Fruehauf, called the EDR at Oncotech, or DRA at

19 ImPath. But it's based on the same technology.

20 May I also respond to the prior

21 question for 15 seconds?

22 DR. FERGUSON: Sure.

23 DR. KERN: I can visually show what Dr.

24 Weisenthal was describing. This is a data set of

25 40 consecutive ovarian cancer patients, all

00194

1 previously untreated. You see patient number one

2 at the top, and tested against five different

3 drugs for ovarian cancer. Patient number one was

4 sensitive to carboplatin, resistant to

5 adreomyecin, so on, sensitive to Taxol. Patient

6 number two was resistant to Taxol. You can see

7 the patterns; it's quite different patient to

8 patient.

9 DR. FERGUSON: What cancer was this?

10 DR. KERN: This is ovarian cancer.

11 Yes?

12 DR. BROOKS: Doctor, the question is,

13 I'm not sure what the question was he asked, but

14 the question I would ask, is ovarian cancer

15 endometrioid, cirrus, mucinous, poorly

16 differentiated, you know, et cetera, et cetera?

17 So amongst the histologies of ovarian cancer,

18 would you have similarity?

19 DR. KERN: Well, what you actually have

20 to do is look at the clinical experience. In

21 other words, the tests, we try to reflect what's

22 actually going on in the clinic. For epithelial

23 cancer, it's different from clear cell cancer, so

24 you test different drugs and you look at, again,

25 it has to be tumor and histology specific, and

00195

1 cancer specific, drug specific testing. And you

2 do find in all different types, a great deal of

3 heterogeneity from patient to patient.

4 DR. FERGUSON: Thank you. One more

5 brief comment. Yes?

6 MR. KIESNER: One comment in relation

7 to Dr. Loy's excellent presentation. I think one

8 of the observations that Dr. Loy made was that he

9 within his region doesn't receive a lot of

10 requests for the coverage. I think the way the

11 laboratory industry bills, if we get tissue from

12 a thousand hospitals around the country, the

13 issue of where it's billed is dependent on where

14 the work is done. And so in our case, when all

15 the tissue comes to Irvine, we have to bill the

16 local carrier in southern California, so that's

17 where all the payment questions are directed.

18 A second question is in relation to the

19 technology, there are burdensome drafting

20 requirements. I recognize that, we all recognize

21 that. We sat through the negotiated rule making

22 session last summer. We did have an oncology

23 work group with about 15 people, including HCFA.

24 We were able to deal with those issues; I'm not

25 sure we dealt with them completely, but it is an

00196

1 issue that can be dealt with.

2 DR. FERGUSON: Okay, thank you. I

3 guess like many other things, it starts in

4 California and goes east; is that right? Anyway,

5 I think on that, we will have lunch, and

6 reconvene at 1:00.

7 (The panel recessed for lunch at 12:03

8 p.m., November 15, 1999.)

9 (End of Volume I)