1

DEPARTMENT OF HEALTH AND HUMAN SERVICES

FOOD AND DRUG ADMINISTRATION

CENTER FOR DRUG EVALUATION AND RESEARCH

ONCOLOGIC DRUGS ADVISORY COMMITTEE

VOLUME II

Friday, March 4, 2005

8:00 a.m.

Gaithersberg Hilton

620 Perry Parkway

Gaithersburg, Maryland

PARTICIPANTS

Silvana Martino, D.O., Acting Chair

Johanna M. Clifford, M.S., RN, Executive Secretary

COMMITTEE MEMBERS

Otis W. Brawley, M.D.

James H. Doroshow, M.D.

Antonio J. Grillo-Lopez, M.D., Industry

Representative

Pamela J. Haylock, RN, Consumer Representative

Maha H.A. Hussain, M.D.

Alexandra M. Levine, M.D.

Joanne E. Mortimer, M.D.

Michael C. Perry, M.D.

Gregory H. Reaman, M.D.

Maria Rodriguez, M.D.

CONSULTANTS (VOTING)

Ralph D'Agostino, Ph.D.

Michael Proschan, Ph.D.

PATIENT REPRESENTATIVE (VOTING)

Sheila Ross - For Iressa

FDA

Martin Cohen, M.D. (a.m. session)

Grant Williams, M.D. (a.m. session)

Amna Ibrahim, M.D.

Nancy Scher, M.D.

Eric Colman, M.D.

Mark Avigan, M.D.

Richard Pazdur, M.D.

Robert Temple, M.D.

C O N T E N T S

Page

Call to Order and Introductions

Silvana Martino, D.O. 5

Conflict of Interest Statement

Johanna Clifford, M.S., RN 7

Opening Remarks

Richard Pazdur, M.D. 10

Sponsor Presentation

AstraZeneca L.P.

Introduction and Regulatory History

Mark Scott, Ph.D. 15

Trial 709

Kevin Carroll, MSc 24

Clinical Actions and Implications

Judith Ochs, M.D. 37

Summary

Mark Scott, Ph.D. 56

Committee Questions 59

Open Public Hearing 89

Committee Discussion 107

- - -

Call to Order and Introductions

Silvana Martino, D.O. 140

Conflict of Interest Statement

Johanna Clifford, M.S., RN 142

Opening Remarks

Richard Pazdur, M.D. 145

C O N T E N T S (Continued)

Page

FDA Presentation

Regulatory History of Zometa and Aredia

Nancy Scher, M.D. 147

Post-Marketing Safety Assessment of

Osteonecrosis of the Jaw: Pamidronate

and Zoledronic Acid

Carol Palmer, R.Ph. 155

Osteonecrosis of the Jaws in Myeloma:

Time Dependent Correlation with Zometa

and Zometa use

Brian Durie, M.D. 173

Sponsor Presentation

Novartis Pharmaceuticals

ONJ Reported in Bisphosphonates Treated

Patients - An Overview

Diane Young, M.D. 188

Clinical Benefit of Bisphosphonates in Cancer

Patients with Metastatic Bone Disease

James Berenson, M.D. 222

Open Public Hearing 229

Committee Discussion 252

P R O C E E D I N G S

Call to Order and Introductions

DR. MARTINO: Good morning, ladies and

gentlemen.

The topic before us this morning is some

additional new data that has arisen relative to the

agent Iressa. Before we start with the topic

itself, I am going to ask the committee to

introduce itself, and we will start on my left with

Dr. Pazdur, please.

DR. PAZDUR: Richard Pazdur, FDA.

DR. WILLIAMS: Grant Williams, FDA.

DR. COHEN: Martin Cohen, FDA.

MRS. ROSS: Sheila Ross, Lung Cancer

Alliance formerly ALCASE.

MS. HAYLOCK: Pam Haylock, Oncology Nurse,

University of Texas Medical Branch in Galveston.

DR. LEVINE: Alexandra Levine, University

of Southern California, Chief of Heme.

DR. RODRIGUEZ: Maria Rodriguez, M.D.

Anderson Cancer Center.

DR. REAMAN: Gregory Reaman, Pediatric

Oncologist, Children's Hospital, Washington, D.C.

DR. MARTINO: Silvana Martino, Medical

Oncology, Cancer Institute Medical Group in Santa

Monica.

MS. CLIFFORD: Johanna Clifford, Executive

Secretary to the Oncologic Drugs Advisory

Committee.

DR. HUSSAIN: Maha Hussain, Medical

Oncology, University of Michigan.

DR. PERRY: Michael Perry, Medical

Oncology, University of Missouri, Ellis Fischel

Cancer Center.

DR. MORTIMER: Joanne Mortimer, Medical

Oncology, University of California at San Diego.

DR. GRILLO-LOPEZ: Antonio Grillo-Lopez.

I am a hematologist/oncologist, a five-year cancer

survivor, and I am here as the industry

representative on this committee. I would like to

state that although I am the industry

representative, I receive no support whatsoever

from industry for my presence here.

DR. PROSCHAN: Mike Proschan. I am from

the National Heart, Lung, and Blood Institute.

DR. D'AGOSTINO: Ralph D'Agostino, Boston

University, Biostatistician.

DR. BRAWLEY: Otis Brawley, Medical

Oncology and Epidemiology, Emory University.

DR. DOROSHOW: Jim Doroshow, National

Cancer Institute.

DR. MARTINO: Thank you.

Next, I would like Ms. Clifford to read

the Conflict of Interest Statement for the group.

Conflict of Interest Statement

MS. CLIFFORD: The following announcement

addresses the issue of conflict of interest and is

made a part of the record to preclude even the

appearance of such at this meeting.

Based on the submitted agenda and all

financial interests reported by the committee

participants, it has been determined that all

interests in firms regulated by the Center for Drug

Evaluation and Research present no potential for an

appearance of a conflict of interest with the

following exceptions:

In accordance with 18 U.S.C. 208(b)(3),

full waivers have been granted to the following

participants. Please note that the following

consulting activities waived are unrelated to

Iressa and its competing products.

Dr. Silvana Martino for consulting for a

competitor, which her employer receives less than

10,001 per year.

Dr. Michael Perry for consulting with a

competitor which he receives less than 10,001 per

year. In addition, Dr. Perry has been granted a

waiver under 21 U.S.C. 505(n) for owning stock in a

competitor, valued between $5,001 to $25,000.

Because his stock interest falls below the de

minimis exception allowed under 5 CFR(b)(2), a

waiver under 18 U.S.C. 208 is not required.

Dr. Maha Hussain has been granted waivers

under 208(b)(3) and 21 U.S.C. 505(n) for owning

stock in a sponsor and a competitor. These stocks

are valued from 25,000 to 50,000 per firm.

A copy of the waiver statements may be

obtained by submitting a written request to the

Agency's Freedom of Information Office, Room 12A-30

of the Parklawn Building.

With respect to the FDA's invited industry

representative, we would like to disclose that Dr.

Antonio Grillo-Lopez is participating in this

meeting as an acting industry representative acting

on behalf of regulated industry. Dr. Grillo-Lopez

is employed by Neoplastic and Autoimmune Disease

Research.

In the event that the discussions involve

any other products or firms not related on the

agenda for which an FDA participant has a financial

interest, the participants are aware of the need to

exclude themselves from such involvement, and their

exclusion will be noted for the record.

With respect to all other participants, we

ask in the interest of fairness that they address

any current or previous financial involvement with

any firm whose products they may wish to comment

upon.

DR. MARTINO: Thank you.

Next on our agenda is Dr. Richard Pazdur,

who will address the committee and give us some

direction for this morning's meeting, please.

Opening Remarks

DR. PAZDUR: Thank you, Dr. Martino.

Iressa was originally approved by the FDA

on May 5th, 2003, as a monotherapy for the

treatment of patients with locally advanced or

metastatic non-small cell lung cancer after failure

of both platinum-based and docetaxel

chemotherapies.

Partial tumor responses occurred in

approximately 10 percent of patients. Iressa was

approved under the accelerated approval

regulations. As discussed yesterday, these

regulations allow approval based on a surrogate

endpoint reasonably likely to predict clinical

benefit and require subsequent studies to verify

and define its clinical benefit.

As an approval condition, AstraZeneca

committed to conduct a randomized trial examining

the Iressa effect on survival in patients with

advanced non-small cell lung cancer who had

received 1 to 2 prior chemotherapies. This is

defined as Trial 0709.

The primary endpoint of this trial was

overall survival and improved survival for

Iressa-treated patients was to satisfy the

requirement for the demonstration of clinical

benefit. For drugs approved under accelerated

approval, the FDA may withdraw approval for the

failure of a post-marketing study to verify

clinical benefit. I should note that there were

several studies that were included in their Phase

IV commitment.

The withdrawal procedure requires a formal

hearing whose composition and procedures are

defined in the Code of Federal Regulations. This

meeting is not that formal hearing.

AstraZeneca notified the United States

Food and Drug Administration on December 17th,

2004, that a large randomized study comparing

Iressa plus best supportive care to placebo plus

best supportive care failed to demonstrate a

survival advantage for Iressa in the treatment of

non-small cell lung cancer.

The results will be reported in detail by

AstraZeneca during this meeting.

The FDA has not received the complete data

set for this trial, especially data that would

allow pharmacogenetic or immunohistochemistry

subset analysis. The FDA management plan is rapid

communication of the above trial results to health

care professionals and patients concurrent with the

expeditious completion of the trial analysis by

AstraZeneca, including the effects of EGFR status

determined by immunohistochemistry and EGFR

mutational status on survival.

We are interested in reviewing the

immunohistochemistry subset analysis since

interesting exploratory findings were included in

the Tarceva label that was recently approved this

year.

The FDA will not make a regulatory

decision on Iressa until the data regarding subset

analysis and the study results are received and

reviewed. In the interim, AstraZeneca has

suspended promotion of Iressa, but will continue to

make the drug available to patients who appear to

be benefiting from Iressa treatment.

Actions have been taken to communicate the

most recent Iressa information to health care

professionals and patients.

These are delineated in the preamble to

the discussion points and include: AstraZeneca

press release of the ISEL study results, Dear

Doctor letters notifying physicians of the study

results and alternative therapies available,

AstraZeneca sales force distribution of Dear Doctor

letters, other Dear Doctor letters being posted on

the AstraZeneca website, patient advocate groups

being notified, AstraZeneca communications to known

patients, information being posted on the FDA

website, abstracts at meetings, journal placements

of the Dear Doctor letters, advertisements on a

continuing basis in all issues of the 10 most

widely read oncology journals urging physicians to

consider options other than Iressa.

A copy of this advertisement is attached

in today's Discussion Points.

AstraZeneca is also tracking total and new

Iressa prescriptions every two weeks to ensure that

the above communications are resulting in decreased

Iressa use.

We are not here today to vote on the

ultimate regulatory fate of this drug. We may be

bringing this question back to future ODAC meetings

after the FDA reviews this study and additional

subset analysis.

The purpose of this ODAC meeting is to

provide transparency of the process that we have

undertaken and to obtain your input on the adequacy

of these steps to date to ensure that patients and

prescribing physicians are aware of the study

results and treatment options other than Iressa

while allowing the drug to be available to patients

who may be benefiting from it.

Thank you.

DR. MARTINO: Thank you, Dr. Pazdur.

A new member has joined us. Dr. Temple,

if you would be so kind as to introduce yourself.

DR. TEMPLE: Good morning. Bob Temple,

Office Director.

DR. MARTINO: Thank you. For the

audience, as well as the committee, I want to

remind everyone that this morning's purpose is not

to decide the fate of this drug, so those of you

who are here thinking that that is what we are

going to do, please relax, that is not the point.

The point this morning is realizing that

there is some new information that needs to be

properly disseminated to the public, both the

medical public as well as the lay public, has that

process taken place and what is that process.

So, those really are the issues before

this committee.

At this point, I would like AstraZeneca to

approach the podium and introduce your speakers, as

well as give us some understanding of what they

will be speaking on please.

Sponsor Presentation - AstraZeneca L.P.

Introduction and Regulatory History

DR. SCOTT: Thank you, Dr. Martino.

My name is Mark Scott and I am the U.S.

Development Leader for Iressa.

As Dr. Pazdur just mentioned, Iressa was

granted an accelerated approval under Subpart H in

May of 2003 to treat advanced non-small cell lung

cancer after failure of two types of chemotherapy.

Subsequent to Iressa's approval, this

committee has discussed in general the terms of the

Subpart H approval guidelines and the need for

rapid completion by sponsors of their

post-marketing trials that are required as part of

such an approval.

During these discussions, an important

question was raised by ODAC, what should be done if

a confirmatory trial does not meet its primary

objective. The ODAC discussion at the time

acknowledged that there would probably be no quick

and easy answer if this situation were to arise.

We are here today because this

hypothetical situation is now real and it applies

to Iressa. The study we are here to discuss is

Trial 709, one of the confirmatory trials for

Iressa which did not achieve statistical

significance for its primary endpoint of overall

survival.

We will describe for you the actions

AstraZeneca has undertaken to communicate the

results of Trial 709 to physicians, so that

informed decisions can be made regarding the

clinical use of Iressa.

Today, we will describe important findings

from Trial 709, how the data from Trial 709 is

actually quite similar to prior clinical data on

Iressa and additional analyses, and clinical trials

that are being conducted or planned to better

understand which patients are most likely to

benefit from Iressa.

We will also outline the timings of

availability for data for FDA review, what has

occurred and the future direction for Iressa,

provide important lessons about drug development,

and accelerated approval in the era of targeted

oncology therapies.

After I cover a brief regulatory history,

Mr. Kevin Carroll will speak more on Trial 709,

then, Dr. Judy Ochs will present the actions

AstraZeneca has taken to inform the oncology

community and the implications for the development

of Iressa. I will then review the timelines that

we have to provide data to FDA.

As posed to the committee by FDA, we look

forward to hearing the Committee's thoughts on the

appropriateness of the communications taken

regarding Trial 709.

Today, we have two experts on lung cancer,

Howard Burris from Sarah Cannon and Mark Kris from

Memorial Sloan- Kettering, and they will be

supporting the AstraZeneca staff here to answer any

questions the Committee may have.

Lung cancer is the most common cancer and

the leading cause of cancer mortality in both men

and women with over 170,000 new patients being seen

each year in the United States.

The disease is complex, most patients are

diagnosed with advanced disease, symptoms are

common, and the prognosis is poor.

Standard first line therapy for advanced

disease was, and continues to be, platinum-based

doublet chemotherapy. Prior to 2003, after failure

of first line therapy, only docetaxel had been

demonstrated to improve overall survival. No

therapy had been approved for use after failure of

both first and second line therapy.

Standard chemotherapies do offer benefits,

but with significant toxicity. Therefore, there

are many lung cancer patients who cannot tolerate

any chemotherapy.

There was a great demand for new, active,

less toxic agents for non-small cell lung cancer.

Now, Iressa is a small molecule inhibitor of the

epidermal growth factor inhibitor tyrosine kinase.

EGFR expression plays a role in angiogenesis,

apoptosis, proliferation in many tumors. Iressa is

thought to mitigate against these factors.

The Iressa Phase I program began in 1998

and doses up to 1,000 mg/day were studied. Among

the 289 subjects enrolled, the most common

toxicities were low-grade diarrhea and rash, and

the dose-limiting toxicity was reversible Grade 3

diarrhea, and this dose-limiting toxicity occurred

at doses beyond 800 mg/day.

Marked anti-tumor activity was seen in

non-small cell lung cancer population that

participated in the Phase I program, and there were

actually 10 of 100 patients where responses were

noted, and these responses occurred across the dose

range.

Because of the safety findings and the

activity findings in Phase I, we chose the doses of

250 and 500 mg/day to be further investigated in

the third line monotherapy setting, as well as in

first line trials in combination with

platinum-based chemotherapy.

I will now focus on the data relevant to

the accelerated approval of Iressa.

IDEAL I and II were trials conducted among

patients where chemotherapy had failed. Both

trials randomized patients between 250 mg and 500

mg of Iressa per day. The primary endpoint in each

trial was objective response, the requirement for

response was at least a 50 percent reduction in

measurable tumor area, or significant reduction in

non-measurable disease, and these decreases needed

to persist for at least one month.

Across doses, response rates seen in IDEAL

I and IDEAL II were 19 and 10.6 percent. Responses

were durable with ranges of 13 months and 7 months

for IDEAL I and II respectively.

Also of note was the variability that was

seen in response across some subgroups. Higher

rates were seen in females, never smokers, those

with adenocarcinoma histology, and of those of

Asian ethnicity.

As you will see in a few minutes, this

same variability in response is suggested for

survival, as well, when Trial 709 was further

analyzed. There were no differences in efficacy

between the two doses, and the survival curves are

presented on this slide where we have collapsed

IDEAL I and II together and looked at 250 versus

500, and the survival curves were completely

overlapping.

As for safety, the most drug-related

adverse events were of low grade, while the most

common adverse events were rash and diarrhea.

There were a greater number of events at the 500 mg

dose. On the basis of these data, the 250 mg dose

was chosen on the basis of its efficacy and

tolerability as part of our application for

accelerated approval as a monotherapy in refractory

disease.

As Dr. Pazdur mentioned, Iressa was the

subject of the ODAC in September of 2002. These

response rate and safety data were reviewed, and

the committee voted in favor of accelerated

approval.

The FDA granted accelerated approval in

May of 2003 in patients refractory to both

docetaxel and a platinum-containing regimen. The

post-approval commitment trial started in July of

2003.

We agreed to conduct and analyze and

report on three additional clinical trials, to

examine the effects of Iressa as a monotherapy in

patient with advanced non-small cell lung cancer

where chemotherapy had failed.

These included Trial 709 where an

improvement in survival was sought, and the

preliminary results will be the focus of Mr.

Carroll's presentation today.

Trial 721 examines whether the survival

seen with Iressa is not inferior to survival seen

with docetaxel. There is a planned interim

analysis of this trial with complete data for this

to be available in June of this year, and with

survival data from this trial available in November

of next year. The results from this trial can

confirm the effectiveness of Iressa.

Trial 710, the third Subpart H commitment,

was a placebo-controlled trial where an improvement

in symptoms was sought. However, the early

availability of results from Trial 709 in December

of last year compromised the ability to recruit

patients.

As a consequence, the independent Data

Safety Monitoring Committee recommended that

further recruitment was not justified because the

trial was unlikely to be completed. In agreement

with FDA, this trial was stopped in September of

last year.

Two other trials featured as additional

commitments that were not linked to the accelerated

approval, we were asked to provide reports on the

SWOG 0023 and BR19 trials.

These placebo-controlled trials seek to

demonstrate a survival improvement for Iressa after

definitive therapy in two settings of non-small

cell lung cancer. Both trials continue to recruit.

In summary, there were three Subpart H

confirmatory trials and two additional trials. One

has been closed, three are ongoing, and I will like

to ask Mr. Kevin Carroll, the statistician for

Iressa, to come and share with you the fifth trial,

Trial 709.

Trial 709

MR. CARROLL: Thank you, Mark.

Today, I will be presenting to you

preliminary data from Trial 709, which is a large

randomized Phase III trial comparing Iressa to

placebo in advanced chemotherapy-failed non-small

cell lung cancer.

The data I will be sharing with you today

are as we saw them for the first time on December

16, 2004, and so are consistent with the materials

in your briefing document.

Since then, the data have been further

validated, in fact, were finalized on the 2nd of

February 2005. There have been few changes to

these data and none that materially affect the

results I will be showing you today.

In Trial 709, 1,692 patients were

randomized to Iressa or placebo on a 2 to 1 basis

in 210 centers across 28 countries. In light of

the approval of Iressa in the U.S.A. in May 2003,

no U.S. sites were included in this trial, as

randomization to placebo was considered infeasible.

Further, to ensure balance between the

treatments at baseline, the randomization was

stratified for histology, gender, reason for

failure to prior chemotherapy, and smoking history.

In terms of key eligibility criteria, the

patients randomized into Trial 709 had advanced

non-small cell lung cancer and had failed 1 to 2

prior chemotherapy regimens.

Furthermore, the patient population

entered into Trial 709 was highly refractory since

the patients had either to be intolerant to their

most recent chemotherapy or had to have progressed

on or within 90 days of their last chemotherapy

cycle.

In Trial 709, as has been said, the

primary endpoint was overall survival. As stated

in the protocol, the primary analysis method was a

stratified log-rank test. As is common in oncology

trials, the protocol also stated that a supportive

Cox regression analysis would be conducted.

There were 2 co-primary populations for

analysis, the overall population and a subset of

patients with adenocarcinoma histology. At least

900 deaths were required overall to provide 90

percent power.

The secondary endpoints are listed on this

slide, being time to treatment failure, objective

response, quality of life, symptoms, and safety.

Several subgroup analyses were pre-planned

with the aim being to examine outcomes in relation

to important clinical and biologic factors, such as

EGFR expression and EGFR mutations, and my

colleague, Dr. Ochs, will say more about this later

in our presentation.

The data I will be presenting today are

all those that accrued up to and including the end

of October 2004. This date was chosen because it

was estimated by this time the 900 deaths we needed

for analysis would have occurred on the database.

So, following data collection, preliminary

data became available for the first time in

mid-December 2004. At this time, median follow up

was 7 months, and we knew of 969 patient deaths.

As can be seen on this slide, patients in

Trial 709 were recruited mainly from Central and

Eastern Europe and then Asia. As I mentioned

before, there were no U.S. sites in Trial 709, and

due to the approval of Iressa in December 2003,

only 1 percent of patients were recruited in

Canada.

This slide shows the baseline

characteristics of the patients in Trial 709. The

median age was 62 years, about two-thirds were

male, one-fifth were never smokers, one-fifth were

of Asian descent, about half had adenocarcinoma

histology, and about half had received one prior

chemotherapy.

In line with our intent to recruit a

highly refractory patient population, 90 percent of

the patients in 709 had progressed on or within 90

days of their most recent chemotherapy. Finally,

as you would expect in a large randomized clinical

trial, the treatment groups were well balanced at

baseline.

I would like to move on now to look at

survival in the overall population. As you can

see, there was some improvement in overall survival

in Iressa-treated patients with the Kaplan-Meier

curves separating after about 4 months. However,

the magnitude of that improvement was not

sufficient to reach statistical significance in the

primary stratified log-rank test, however, the

supportive Cox regression analysis did suggest

statistical significance.

Here are the survival curves for the

co-primary population of patients with

adenocarcinoma histology. Again, there was some

improvement in overall survival in Iressa-treated

patients, but the magnitude of that improvement was

not sufficient to reach statistical significance on

the primary stratified log-rank test.

Again, here, the supportive Cox regression

analysis did suggest statistical significance.

Moving on now to secondary endpoint data,

tumor shrinkage in terms of response rates was

significantly greater in Iressa-treated patients

compared to placebo.

In terms of the time to treatment failure

being the time from randomization to the first

event that led to the cessation of randomized

treatment, there was a statistically significant

difference between the treatments with the risk of

treatment failure being 18 percent lower in

Iressa-treated patients compared to placebo.

The reasons for treatment failure are

shown on this slide. As can be seen, the primary

driver for treatment failure was progression be it

either symptomatic or radiographic, with

approximately 56 percent progressing on Iressa

compared to 70 percent progressing on placebo.

As you would expect, Iressa failed more

often due to adverse events than placebo, and Other

on this slide refers to a number of items including

lost to follow-up, noncompliance, and withdrawal of

consent. As you can see, there was no difference

between the two treatments in this regard.

Turning now to quality-of-life data, the

analyses of these data is currently ongoing, but I

can share with you some initial results. As you

can see, the primary quality of life endpoints

being symptoms, overall quality of life, and trial

outcome index, all tended to favor Iressa-treated

patients although the treatment differences were

relatively small.

As I mentioned before, several subgroup

analyses were pre-planned. Now, before I run

through these data with you, it is important to

emphasize that these analyses are not

retrospective, nor are they data driven.

The subsets were identified in advance

based on what we saw in our Phase II trials and

based upon findings on other drugs in the same

class.

Furthermore, in analyzing these subsets,

we have applied a rigorous statistical approach

whereby we looked first for evidence of a subset by

treatment interaction to give us confidence that

the subsets are truly behaving differently, and if

evidence exists, then, we go on to look at detail

at the subsets.

It is important to recognize that this is

a harder test to pass than simply having a list of

subsets and looking for p less than 0.05. So, if

we do see differences in Trial 709, we can be

reasonably confident that they are more likely due

to a real drug effect than due to chance alone.

This is the first of two slides that show

subset analyses. For each subset analyzed, you can

see the hazard ratio and its confidence limits and

the response rate in Iressa-treated patients.

As you will recall, the hazard ratio

measures the risk of death on Iressa-treated

patients to placebo-treated patients, and

therefore, a hazard ratio of less than 1 to the

left of the vertical line shows a treatment effect

in favor of Iressa, and a hazard ratio to the right

of the vertical line shows a treatment effect in

favor of placebo.

So, now while no subgroup favored placebo,

there was clearly some variability in survival

outcome. This was most marked in terms of smoking

history where outcomes in never smokers was

statistically different than outcomes in ever

smokers.

This is the second slide showing data in

subsets, the same format as the previous slide.

Again, you can see there was variability in

outcomes with, in this instance, it being most

marked in terms of ethnicity where patients of

Asian ethnic origin have statistically different

outcomes to patients of non-Asian ethnic origin.

Now, while the credibility of subset

analyses is always a matter of debate in any

clinical trial, in 709, the rigorous approach we

have taken provides us with confidence that the

differences we have seen are most likely due to a

real effect of the drug, and less likely due to

chance.

So, the findings we have seen in Asians

and on smokers are therefore supported

statistically by the presence of subset by

treatment interactions and also clinically by prior

Phase II data that have consistently shown

increased response rates in these populations.

Furthermore, Trial 709 is internally

consistent with respect to these subsets, with

better time to treatment failure and a two-fold

improvement in quality of life in Iressa-treated

patients.

This slide shows survival curves for never

and ever smokers. As you can see, there was a 33

percent reduction in the risk of death in never

smoking patients treated with Iressa compared to

placebo. There was no significant difference in

ever smokers.

Similarly, this slide shows survival

curves by ethnic origin. Again, you can see there

was a 34 percent reduction in the risk of death in

Asian patients treated with Iressa compared to

placebo, and there was no significant difference in

non-Asian patients.

I would like to move on now to look

briefly at the safety data in Trial 709. I should

note these data have become available since we

compiled the briefing document, so they won't be in

your papers.

The adverse event profile in Trial 709 is

consistent with the established safety profile for

Iressa with the most common adverse events being

rash and diarrhea.

Notably, there was little difference

between the treatments in terms of serious adverse

events, adverse events leading to withdrawal, and

the incidence of interstitial lung disease.

Here is a summary of the most common

adverse events in the trial ordered from highest to

lowest frequency in Iressa-treated subjects.

As you can see, with the exception of rash

and diarrhea, which I just mentioned, there is

little difference between Iressa and

placebo-treated patients in terms of the adverse

event reporting. In particular, there were

relatively few Grade 3/4 adverse events in

Iressa-treated subjects.

This list of adverse events continues on

this slide where again it can be seen there is

little difference between Iressa and

placebo-treated subjects.

As I mentioned at the outset, the

preliminary data we saw on December 16th were

validated and finalized as of the 2nd of February

2005. These final data confirmed a total of 976

deaths occurring on or before the October 2004 data

cutoff. With only 7 additional deaths, it is

obviously not surprising that the findings based on

the preliminary data remain unchanged.

On reviewing the data in December, the

Independent Data Monitoring Committee recommended

that further follow-up of Trial 709 should be

obtained. Having seen somewhat late separation in

the Kaplan-Meier curve, they were unwilling to rule

out that further separation could occur with more

follow-up.

Hence, survival data were updated as of

the end of January, which provided for a further 3

months of follow-up, taking median follow-up to 10

months and overall mortality in the trial to 70

percent.

As you can see, these further data are

consistent with the planned protocol analysis, and

despite increased crossover in the placebo arm to

Iressa, variability in survival outcomes continues

to be seen.

To briefly summarize what we have shared

today, the data seen on December 16 showed some

improvement in survival in Iressa-treated patients,

but the magnitude of that improvement was not

sufficient to reach statistical significance in the

primary stratified log-rank test.

Overall, however, considering both primary

and secondary endpoints, these data showed that

Iressa was efficacious in the population study, but

there was marked variability in survival outcomes.

So, with that, I would like to thank you

for your attention and hand over to my colleague,

Dr. Ochs. Judy.

Clinical Actions and Implications

DR. OCHS: Thank you, Kevin.

In this part of our presentation, I would

like to briefly summarize AstraZeneca's actions to

communicate the results of Trial 709 to the

oncology community. Following this, I would like

to give an overview of the clinical implications of

the Trial 709 data, review some of the immediately

relevant emerging science, and conclude with our

proposed or ongoing development proposals.

In agreement with the FDA, AstraZeneca

concluded that it was in the best interest of

patients that the information on Trial 709 be

rapidly, extensively, and clearly communicated.

On December 17th, a Dear Doctor letter

approved by the FDA was distributed by AstraZeneca.

This communication provided physicians with the

needed information to enable them to make the most

appropriate treatment decisions. The expectation

was that this communication would greatly reduce

the number of patients receiving Iressa for the

first time.

In addition, AstraZeneca would provide to

the FDA, prescription data every two weeks to be

able to assess the continuing impact of the

communications.

It was also agreed that a key goal was to

maintain Iressa availability to those patients

already benefiting who would wish to continue and

had concerns about possible Iressa availability.

A commitment was given to the FDA that

AstraZeneca would rapidly provide them with all of

the data as it became available to allow them a

thorough and informed analysis.

Upon public release of the top line Trial

709 survival results a series of extensive

communications were simultaneously begun and are

listed on this slide, and were previously mentioned

by Dr. Pazdur.

Taken as a whole these actions were

designed to ensure that relevant physicians would

be aware of the results and be reminded that

alternative therapeutic options with proven

survival benefits should be considered.

On January 6, the FDA and AstraZeneca met

and agreed upon the following steps for continuing

communication of the Trial 709 data. A public

disclosure of the then available results would be

made at the first available scheduled ODAC meeting,

today, acknowledging that the further trial data

would still be pending.

Ongoing communication of the Dear Doctor

letter was to be done using journal placement and

the full clinical data would be submitted and

presented at scientific meetings and published in

refereed scientific journals as soon as possible.

Abstracts have been submitted to the AACR

meeting, as well as the World Lung Cancer

Conference. A full publication submission is

planed in the May-June time frame.

Here is a copy of the Dear Doctor letter,

which I realize you cannot read. The letter,

however, does include the survival results in the

overall and adenocarcinoma subpopulation along with

median survival and respective hazard ratios.

The sentence highlighted in red above is

included in the body of the letter and urges

physicians to consider other treatment options.

This is how the letter is being displayed in the 10

most widely read oncology journals, and a list of

these journals is shown in the next slide.

The impact on Iressa usage has been marked

in the 10 weeks since the Dear Doctor letter was

first sent out. There has already been a

significant reduction in the prescriptions written

for Iressa, and our internal AZ usage data also

indicates marked reduction.

Market research, that we have just

obtained from 100 community oncologists, indicates

that the great majority are aware of the data

contained in the Dear Doctor letter and have

modified their treatment practice accordingly.

Thus, all of the agreed upon communication

actions have been set in motion, and the available

information suggests that the oncology community is

aware of and acting on the information.

The larger question is now being asked:

What are the clinical implications of the Trial 709

data, and what are the next steps? These are

clearly important questions for oncologists and

patients since Iressa possesses significant durable

anti-tumor activity which has greatly benefited

some patients and some patient subsets.

Yet, in Trial 709, Iressa did not meet the

statistically defined survival endpoint in an

unselected patient population.

Advances in understanding of the molecular

biology in this area of EGFR inhibition, as well as

in the area of non-small cell lung cancer, are

occurring rapidly and have the potential to better

select or predict those patients who would benefit

beyond, or in addition to, clinical

characteristics.

What are the questions that we are asking

as we seek to understand the Trial 709 outcomes,

and not wrongly or prematurely make conclusions

about the actual role or place of Iressa, an agent

with anti-tumor activity in the treatment of a

disease with a continuing poor prognosis? Why did

this result occur?

How does this result compare with our

other data on Iressa in non-small cell? Were the

findings in our trial due to play of chance? Was

the dose selection appropriate? Were there

methodologic issues, such as the trial population

and where the trial was conducted of any potential

impact on the findings?

What biologic data may be available now

and in the future to help better understand the

clinical outcomes, and what further relevant

clinical data in the recurrent non-small cell lung

cancer setting are expected?

Firstly, how does the survival outcome

seen in Trial 709 compare to other data with

Iressa? As was previously mentioned by Dr. Scott

in our Phase II program, a striking and

unanticipated finding was the apparent high rate of

response in patients with certain clinical

characteristics.

It can be seen if one compares these Phase

II response rates with those in Trial 709, and the

Phase II results are in the right-hand column in

yellow, and the 709 results in the middle column in

white, that the same patient groups continued to

show higher response rates.

In addition to these higher response

rates, the subgroups having the highest response

rates experienced the greatest benefit in survival.

The patient subgroup with the highest response rate

were the never smokers, and as previously noted,

the survival in this subgroup was significantly

increased.

Similar trend, although not of the same

magnitude, of survival benefit was seen in women

and with the adenocarcinoma group.

Continuing with this line of inquiry,

higher response rates and statistically significant

survival results and benefit were seen in those

patients of Asian descent.

Could chance have played a role as the

defined survival endpoint was so narrowly missed?

Trial 709 and the erlotinib trial BR21 are the only

two Phase III survival trials which compare an oral

EGFR inhibitor with placebo in the recurrent

non-small cell lung cancer patient population.

Both Iressa and erlotinib have similar

overall response rates as can be seen in the

right-hand portion of the slide. The erlotinib

trial did reach statistical significance for the

overall population.

Juxtaposing overall survival hazard ratios

as we have done in this slide shows that while the

point estimates differ, there is a high degree of

overlap in the confidence intervals. The small

confidence interval in Trial 709 reflects the

larger trial size in 709, which is almost twice

that of the BR21 trial.

Dose selection. Since there appears to be

a difference in magnitude of survival benefit in

BR21 compared to Trial 709, questions about the

adequacy of the Iressa dose have arisen

irrespective of the data used to support its use in

this trial.

The erlotinib dose used was at the maximal

tolerated dose, while the Iressa dose is one-third

the maximal tolerated dose, reflecting different

development strategies.

As you might guess, we have gone back and

re-evaluated our prior experience in light of the

current data. Our extensive Phase I program had 280

patients, and these patients received doses ranging

from 50 mg to 1,000 mg.

Responses and durable stable disease first

were seen at the 150 mg dose level. There was no

dose response evident from 150 mg through 1,000 mg

with respect to partial response rates, partial

response rates plus stable disease rates, or the

duration on Iressa therapy.

In our Phase II trials, as previously

mentioned, we formally compared the 250 and 500 mg

dosage. 250 mg was chosen as it was above the 150

minimum dose that we saw responses and stable

disease at, and 500 mg dose was chosen in part

because of minimizing the amount of patient

interruptions of therapy due to toxicity.

We found no difference in efficacy

including survival although the adverse events and

therapy interruptions were more frequent at the

higher 500 mg dose.

Admittedly, however, we have not

rigorously evaluated doses above 500 mg, and it is

unknown if doses above 500 mg would achieve better

overall or patient subset survival outcomes. Due

to the lack of data, we cannot rule this out

entirely.

Speculatively, can the inability to

achieve statistically significant survival be

explained by too few patients likely to benefit

based on their advanced disease status with

refractoriness as specified in our patient

inclusion criteria.

Another area to further explore are the

impact of environmental factors, such as smoking,

as it relates to various geographic regions where

the trial was conducted.

As Mr. Carroll showed, over one-third of

the patients on Trial 709 were from Eastern Europe

where the median pack year exposure was very high.

Patients with the highest smoking exposure appear

less likely to benefit from EGFR inhibitor therapy.

We have looked at our data and found a

continuous spectrum in terms of survival benefit,

with the greatest survival benefit appearing in

never smokings, but it continues with the amount of

exposure to smoke.

So, what can we conclude at this point?

Iressa is an active agent, the response data are

consistent in our Phase II and III trials. The

patients most likely to benefit are those patients

who never smoked and those of Asian ethnicity.

With these consistent findings, using an

agent that inhibits a specific receptor and

pathway, it is logical to assume that there is an

underlying biologic basis. In the last 10 months,

two areas of translational research have been

fruitful and may be useful in better understanding

the clinical data in our Phase III program in Trial

709, as well as guide therapy in our future

development.

The two biomarkers of most promise

currently are EGFR expression and the EGFR

mutations. Published Iressa Phase II data did not

appear to show definitive correlation of EGFR

expression with response, but tumor samples were

not available from all patients, and the trials

were not controlled.

Recently, however, results relating EGFR

expression to survival outcomes were included in

the erlotinib label.

The second promising biomarker are

activating mutations. These were first described

approximately 10 months ago in responding Iressa

patients. There are other promising, but more

exploratory biomarkers that are included in the

Iressa science program including gene copy number

and dimerization patients, but again these remain

more exploratory.

What I would like to do now is show you

from the erlotinib label--and I have included the

three graphs they have relating to EGFR

expression--and to ensure perfect synchronicity and

accuracy, I am going to read the portion for you

for all of those of you who can't read the lower

right-hand column.

What we see here are three graphs. The

graph to the upper far left is the graph of the

patients who had positive EGFR expression in their

tumors, with the lower part of the Kaplan-Meier

showing the patients treated with placebo.

The graph to your far right, on the

upperhand side is the patients who were EGFR

expression-negative compared to placebo. The lower

lefthand is those patients that they did not have

EGFR expression data on.

As stated in the label, Tarceva prolonged

survival in the EGFR-positive subgroup and the

subgroup whose EGFR status was unmeasured, but did

not appear to have an effect on survival in the

EGFR-negative subgroup. However, the confidence

intervals for the EGFR-positive, negative, and

unmeasured subgroups are wide and overlap, so that

a survival benefit due to Tarceva in the

EGFR-negative subgroup cannot be excluded.

It needs to be said that a positive EGFR

expression status in this study was defined as

having at least 10 percent of cells staining

positive for EGFR in contrast to the 1 percent

cutoff specified in the DAKO EGFR pharmDx kit

instructions.

The use of the pharmDx kit has not been

validated for use in non-small cell. Accordingly,

the data to date are inconclusive, but tantalizing

as to the predictive nature of EGFR testing.

In this trial, as in Trial 709, the tumor

sample collection was not mandatory, and thus the

number of samples is less than the number of

enrolled patients.

This is a busy slide and summarizes a very

busy area of research in the 10 months since

mutations were first described. As noted on this

slide, mutation appears to occur almost exclusively

in non-small cell lung cancer. The mutation is

activating and in the ATP-binding site, which is

where Iressa's activity occurs.

I mentioned that the mutation was first

described in patients with rapid, dramatic and

prolonged responses to Iressa. The increased

frequency of the mutation occurs in patient subsets

where Iressa responses are most frequent and where

the survival benefit is most likely to be seen,

that is, those patients who were never smokers,

patients of Asian descent, women, and

adenocarcinoma histology.

There are actually two papers out this

week looking at smoking status in relationship to

the presence of these activating mutations, and

depending on the paper, a minimum of 25 percent to

75 percent of patients in different geographic

regions who were never smokers have the mutation

present.

While patients whose tumors possess this

type of somatic mutation appeared to be much more

likely to have a response, all patients with

mutations do not have a response. We have recently

looked at our IDEAL II data, and in the small

subset with 14 mutations that we detected, 6 of

these patients had prolonged partial responses.

Again, where are we? EGFR expression

appears to be associated with increased survival.

EGFR mutations appear to explain some, but not all,

of the responses to Iressa.

Outcomes in Trial 709, comparing Iressa to

placebo, will be explored in terms of EGFR

expression, activating mutation, and other

biomarkers.

We anticipate that this data will be

available in June 2005. We have collected close to

600 tumor samples. Approximately 400 of them we

estimated based on our past experience will be

fully evaluable for EGFR expression, and 200 for

mutation status.

It is hoped that these results may provide

further insight into the clinical outcomes that we

have seen in Trial 709.

Thus, with these current clinical and

translational data, what prospective studies are

underway or could be considered?

One proposal would be to evaluate patients

with metastatic disease and compare the outcomes of

Iressa with chemotherapy. Mandatory tissue

collection is an obvious requirement to evaluate

the utility of biomarkers with respect to both

outcomes in both the chemotherapy-treated patients

and in patients with EGFR expression or

overexpression.

Targeted studies in patient populations is

another obvious way to proceed. We have an ongoing

Phase II trial which is enrolling patients who are

mutation-positive, another trial that is a trial

that should be considered is that in patients who

are never smokers.

Never smokers represent 20 percent of the

U.S. population of non-small cell lung cancer

patients.

Finally, specific trials in the Asian

populations to define the role of Iressa in the

first line setting appear warranted. Here, too,

translational studies would be integral to the

trial. There are already several trials being

conducted in Asia, as you might anticipate.

A clinical question of increasing

relevance that hasn't been answered to date is that

of comparing both survival outcome and toxicities

of Iressa or any EGFR inhibitor with single agent

chemotherapy.

Trial 721, as previously noted by Dr.

Scott, is a randomized Phase III post-approval

commitment trial which compares Iressa to

docetaxel. This trial will complete patient

enrollment by the end of the summer, and an interim

survival analysis is expected this May or June.

Trial 721's principal investigators and

steering committees have reviewed the Trial 709

data and continue to support this trial. Another

similar trial is being conducted entirely in Japan.

Some clinical support to continue at this

dose is mature Phase II data in a Caucasian and

Hispanic patient population, which has recently

matured and become available. In addition to

showing comparability with the primary symptom

endpoint, comparable outcomes were seen with

response rates, time to progression, and overall

survival.

The next slide is a Kaplan-Meier survival

curve from this trial, and it is easy to see the

comparability of these trial results. With a

median follow-up of 9 months and 55 percent overall

mortality, there are no differences between Iressa

and docetaxel.

The overall survival with docetaxel is

consistent with that previously reported with this

agent and in this clinical setting. Although the

trial is small, if Iressa was behaving as a

placebo, then, one would have expected Iressa to

have performed substantially worse in both time to

progression, as well as overall survival.

Back to our original question: Where are

we now?

Well tolerated agents in the EGFR

inhibitor class of agents are now an accepted

addition to the therapeutic armamentarium of

practicing oncologists and clinical trial

investigators.

Clinical and translational data are

pointing the way to the most appropriate and

optimal use of Iressa. AstraZeneca and our

clinical investigators remain committed to this and

other biologically targeted agents as the way to

the future.

Thank you.

DR. PAZDUR: Silvana, I am sorry, I didn't

realize there was more from AstraZeneca, but if we

want to have some discussion or clarification

before the open public hearing, that would be

appropriate.

Summary

DR. SCOTT: Thank you, Dr. Ochs.

As you have heard from both Mr. Carroll

and Dr. Ochs, there is a lot of work ongoing to

fully understand Trial 709, and there are other

trials, such as Trial 503, that provide supportive

information, and Trial 721, which is also part of

our Subpart H commitment to the FDA.

This slide summarizes some key milestones

that will be occurring. It is expected that the

complete data from Trial 709 and Trial 503 will be

with the FDA in June for their review. After that

time, we expect to discuss labeling updates as

appropriate based on the final data findings.

It is expected that the Subpart H

commitment trial, Trial 721, will deliver its final

survival data in November of 2006.

While the drug development road for Iressa

has not been straightforward or without its

surprises, the development program for this agent

has provided a great deal of valuable information

about non-small cell lung cancer and the EGFR

target.

Iressa is an active and well tolerated

agent, and the lung cancer community has urged us

to continue the development of this drug, and we

are committed to doing so.

Trial 709 has provided important patient

selection information in a controlled randomized

setting that may in the future help us write

appropriate labeling to guide the clinical use of

Iressa.

You have also heard today the critical

information regarding EGFR expression and mutation

status is yet to be delivered from this trial.

Trial 721, the head-to-head trial versus

docetaxel can provide confirmatory evidence of the

effectiveness of Iressa. As outlined, the

development program for Iressa will help in

identifying those patients who are most likely to

benefit from Iressa.

AstraZeneca rapidly and thoroughly

disseminated information to oncologists about Trial

709 to ensure informed treatment decisions would be

made while further analysis were underway.

As the patients responsive to Iressa will

tell us, and their physicians will support, Iressa

remains an important treatment option for non-small

cell lung cancer.

We thank the committee for their attention

and welcome any questions at this time.

Committee Questions

DR. MARTINO: Dr. Ochs, do one thing for

me before you leave. A slide was shown by--Dr.

Ochs actually had the slide up--where you

demonstrated what has been done to disseminate this

information, if you would just flash that one more

time.

I will allow the committee to ask

questions. We actually have no time allotted for

that, so please keep your questions pertinent to

today's issue, which really is has this information

been appropriately disseminated.

The slide that I want you all to just

notice are the things that they have, in fact, done

to disseminate this information. Rick, can you

simultaneously just remind the group what the FDA

has done from its side in terms of disseminating

the information, so that everyone is sort of up to

date.

DR. PAZDUR: We have notified shortly when

we were in receipt of this information, an e-mail

went out from the FDA to ASCO members notifying

them on that day that we received the information

of the study results and alternative treatment.

We have a letter posted on our website

that is included in your packet.

DR. MARTINO: So, then, from the FDA's

side, the information has gone out to physicians

primarily, as well as the website.

DR. PAZDUR: Correct.

DR. MARTINO: And from AstraZeneca,

information has been provided to physicians, as

well as to the lay public.

DR. OCHS: Yes.

DR. MARTINO: At this point, I will take

some questions, but please keep them brief and

succinct.

Dr. Hussain, you are first.

DR. HUSSAIN: It is a question to either

Dr. Mark or Dr. Ochs. When you pointed out that

you are possibly thinking about targeted

population, I saw that there were no women

mentioned and no adenocarcinoma.

Does that mean if you were a smoker and a

woman, that the smoker component takes over as far

as your potential benefit, or that if you are an

Asian and a smoker, then, the smoker takes over?

DR. OCHS: I think I will let Dr. Carroll

discuss that particular issue since there is a lot

of interconnection and interplay.

MR. CARROLL: Thank you for your question.

Of course, it is very important, I mean it is one

that we need to look at more closely, what is the

interplay between the factors of interest, be they

Asians, be adenocarcinoma, gender.

The data, as I said, were finalized

only--I am not sure--four or so weeks ago, and that

kind of analysis requires multivariate analysis to

actually see which factors are contributing, which

are the ones that are predicting the treatment

effect.

That is something that we do plan to do in

the next coming weeks and months to provide that

data to the FDA, so we can answer the very

important question that you have raised, because I

am not sure we have the answer to that today.

DR. MARTINO: Dr. Perry.

DR. PERRY: I am not sure who gets this

question, but I have been under the impression that

in Europe particularly, the incidence of squamous

cell carcinoma was considerably higher than in the

United States, so I am somewhat surprised of a 48

percent incidence of adenocarcinoma histology

worldwide, particularly when two-thirds of the

patients seem to be from Europe.

How do you know that these are

adenocarcinomas, is this the local pathologist's

interpretation, and are they inclined to overread

them as adenocarcinomas rather than as non-small

cell carcinomas not otherwise specified?

DR. SCOTT: I will ask Dr. Alan Barge to

come and speak to that point.

DR. BARGE: Thank you. Alan Barge,

AstraZeneca.

We have not done central pathology review.

All of the diagnoses were the ones that were

confirmed by the hospital pathologists, so we

couldn't answer your question directly, I am

afraid.

DR. MARTINO: Dr. Rodriguez.

DR. RODRIGUEZ: I just wanted some

clarification about the actual trial design, and

just have a few questions which might be relevant

because this was done by a variety of cultural

groups.

Were the patients and the investigators

both blinded to the assignment to placebo?

DR. SCOTT: Yes, it was a randomized

double-blind trial.

DR. RODRIGUEZ: How was compliance

confirmed in the participants?

DR. SCOTT: Nick Botwood will come to the

stand.

DR. BOTWOOD: Thank you. Nick Botwood,

AstraZeneca. We did look at compliance on this

trial and found that over 90 percent of the

patients were compliant and had taken at least 95

percent of their medication.

This was based primarily on data that we

collected in the CRF in terms of any documented

dose interruptions for whatever reason, and then we

went on to further validate that, to actually look

at the number of tablets that were returned and

looked at the number of tablets that had actually

been prescribed to validate that what was in the

CRF was actually the correct information.

DR. RODRIGUEZ: Along those lines, was

there a required or concurrent diarrhea prophylaxis

program, and was compliance to that also monitored?

DR. BOTWOOD: That wasn't, no.

DR. RODRIGUEZ: It is interesting because

your failure to treatment has a significantly

different profile with regards to symptoms and

adverse events. It seems that the patients on the

Iressa arm, a higher proportion were taken off

study because of those problems, is that correct?

That is what your bar graph seemed to show.

DR. BOTWOOD: Kevin Carroll can answer

that question, please.

MR. CARROLL: If I am correct, you are

asking whether there was a difference in withdrawal

due to--

DR. RODRIGUEZ: Yes, side effects.

MR. CARROLL: Side effects. As I showed

when we went through the adverse event data, there

was very little difference between the two

treatments in terms of withdrawal due to adverse

events, and in terms of the data that we obtained

on time to treatment failure, there were, in fact,

fewer--Iressa failed fewer patients due to

progression than placebo, so I don't think the

difference was there in the way that perhaps you

think.

DR. MARTINO: Dr. Levine.

DR. LEVINE: I also have several

questions. First, you mentioned crossover. How

many of these placebo patients did cross over to

Iressa?

DR. SCOTT: Dr. Botwood.

DR. BOTWOOD: Yes, thank you. The rate of

crossover from placebo to Iressa in this trial was

only 3 percent.

DR. LEVINE: Three. Do you have data on

other treatment beyond, you know, crossover to

anything?

DR. BOTWOOD: Yes, we do. The number of

patients that went on to receive any subsequent

chemotherapy was 10 percent, and this was balanced

between the Iressa and placebo arm.

DR. LEVINE: Just to further that a little

bit, even complementary therapies in Asia, and so

forth, do you have data on that, green tea?

DR. BOTWOOD: It was extremely small.

DR. LEVINE: My other question related to

the concept of secondary smoke. In Eastern Europe

and in Asia, where so many of the population smoke,

even individuals who say that they weren't smokers

may have been exposed, and therefore, did you look

at cotinine levels or anything? That might be

something to explore, or did you look at that?

DR. BARGE: I am afraid we haven't looked

at that. When we looked at the smoking demography

of the patients from Eastern Europe, approximately

85 percent of the patients from Eastern Europe were

heavy smokers, and they had a higher median year

exposure than the patients from other regions, but

that is as far as we got, I am afraid.

DR. MARTINO: Dr. D'Agostino?

DR. D'AGOSTINO: Yes. I am having a hard

time keeping my questions solely to the material

that has been circulated as opposed to asking a

million questions about the study, but the question

I do have in terms of the reporting of the data,

you may have said it, and I am sorry if I missed

it, you did have the expected number of deaths, you

wanted 690 or whatever, 960, and you got more.

I understand that the study did run its

course, and then you did an analysis with unclean

data or not completely clean data, and then later

on had clean data, or was the analysis you are

reporting an interim analysis?

DR. SCOTT: The analysis that we reported

on in December was not an interim analysis. It was

based on final survival data, but it had been yet

to be validated.

DR. D'AGOSTINO: Okay. So, it was the

validation. Thank you.

DR. MARTINO: Dr. Proschan.

DR. PROSCHAN: You mentioned that the

Phase II trials identified subgroups, and ethnicity

was one of them. I am wondering if, at that time,

you specifically decided to break it into Asian

versus non-Asian, and why do you think there is a

difference?

DR. SCOTT: We can have Mr. Carroll talk

about the rationale behind the subgroups that we

planned for Trial 709, and then perhaps have Alan

Barge talk about why we think so.

MR. CARROLL: The subsets that we have

looked at in Trial 709, all of them we have shared

with you today, I have not shared a subset of the

subsets, of course, and they were determined

primarily by what we saw in our Phase II data, and

also information that came out in June of this year

on the BR21 trial, as described by Dr. Ochs.

There, there was an evaluation of Asians

and non-Asians, and that, in addition to our

findings in the IDEAL trials where our Japanese

patients had a much higher response rate was a

motivation to look at that subset amongst others

that were deemed to be clinically relevant.

Perhaps I can now turn to my colleague,

Dr. Barge, to answer the second part of your

question.

DR. BARGE: Yes, thank you. There is a

good deal of speculation as to why patients of

Asian ethnic origin appear to do better on this

class of drug. There have been some quite

interesting publications very recently. In fact,

this week there was a publication from Dr. Gazda

[ph] at UT-Southwestern. His group showed that the

frequency of activating mutations of the kind that

Dr. Ochs described is much higher in Asian

populations, particularly female Asians, and

particularly female Asians with adenocarcinoma.

The Phase II studies that we conducted in

various Asian countries all show that the frequency

of responses are much higher in those populations,

and we have seen response rates as high as 60 or

even 80 percent in selected populations of Asian

nonsmoking females.

Whether or not that is all driven by

activating mutations we don't know, but that is

certainly a very strong hypothesis at the moment.

DR. MARTINO: Mrs. Ross.

MRS. ROSS: Thank you, Madam Chair.

If I understand correctly, the primary

purpose of this hearing is to evaluate or just to

discuss the transparency of the post-approval

process and the adequacy of the notifications.

In that regard, I would like to advise the

rest of the panel of other steps that were indeed

taken by both AstraZeneca and the FDA, and I would

like to thank Dr. Pazdur in particular for his help

on this.

As the only lung cancer advocacy

organization nationwide, we started receiving many

phone calls from patients who were somewhat

panicked when they heard the news in the press that

Iressa might be pulled. The press always leaps to

the worst possible conclusion, as you all know.

These people were discussing stockpiling

drugs, buying them in Japan. There was a lot of

panic out there. Dr. Pazdur responded, and

AstraZeneca did, by helping us draft more

information, more plain English information to put

up on our website and to tell people over the phone

when they called in a state of panic about Iressa.

I think that should be noted. I think

that overall, the process was extraordinarily

transparent and more than adequate in dealing with

the situation. Again, I would just like to thank

FDA and AstraZeneca for all they did.

DR. MARTINO: Dr. Temple, did you want to

make a comment?

DR. TEMPLE: Just one question. We

certainly never at any time thought that someone

who had apparently responded to the drug should

lose access to it. That was never in doubt.

But I wanted to ask you about where

AstraZeneca is at the moment. This was, shall we

say, an optimistic presentation. The study, after

all, failed. You had opportunities to identify

subsets before the study that would be your primary

analysis, but you didn't think that they were good

enough to do that.

So, these are now--it's an important

distinction, Ralph may want to comment more--these

are after-the-fact subset analyses in a study that

did not win. That is different from subset

analyses in a study that did win.

But what I really want to know is where do

you come out on the question of new patients with

non-small cell lung cancer being started on Iressa

now. The material you put out says you should

consider other drugs. Fine.

But would it be your view that at the

present time, optimism about the future and data

that might come forward notwithstanding, a person

with this disease should really not be started on

Iressa, would that be your view, or is that not

your view anymore?

DR. SCOTT: Our view is that what was

stated in the Dear Doctor letter then is what is

today, that physicians should consider other

options armed with the information from this

particular trial.

If I could ask Dr. Kris to come up and

talk about how this has played out in his practice,

maybe Dr. Burris, as well.

DR. KRIS: To answer your question, Dr.

Temple, I think the most important thing is to put

this into a context of what is available for a

patient with advanced non-small cell lung cancer

particularly after the failure of initial therapy.

I think that the information that we have

today is that there are some patients, those with

an EGFR mutation, that have, and the literature

today says that they have an 89 percent chance of

having a response, and in those patients, when you

look at their duration of response and survival, it

is clearly prolonged. In the trials that looked at

survival in mutation positive and negative people

being treated, it is much better with treatment.

So, as a clinician, my first point is to

find those people that have that extraordinary

chance of benefit, that is, mutation-positive

people, and the two surrogates for positive

mutation we have today, that is, never smoking

status for U.S. population and worldwide, is

probably Asian, and it is not simply Japanese.

There are reports now from Taiwan, from China,

Singapore.

DR. TEMPLE: Let me be clear, though. You

are looking at the mutation status of the people

in--some of the people anyway, about 200 you

said--in the trial, and maybe that will be

overwhelming and knock everybody's eyes out.

But at the moment you have no prospective

data on that subgroup for survival.

DR. KRIS: The only prospective data that

exists on the treatment of mutation-positive

patients is, frankly, an extrapolation to the never

smoking patients.

DR. TEMPLE: I understand.

DR. KRIS: But those are

placebo-controlled trials.

DR. TEMPLE: But what I am really asking

is what you really mean, and we will probably have

to have subsequent discussions, one might say that

you should use the drug with very similar

properties, similar mechanism, et cetera, that has

actually been shown to improve survival.

Are you saying something to the contrary

or not? I don't think it is clear yet. I sort of

thought it was clear, but from your presentation, I

don't.

DR. KRIS: Well, I frankly think that the

most critical slide there was looking at the hazard

ratios for the two substances, for gefitinib and

erlotinib. I am putting my clinician hat on, it is

not an AstraZeneca hat right now, and that

clinician's hat is that there is effect there.

You can argue the p value of 0.04 versus

0.07, and there are people here that can do that

much better than I, but from the clinician

standpoint, you have to make that choice. But you

must remember that this isn't--you also have a

patient, you have a man with a squamous cancer

sitting in your office that is smoking today, and

his likelihood of benefit by the literature is

extraordinaril