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