September 26, 2024 Meeting of the Oncologic Drugs Advisory Committee (ODAC)

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e right good morning and welcome I would first like to remind everybody to please mute your line or microphone when you're not

speaking also a reminder to everyone that please silence your cell phones smartphones and any other devices if you have not already done so slide two please for media and press the FDA press

contact is Lauren J McCarthy her email is currently displayed slide three please my name is Dr Christopher Lou and I'll be chairing this meeting I will now call the morning session of the

September 26 2024 oncologic drugs advisory committee meeting to order we'll start by going around the table and introducing ourselves by stating our names and affiliations we will start with the FDA to my left and go around

the table uh Richard Pastor director of the oncology center of excellence FDA Steph Lem director of the division of oncology

3 Sandra kasak team leader division of oncology 3 Robert Kumar clinical reviewer division of oncology 3 iming John statistical reviewer division

of biometric 5 Dr Van LOM I'm Catherine van Lon gastrointestinal oncologist and professor of

medicine Bill gradisher professor of medicine rest on colleges Northwestern Dan Brad professor and chair of radiology at uh sidon Case

Western Reserve University Robbie Madden medical oncologist National Cancer Institute Chris L GI medical oncology University of Colorado Joyce ver pong designated

Federal Officer FDA Neil vson R oncologist at Columbia University Lori Dodd chief of the clinical trials research and statistics Branch at the National Institute of

allergy and infectious diseases uh James Randolph Hillard patient representative a survivor of

metastatic stomach cancer due to trusa M Randy Hawkins Internal Medicine pulmonary medicine Charles University um

consumer representative Michael Gibson uh aerodigestive and upper G GI medical oncologist at the vanderbild Ingram

Cancer Center Heidi mcken Community oncologist at AIC Cancer Institute sou Fall South Dakota Jeff meart GI medical oncologist

Dana Farber Cancer Institute Hannah anof GI medical oncologist University of North Carolina thank you for topics such as those being discussed at this meeting

there are often a variety of opinions some of which are quite strongly held our goal is that this meeting will be a fair and open forum for discussion of these issues and that individuals can express their views without interruption

thus as a gentle reminder individuals will be allowed to speak into the record only if recognized by the chair person we are looking forward to a productive meeting in the spirit of the federal advisory committee Act and the

government and the Sunshine act we ask that the advisory committee members take care that their conversations about the topic at hand take place in the open for of the meeting we are aware that members of the media are anxious to speak with

the FDA about these proceedings however FDA will refrain from discussing the details of this meeting with the media until its conclusion also the committee is reminded to please refrain from discussing the meeting topic during

breaks or lunch thank you Dr frong will read the conflict of interest statement for the meeting thank you the Food and Drug Administration is convening today's meeting of the oncologic drugs advisory

committee under the authority of the federal advisory committee Act of 1972 all members and temporary voting members are special government employees sges are regular federal employees from

other agencies and are subject to Federal conflict of interest laws and regulations the following information on the status of this committee's compliance with Federal ethics and

conflict of interest laws covered by by not those by not limited to those found at 18 USC section 208 is being provided to participants in today's meeting and

to the public FDA has determined that members and temporary voting members of this committee are in compliance with Federal ethics and conflict of interest law laws under 18 USC section 208

Congress has authorized FDA to Grant waivers to special government employees and regular federal employees have potential Financial conflicts when it's determined that the agency's need for special government employee services

outweighs their potential Financial conflict of interest or when the interest of a regular federal employee is not so substantial as to be deemed likely to affect the Integrity of the services which the government may expect

from the employee related to the discussion of today's meeting members and temporary voting members of this committee have been screened for potential Financial conflicts of interest of their own as

well as those imputed to them including those of their spouses or minor children and for purposes of 18 USC section 208 their employers these interests may

include Investments Consulting expert witness testimony contracts grants cras teaching speaking writing patent and royalties and primary

employment today's agenda involves the discussion of the use of immune checkpoint Inhibitors in patients with unresectable or metastatic gastric and gastro sophal Junction at a carcinoma

the current labeling for approve the current labeling for Approved checkpoint Inhibitors in this indication reflects broad approvals in the intens Tre population agnostic of program death

cell Lian 1 pdl1 expression cumulative data have shown that pdl1 expression appears to be a predictive biomarker or treatment efficacy in this patient population

however clinical trials have used different approaches to assess pdl1 expression and different thresholds to Define pdl1 positivity FDA would like the committee's opinion on the following

adequacy of pdl1 expression as a predictive biomarker for patient selection in this patient population differing differing risk benefit Assessments in different subpopulations

defined by pdl1 expression and adequacy of C data to restrict their approvals of immune checkpoint Inhibitors based on pdl1 expression the committee will discuss

the existing supplemental biologic applications SAS which were approved for patients with previously untreated her to negative unresectable or metastatic

gastric or gastro stoal ATAC carcinoma will be Spa 12554 s-91 foro NOAB

injection submitted by Bristol Myers squ anda1 12554 s143 for Kuda PAB injections submitted

by Merc sharp and Dome LLC a subsidiary of Merc and Co Company Incorporated the committee will also

discuss bla 761 417 for tsis andab injections submitted by beine USA Incorporated for the same proposed indication this is a particular matters meeting during which specific matter

matters related to Bristol Meers sbsa MC's and beijing's NDA will be discussed based on the agenda for today's meeting and all Financial interests reported by the committee

members and temporary voting members no conflict of interest waivers have been issued in connection with this meeting to ensure transparency we encourage all standing me standing committee members and temporary voting

members to disclose any public statements that they have made concerning the product at issue we would like to remind me members and temporary voting members that if discussions involve any other products or firms is

not already on the agenda for which an FDA participant has a personal or imputed financial interest the participants need to exclude themselves from such involvement and the exclusion

will be noted for the record FDA encourages all other participants to advise Committee of any Financial relationships they may have had with the firm at issue thank

you thank you Dr frang we will now proceed with FDA introductory remarks starting with Dr Steven lemy wait sorry the clicker clicker is

different than the practice one oh got it okay good morning my name is stevenh lemy I'm a medical oncologist and director of the division of oncology 3

I'm here to today to set the stage for what will be important discussion regarding the optimization of treatment using pd1 inhibitors for the treatment of patients with gastric or gastro sophal Junction at no

carcinoma I would like to acknowledge the herculan efforts by the FDA review teams involved in both of today's meetings we wanted to look at the data with fresh eyes um when we embarked on

the need for these advisory committees with the intent of making the most scientifically and appropriate decision- making decisions for patients we're holding this meeting today in an attempt to bring order to confusing

situation p21 expression by IHC and gastric cancer is not a perfect biomarker however we would like to optimize the risk benefit for patients and Foster consist consistency in the

treatment of gastric cancer as well as the developmental landscape of new drugs studied for patients with gastric cancer pd1 appears to have utility in identifying which patients are more

likely to benefit however because different Studies have used different tests and different cut offs it can be difficult to assign the clinical effect to different pdl1 levels particularly as P1

expression increases above one nevertheless patients who were P1 intermediate or between 1 and 10 using CPS or tap appear to benefit to a lesser extent and there's uncertainty regarding

these treatment effects at the conclusion of this meeting we will ask the committee to consider whether class labeling at a PDL pdl1 level of less than one would be

appropriate we acknowledge however that arguments can be made for the selection using selection using different cutoffs we are open hearing the committee's opinions on this matter

FDA has approved two pd1 Inhibitors pisab andab in combination with chemotherapy for the first line treatment of patients with gastric cancer an application has also been

submitted for a third drug to cmab for a similar indication although FDA has granted approvals to pd1 Inhibitors in patients with previously treated gastric cancer we will focus the discussion today on

the first line setting which is most relevant to current practice all three drugs demonstrated improvements in overall survival in the first line setting both in the intent to treat patient populations highlighted by

the red box as well as in pre-specified subgroups of patients based on pdl1 expression highlighted by the purple boxes although FDA approvals in the intent to um although FDA granted

approvals in the intent to treat populations for nalab and pmab data in the pd1 low groups shown in the more heavily shaded columns were included in

product labeling to to facilitate decision- making the data appeared to show a smaller treatment effect when compared to patients with higher pdl1 expression uh which is shown in the

lighter shaded columns although not approved the data for tisab are also provided although FDA granted the gastric cancer approvals in the it patient populations professional Society

guidelines have recommended using pdl1 to select patients for treatment with pism ab and N volum AB likewise International regul Regulators have also taken such an approach of note

the ASCO and nccn category 1 recommendations were based on the tests and statistical designs used used in each of the individual studies nevertheless the guidelines do not

specifically describe or require the use of the individual test kits that were used in the clinical trials to illustrate how we got here from the agency's perspective I will highlight the challenge of the sub

subgroup analyses with the original approval of Nali map in 2021 although clearly the largest treatment effect was in patients with

CPS uh pd15 or greater which is circled in red the CPS low data were less clearcut with a questional questionable intermediate effect in patients with CPS less than one as compared to the CPS

less than five Group which is highlighted by the purple box additionally the CPS low groups were not powered to demonstrate a treatment effect leading to more uncertainty in these groups of patients

about 16 years ago the FDA held an advisory committee meeting to discuss the use of subgroup analyses to support decision-making at the time accumulating

data across at least seven trials um in patients with uh uh Chas immune colal cancer um appear to show no benefit for egfr Inhibitors panit tube maab or

stab during that advisory committee meeting when considering retrospective subgroup analyses members found that application of results across multiple trials strengthen inferences furthermore sample

ascertainment was deemed important to ensure analyses represented the populations enrolled in the trials biological posibility was another Factor when considering these subgroup analyses that designed considerations

could include stratification and pre-specification in each of the three Trials of anti- pd1 Inhibitors under discussion today there was certain there's

pre-specification of certain pdl1 High subgroups but not pre specification of the converse C pd1 low groups although one could argue that the subgroup effects in each of clinical

studies were underpowered now we have the results of at least three trials with generally consistent effects results in pd1 subgroups from the three applications in the first line her two negative metastatic setting

consistently show that the largest treatment effect appears to be in patients with CPS or tap pd1 greater than 10 these results will be be shown in subsequent presentations conversely

there's less convincing evidence of a treatment effect in patients with pdl1 CPS or tap less than one which is highlighted by the red box modest or more inconsistent effects have been

observed in patients with pdl1 intermediate disease in addition to the three trials to be discussed today to be discussed today an external trial level meta analysis of the literature has been

published by Harry unit all of uh that included 10 gastri or esophageal adnoc carcinoma studies including studies in both the first and second line settings negative trials and trials conducted

Solon Asia in this analysis CPS high was based on the different levels pre-specified in each trial with trial specific pd1 testing and pd1 appeared to designate patients as more likely to

benefit of note I'm also highlighting MSI high on this slide like our own analyses patients with MSI High tumors appear to have a very high likely a benefit following treatment with

checkpoint inhibitors for the purposes of this odac we we will be limiting the discussion to patients with microsatellite stable disease as we would not propose to modify the indication for patients with MSI High

tumors irrespective of pdl1 status when we talk about lack of benefit it's important not to forget about safety if a drug is not effective patients may be exposed to life altering toxicity the table on the left is a

summary of the inance of Select immune related Adverse Events or imrs across four clinical trials that assess pisab or nalab in general grade three or greater

imrs occur at a rate of 3 to 11% depending on the clinical trial although many imrs are treatable imrs can become chronic par particularly for endocrine

lung neurologic cardiac arthritis um or arthritis these Adverse Events or even steroids used to treat imrs of less severity um Can can greatly compromise the patient quality of life which are

important to patients with endstage gastro cancer I would like to transition to how we move forward if the right thing to do is to limit the indication at a specific

pd1 cut off one approach to take would be to limit the indications based on P1 positive cut offs used in each clinical trial although this may be a reasonable approach to statistically is not

necessarily biologically based in other words what would be the optimal cut off to maximize benefit and reduce risk such an approach may also unnecessarily exclude patients from treatment if we

selected a cut point that was too high we now have data from multiple clinical trials that can be considered to assess whether a classwide approach would be more appropriate in practice oncologists do

not necessarily use the specific CPS or tap diagnostic test that were developed for each individual monoclonal antibody the populations in the clinic May differ as compared to clinical

trials perhaps more importantly companies wanting to study dual checkpoint Inhibitors other add-on drugs or drugs intended to treat Target other biomarkers may have to link their clinical trial to P1 levels related to a

checkpoint inhibitor rather than a biological principle which doesn't seem to be the most rational approach one more important consideration however is that all three studies use different diagnostic tests

to assess pdl1 status as a hyp purely hypothetical example if a cut off of either five or 10 were selected any of test theoretically could be used however the number of eligible patients would

greatly be affected depending on the test importantly irrespective of the test use Patti patients were pd1 negative or less than one appear not to or were less likely to benefit

designating a patient as pd1 negative may be less variable from test to test as compared with the challenges of designating the specific score for example p14 8 or 10 of not only a

minority of patients are P1 negative regardless of the asset used if a cut off of one were recommended depending on the test 80 to 90% of patients with gastric cancer would still be eligible to receive a checkpoint

inhibitor I'll summarize by providing a snapshot of the data in the pd1 less than one in 10 subgroups so Dr Kumar will provide a much more complete review of the data including data above and

below different pdl1 cutoff levels Dr Kumar will also provide FDA exploratory pooled analyses limited to patients with microsatellite stable disease to provide additional cont context regarding each clinical

trial it is not shown here but the data in patients with CPS or tap pd1 greater than 10 show clear benefit in patients with P1 of less than 10 or between 1 and 10 there are more

inconsistent effects however there may be a hint of a plateau in two of the cap M curves as shown by the purple arrows and the upper bound of the 95% confidence intervals in the purple ovals are below one or close to

one nevertheless this should not be considered as definitive evidence of either benefit or lack of benefit when one views either the median overall survival haard ratio or Capital my curves at the pl1 less than one cut

off and the red boxes there appears to be less convincing evidence for benefit and not even a hint of a possible plateau in the Capal M curves when compared to the control

arms Dr Kumar's presentation will show a stark contrast of these capm curves compared to those in patients with pbl1 tumors greater than 10 for each drug it is important to remember that although a

minority of patients will develop severe or life-threatening toxicity for those who do the quality of their lives can be greatly altered I would Al like to point out the available results for keyot 811 to

further support the biological plausibility with respect to pdl1 and gastri cancer Kino 811 is a firstline study in patients with her two positive gastric cancer where pisab was administered in combination with trab

and chemotherapy please note that I'm only providing this information for contexts for the committee mad received accelerated approval for this indication based on the pre-specified inter

analysis inter analysis of response rate that demonstrated statistical significance in a subsequent prespecified interim analysis however there appear appear to be a potential for detriment and survival in patients

with P1 CPS tumors less than one and current labeling limits pism AB to patients who are ped1 greater than one based on the data that I've shown so far there does appear to be a general

replication of results across clinical trials which is consistent with a retrospective approach used for egfr Inhibitors based on Ras mutations in color rectal

cancer this reputation was seen both in the first line gastric cancer trials in combination with K therapy as well as additional trials and the Harry Yun metaanalysis sample ascertainment for

CPS or tap was high in each of the three trials to be discussed today with the results available from the vast majority of patients with respect to biological posibility although pd1 has had variable

utility as a biomarker in different tumor types it does appear useful in select disease settings finally all studies all studies designated specific P1 High populations at different

thresholds however none of none of the studies were specifically designed to test for the pdl1 negative groups again all three studies use different pdl1 testing methodology I would like to summarize my

interpretation of the available data clearly there's an important benefit to checkpoint Inhibitors in patients with gastric cancer who have uh their tumors P1 score of 10 or greater and patients

for micrite instability high again we're excluding MSI high from our assessment of risk and benefit in patients with P1 intermediate gastric cancer for example 1 to 10 there may be a modest benefit however it's

difficult to convinced demonstrate or exclude such an effect an additional consideration regarding uncertainty in this group may involve the accuracy of classification of pdl1 for example

differentiating 9 from 10 in patients with P1 negative disease although there may be some uncertainty based on the smaller number of patients irrespective of the Asus there there does not appear evidence of benefit and patients may be

at risk for harm additionally uncertainties regarding testing may be mitigated as pd1 P1 standing will either be present and positive and negative absent or

negative again selection of a P1 cut off of one would result in 80 to 90% of patients being eligible for checkpoint Inhibitors and would allow consistent approach to treatment in the clinic and

in clinical trials going forward following all the presentations we will ask the committee to discuss the use of pdl1 as a predicted biomarker for the selection of patients with gastric

and gastrosoph adoc carcinoma we welcome the viewpoints of the committee on this challenging topic following the discussions we will ask the committee to vote on whether the risk benefit assessment is favorable in

patients with gastric cancer who have P1 expression less than one please note that of the FDA review staff has had extensive internal discussions following the review of the totality of data prior

to finalizing this question FDA reviewers consider not just the data but the landscape of testing and the landscape of treatment of gastric cancer although we are specifically asking about the cut off of one we invite you to express your opinion if you believe a

different cut off would be more appropriate thank you thank you Dr Lem both the Food and Drug Administration and the public believe in a transparent process for information gathering and decision-

making to ensure such transparency at the advisory committee meeting FDA believes that is important to understand the context of an individual's presentation for this reason FDA encourages all

participants including industry's non-employee presenters to advise the committee of any Financial relationships that they may have with industry such as Consulting fees travel expenses

honoraria and interests in a sponsor including Equity interests and those based upon the outcome of the meeting likewise FDA encourages you at the beginning of your presentation to advise the committee if you do not have such

Financial relationships if you choose not to address this issue of financial relationships at the beginning of your presentation it will not preclude you from speaking we will now proceed with our first presentation from Bristol

Meers squib good morning my name is Ian Waxman and I'm part of the late development oncology organization at brist Meer squib I'd first like to thank the advisory committee members and the FDA

staff for this opportunity to discuss the data for updo in combination with chemotherapy and firstline gastric cancer these data come from the

Checkmate 649 study and resulted in FDA approval for this indication in April of 2021 this marked the first approval of a new treatment for firstline her 2

negative gastric cancer since chemotherapy became the standard of care by way of background ABDO was first approved in the US in 2014 for the

treatment of melanoma and is now approved in 11 cancer types as shown here here is the indication for ABDO in combination with chemotherapy for

firstline patients with gastric cancer GE Junction cancer or esophageal adenocarcinoma and it's important to highlight two things first the approval

was granted regardless of pdl1 status and second since the initial approval our interpretation of the study results has not changed with longer

followup although the indication is not limited to a pdl1 positive population clinical data by pdl1 expression level are included in section 14 of the uspi

these data are included to ensure that treating Physicians have sufficient information regarding the impact of pdl1 positivity when discussing treatment

options with their patients since approval results from additional gastric cancer Studies have been reported with different sponsors incorporating different methods for

measurement of pdl1 as well as different cut offs to determine positivity nccn has managed the situation by giving a category one recommendation for patients with higher pdl1

expression while also giving a category 2B recommendation for patients with lower pdl1 expression ASCO and esmo guidelines have taken a similar approach

for noolab nccn guidelines are consistent with the current FDA label with no restriction on treatment but with information provided to highlight the importance of pdl1 expression level

in determining likelihood of clinical benefit given the current uspi and nccn recommendation it's important to understand what Physicians are doing

about testing in the real world when we look at testing patterns in the US based on flat iron data we see that phys Ians have indeed received the message around the importance and impact of pdl1

expression in this disease what we see on the left is that approximately 60% of all patients treated with any regimen in the first line setting are already being

tested for pdl1 expression even without a requirement to do so and when we look at the middle pie chart over 70% of those treated with noolab are being

tested regarding treatment patterns also based on flat iron data we see that the pdl1 test result is influencing treatment decisions in the real world

today among patients known to be pdl1 positive on the left about 50% receive an IO regimen in contrast among the much smaller proportion of patients known to

be pdl1 negative in the middle pie chart fewer than onethird receive an IO regimen another way to think about this is that among all treated patients less

than 5% are treated with IO and known to be pdl1 negative on the far right hand side we see that many patients are not tested or

have an unknown test result and about onethird of these patients are treated with an iio regimen this high percentage of patients without a test result is not surprising when we consider testing

rates for her too which is another established biomarker in gastric cancer approximately onethird of firstline patients are still not tested for her two in clinical practice

despite the longstanding availability of her two directed therapy given the high prevalence of pedial one positivity most patients without a test result would be

considered positive if tested we believe these patients should remain eligible for an IO containing regimen in the absence of a comorbidity precluding its

use with this information in hand we're here to discuss whether any label changes for ABDO and gastric cancer are needed our goal is is to ensure that each firstline gastri cancer patient has

every appropriate therapy available to them along with clear guidance to inform choice of treatment a review of subgroup analyses

by pdl1 expression level from Checkmate 649 and a summary of challenges associated with interpretation of a pdl1 test result are important topics to

discuss when considering this goal once we have covered these additional areas I'll turn to a summary of potential op for labeling also

briefly described here one option is to modify the indication to only include patients with any level of pdl1 positivity this would limit treatment to patients more likely to benefit based on

the clinical trial data but could leave some patients without a potentially important treatment Choice the second option is to leave the indication as is so that Physicians can continue to make

treatment decisions informed by the data as currently described in the uspi and consistent with CCN guidelines additional considerations for each of these approaches are shown here and will

be discussed in more detail in the next parts of this presentation here's the agenda for the remainder of our time first Dr Dana Walker from the drug development organization at BMS will review the

relevant efficacy and safety data from Checkmate 649 then Dr Robert Anders an expert pathologist from the Johns Hopkins University will discuss the realities of pdl1 testing and clinical practice and

then finally I'll return to review the proposed options for labeling thank you and I'll now turn it over to Dr Walker thank you my name is Dana Walker and I'm the global program lead for

opdivo and yvo for GI and gu cancers I'll be presenting efficacy and safety data that will highlight the benefit risk profile of novad plus chemotherapy

in the Checkmate 649 study across pdl1 subgroups Checkmate 649 is a randomized phase three study that included patients with previously untreated unresectable

Advanced or metastatic gastric or esophageal adoc carcinoma regardless of pdl1 expression randomization was stratified according to tumor cell pdl1 expression

this study had dual primary endpoints of overall survival and progression-free survival in CPS 5 or higher 1,581 patients were concurrently

randomized to the Nomad plus chemotherapy versus chemotherapy arms of which 60% had pdl1 CPS of five or

higher with a minimum of 12.1 months of followup check meet 649 demonstrated both a statistically significant and clinically meaningful overall survival benefit in the primary and secondary

analysis population the primary endpoint was in patients whose tumors Express pdl1 CPS of five or higher which demonstrated a

hazard rate ratio of 0.71 and a 3.3 month Improvement in median overall survival versus chemotherapy of note there was an early and sustained separation of the overall

survival curves a similar overall survival benefit was observed with Neo Plus chemo in the CPS one or higher and the all randomized

population shown here is the overall survival data in pdl1 subgroups that was available at the time of the initial approval the data in the purple boxes highlights the pre-specified primary and

secondary analysis population the other CPS subgroup analyses were exploratory there was an increased overall survival benefit observed with

Neo Plus chemo at higher pdl1 cut off patients with CPS less than one did not derive an overall survival benefit however in the subgroup with CPS greater than or equal to one the hazard

ratio was 0.76 with 95% confidence interval of 0.67 to 0.87 now let's take a look at four-year

follow-up data which are presented in our briefing document shown here are all of the subgroups requested by FDA keeping in mind that these are exploratory subgroup analyses and should

be interpreted with caution these data are consistent if not continuing to improve across subgroups relative to the initial clinical trial data in particular I I would like to

point out that the subgroup with CPS greater than or equal to 1 to less than 10 has a hazard ratio of 0.88 improved from a hazard ratio of

0.95 at the initial database lock of note the FDA meta analysis does not include these updated data the safety profile of noola mad

plus chemotherapy was consistent with the known safety profile of the individual drug components with no new safety signals identified as expected the addition of noolab to standard

chemotherapy was associated with added toxicity grade three4 treatment related Adverse Events and those leading to discontinuation of any treatment component were numerically higher in

patients receiving NAD plus chemotherapy of note in the Neo Plus chemo arms the majority of immun mediated events were low were low grade manageable with

established treatment algorithms and reversible importantly the safety profile of no Neo Plus chemo did not differ based on pdl1 expression and was

consistent across all pdl1 subgroups evaluated in summary based on the data from Checkmate 649 Neo Plus chemo demonstrated both a statistically

significant and clinically meaningful overall survival benefit in the CPS greater than or equal to five CPS greater than or equal to one and all randomized

population exploratory analyses showed a higher likelihood of overall survival benefit in all pdl1 positive subgroups and the long-term follow-up data are consistent with the data available at

the time of approval and provide a clearer picture of the overall survival benefit and the CPS greater than or equal to one population the safety profile was

consistent with the known safety profile of the individual drug components and similar regardless of pdl1 status overall Neo Plus chemo demonstrated a positive benefit risk

profile in all pdl1 positive subgroups thank you I will now turn it over to Dr Anders thank you Dr Walker my name is Robert Anders I'm a professor of pathology at John's Hopkins University

and a paid consultant for BMS today I'll be sharing you with you my 17 years of experience with the technical aspects of pdl1

testing as you heard about 60% of patients with gastric cancer are tested for pdl1 expression and at my institution most patients with gastric

cancer are tested it's worth noting that most gastric cancer patients with a test are

pdl1 positive as defined by a CPS of greater than or equal to one the graph on the left shows data from Checkmate

649 where 82% of patients were pdl1 positive most variable multiple variables can affect the results of pdl1

scoring such as the type of tumor tissue spatial heterogen and temporal changes in pdl1 expression shown here is pdl1

staining of a full thickness gastri cancer reection for my practice pdl1 positive areas are stained Brown and are circled with solid black lines while

pdl1 negative areas are circled with dashed lines and you can clearly see there's heterogen of pdl1 expression on top of that I super

composed an h& stained mucosal biopsy to give you an idea of the depth of a typical endoscopic biopsy these are the most common types of tissue samples I

see they're more amenable to pdl1 scoring because they have fewer cells to count but they may under represent the

tumor and may lead to false negatives a surgical reection better represents the entire tumor but presents a challenge because the heart of CPS is

counting cells both immune and tumor cells and frankly reections have too many cells to count we also know that pdl1 expression

varies as a function of time and both biopsies and resections Are One Moment In Time let's move to analytical

consideration there are three approved antibodies that recognize pdl1 what you see here from a recent publication are serial sections of the same tumor stain

with the three antibodies the staining is similar but not identical as a result Pathologists May count fewer or more positive cells

depending on which antibod is used this is relevant at the patient level because differences of a few cells changes the CPS this could be a change across a critical

threshold so minute differences could mean the difference between a patient receiving firstline IO therapy or not it's worth noting that registrational

trials each use different antibodies which has implications for harmonization as a result of these variables I just described CPS is a highly subjective test with high

interobserver variability I recently published a paper with Marie robar from Yale on the level of agreement between 12 expert

Pathologists from around the globe shown here are the CPS scores for 100 gastric cancer biopsies using 22 C3

the score from each of the Pathologists is represented by a different color dashed line we report an interclass correlation

coefficient a measure of statistical Observer agreement of about 0.5 which is fair to poor agreement for

reference the ICC for her two testing and breast cancer ranges from 0.8 to 1 here we've taken a closer look at the

individual scores from 30 of those gastric cancer biopsies scores are indicated by a DOT for each pathologist and we've marked

the CPS cut offs at 1 15 and 10 as horizontal dashed lines it's easy to see not only High Vari in scores but tremendous variation

across the cut points often times with just as many dots above the line as below the implications are significant because even these worldclass experts would disagree on whether a patient

should receive therapy simultaneously Dr Rim from Yale did an identical study with us-based Pathologists and came to identical

conclusion that CPS is a highly subjective test our conclusion from these studies is that pdl1 CPS may be useful biomarker at the

population level but an imperfect biomarker at the individual patient level in conclusion in my experience

pdl1 expression by CPS is Complicated by spatial heterogeneity of expression endoscopic biopsies that have the potential for false negatives we also

have different antibodies with different assays and un acceptably High interobserver uh variability I believe Pathologists can reliably determine if

there is pdl1 expression or not however it's much more difficult to precisely quantify pdl1 expression which

results in the high variability we see in CPS scores thank you and I'll turn it back now to Dr Waxman thank you Dr Anders as I summarize the data just presented I'd

first like to acknowledge that this is an important issue without one clear-cut solution the FDA has asked you to consider whether the data support the use of pdl1 expression as a predictive

biomarker the clinical trial data show that there is an overall survival benefit in patients who Express pdl1 including at the level of CPS of greater than or equal to one further enrichment

for OS Improvement at CS higher than cps1 is also likely but we must consider some practical challenges when choosing to restrict the indication to a specific

higher pdl1 cut off as you heard from Dr Anders pdl1 is a dynamic biomarker that exhibits significant temporal and spatial heterogenity and pdl1 scoring is

challenging for Pathologists and subject to a high degree of variability the quality and availability of tissue can also be a barrier to testing we do support testing for pdl1

whenever possible since this result is important in informing benefit risk and we're assur reassured that such testing is already occurring in the majority of patients today but we recognize there's

a downside to requiring a test result given the numerous testing challenges just described therefore we still support a broad indication however if

the label were to be restricted a cut off based on cps1 is the most reasonable Choice based on the totality of clinical data and testing

considerations to summarize the first option is to ify the indication and require cps1 a higher cut off would not be optimal given the challenges with precise quantification of CPS in

individual patients however this option would leave some patients without the potential to benef with potential to benefit untreated the second option is to keep

the existing indication given that details regarding the impact of pdl1 expression are already captured in the label this approach accounts for the uncertainties associated with pdl1

testing and leaves informed decision-making in the hands of the treating physician where it lies today this approach also provides greater opportunity for patients without

a test result to benefit from immunotherapy although both proposals are reasonable we consider the option that provides flexibility for patients regardless of their pdl1 test result to

be most appropriate and thank you once again for your time and attention thank you so much we will take a very brief five minute break to allow for the next presentation to set up

panel members please remember that there should be no discussion of the meeting topic during the break amongst yourselves or with any member of the audience we will resume at 8:50 a.m. recording

a.m. recording

stopped e e e e e e e

e e hey welcome back uh we will now proceed with our second presentation from mer sharp and Dome Incorporated good

morning I am Kathy Panza vice president of clinical research in late stage oncology I'm a medical oncologist and prior to joining MC I was an attending

physician at Memorial phone kin Cancer Center thank you for the opportunity to share evidence supporting the positive benefit risk profile of Kuda in patients

with herto negative gastric and gastro sophal Junction cancer which we will refer to as gastric cancer after my introductory comments Dr

Pua bagya will share data from keyote 859 and Dr yena jigan will share her clinical perspective Kuda helps fulfill a critical need for the the treatment of

patients with metastatic gastric cancer who have a poor prognosis with only 7% surviving 5 years before immunotherapy the only treatment for firstline

metastatic disease was chemotherapy with few biomarkers or targetable molecular aberration we Face a dur of therapeutic options rigorous

study design and conduct gives confidence in the positive results of keynote 859 which met success criteria for all primary and key secondary end

points in the intention to treat population the Kuda label includes information about pdl1 subgroups empowering Physicians to work with

patients to make the best choice for therapy the indication for Kuda and her two negative gastric cancer should be retained based on the efficacy and

safety data for pism app the mechanism of action of pism app is well known increased expression of pdl1 enriches for response with pism when given as

monotherapy in many tumor types pdl1 expression is tumor type specific and interpretation is dependent on the assay and scoring method

used we know that chemotherapy has pleotropic immunomodulatory effects both promoting and impairing the anti-tumor response adding an anti- pd1 inhibitor

to chemotherapy can enhance the positive and reduce the negative immune effects the complimentary effects of the combination of pemis map and chemotherapy can benefit patients with

tumors across a broad range of pdl1 expression as observed in numerous indications I will now describe the comprehensive training and validation

methodology used for pdl1 testing in keynote 859 clinical samples are processed using

the pdl1 IHC 22c 3 Farm DX assay and interpreted using combined positive score known as CPS which captures pdl1 expression on tumor cells lymphocytes

and macrofagos this is clinically important in gastric cancer as it has a significant immune infiltration Merc clinical studies are

used as training sets to determine Cup points for each tumor type once these cut points are identified we collaborate with our diagnostic partner and testing

laboratory to validate these Cup points Pathologists were trained to use the pre-specified Cup points during patient screening for keynote 859 the

validation test set for assessing pdl1 expression in the study all pdl1 expression was performed in a central

laboratory this rigor allows for informative pre-specified pdl1 subgroup analyses importantly while higher pdl1

Cup points can enrich for pemis map monotherapy efficacy in gastric cancer we cannot predict who will benefit especially when chemotherapy is added to

pema's map robust pdl1 evaluation in keynote 859 supports the indication under discussion we acknowledge that pdl1

testing outside of clinical trials is variable the biology of the disease the treatments mechanism of action and the possible impact of combination establish

the foundation of merc's phase three trials even when we anticipate benefit across a broad range of pdl1 expression we design our studies with the potential

for biomarker enrichment to increase the likelihood of successful outcomes insights for Merk clinical trials and emerging Knowledge from

external sources further inform the design of reg ational studies of course labeling reflects the results as well as the statistical rigor and methodologies

of phase three studies when considering labeling changes the same statistical principles apply post Hawk subgroup analyses at cut

points that are neither carefully assessed nor pre-specified with type one error control may lead to spous findings

additionally a value Val ating numerous subgroups May demonstrate randomly high or low treatment effect

estimates the scca's pooled analysis also has inherent limitations it assumes that the different immune checkpoint Inhibitors have an identical treatment

effect it ignores differences between the therapies trials pdl1 assays and defined Cup points most important

patient selection reflects three different tests whose interchangeability has not been established using the same numeric value across different tests

does not mean that the values are equivalent combining potentially different populations treated by different drugs estimates a quantity

that does not represent any actual drug in combination with any test such an analysis does not meet FDA standards for labeling

postt talk subgroup and pulled analyses should not supersede the findings of the phase three randomized trial with a diagnostic specifically developed for

use with pism app since the approval of keynote 859 there have not been any new efficacy or safety data that changed the benefit

risk profile for pemis map in gastric cancer the pdl1 IHC 22 22 C3 Farm DX assay is specifically studied for pema's

map in the approved indication there are key differences in considering a restriction of this indication by pdl1 Cup point compared to those for sumab or

panitumumab and olaparib molecular alterations such as K and braam mutations strongly predict response

whereas pdl1 expression is a Continuum it can be modulated by other therapies like chemotherapy and is not always predictive of immunotherapy response for

camab panitumumab and olaparib the outcome of each study was individually evaluated as opposed to a pulled

analysis across three different studies nccn ASCO and esmo guidelines specify granularity around the strength of efficacy at various cut points

Physicians use these clinical guidelines the label and importantly the individual patients characteristics to determine appropriate treatment Dr puaa will now

share data from the phase three study in her2 negative gastric cancer that led to the approved indication thank you Dr Panza my name is Puja bhagya I am the upper GI cancer

clinical lead at MK and I will present safety and efficacy data from keynote 859 keynote 859 supported the full

approval of Kuda for the first line treatment of adults with her2 negative gastric cancer in keynote 859 patients had herto

negative metastatic or locally Advanced unresectable gastric cancer regardless of pdl1 status pdl1 CPS less than one

versus greater than or equal to one was one of the stratification factors keot 59 was designed based on

pisab monotherapy studies which demonstrated activity across all levels of pdl1 expression with potential for increased efficacy in the CPS greater

than equal to one population the statistical plan was designed to test both the it and the CPS greater than equal to one

populations 78% of the population was CPS greater than equal to 1 following initiation of keyote 859

results from another phase 3 study in first line gastric cancer indicated a potential for further enrichment at CPS greater than equal to 10 we adjusted the

statistical plan to formally test endpoints in this population as well 35% of the population was CPS greater than

equal to 10 now let's review the data support in full approval in the intention to treat population keynote 859 met statistical

success criteria for all its end points overall survival progression-free survival and objective response rate the

overall survival curve favors fism app with a 22% reduction in the risk of death progression free survival curve also favors fism app reducing the risk

of progression or death by [Music] 24% at 2 years 28% of patients in the

Pisa plus chemotherapy arm remained alive versus 19% in the chemotherapy arm notice the tail of the curve which is

characteristic of femoralis app the safety profile of the investigational arm is consistent with the established safety profiles of femoralis ab and

chemotherapy the addition of fism ab adds immune mediated AES and infusion reactions which were mostly low grade and

manageable it is known that some immune mediated AES such as endocrinopathies will require long-term hormone replacement these data highlight the

favorable benefit risk profile of pisma plus chemotherapy for all patients to address the fda's questions

we will now look at different pdl1 cut points in all pdl1 subgroups patients experience a benefit with Hazard ratios

below one for both OS and PFS a higher magnitude of benefit is seen with increasing pdl1 expression the CPS greater than equal to

one subgroup was formally tested with Alpha control the point estimate of Hazard ratio for OS and PFS is 0 .73 and

.73 and 0.72 respectively with a statistically significant and clinically meaningful benefit in this group the CPS greater than equal to 10

subgroup also shows a clinically meaningful benefit the CPS less than one subgroup was not pre-specified with formal

statistical testing the magnitude of benefit is less in the CPS less than one subgroup with the point estimate of the OS Hazard r IO being directionally

consistent with the it importantly the median PFS was approximately 1.5 months longer with increased or and longer

duration of response suggesting that some patients do derive benefit in this subgroup in patients with CPS between 1

and less than 10 we see benefit with an OS Hazard ratio of 0.83 with narrow confidence interval overlapped with the it and the upper

bound is less than one this indicates that the benefit is not driven by CPS greater than 10 at a threeyear followup the benefit

of pism remained consistent with the primary analyses underscoring a tenet of immunotherapy the safety profile is in

general similar across CPS cut points and there is no biological rationale to suggest that the safety profile of fism ab would change based on pdl1

expression in conclusion there is a high unmet need in firstline metastatic gastric cancer pisab added to chemotherapy provided a statistically

significant and clinically meaningful Improvement in OS PFS and Orr in all patients and the magnitude of benefit increases with higher pdl1

expression some people patients with lower CPS scores also experienced benefit highlighting that CPS expression alone cannot predict which patients will

benefit from the combination of fism app and chemotherapy health related quality of life remained stable during treatment was similar between arms and consistent

across CPS subgroups the manageable safety profile reflects the known safety profiles of the components and is generally similar

across CPS subgroups moreover the label includes information on efficacy by pdl1 cut points and supports a benefit risk

discussion between Physicians and patients thank you and I will now invite Dr jigan to the podium good morning thanks Puja and it's

such an honor to address this audience I'm a medical oncologist and I'm here on behalf of clinicians treating this disease certainly in our patients and

caregivers next slide these are my disclosures uh next slide and so um I'm an expert in the field I yesterday in clinic I I saw 30 patients with this

disease and my research practice uh is focused on this most patients in United States are treated outside of tertiary Cancer Center so we need to really be

there in the clinic with our uh practitioners to understand what they're facing most patients present with stage four disease it's an orphan illness in United States so an oncologist sees

maybe at best five uh gastric cancer patients a year so they need to act fast and we know that there's a narrow window of improvement of their quality of life

and also the likelihood that they will these patients will respond to therapy will predict uh the likelihood they will get on second line third line and so forth so we're in it now for the long

game so the response is important these patients tend to uh come in from centers that they're not getting their inial diagnosis in so they may or may not bring unstained slides in and the

therapeutic options are um really driven by uh the clinician as you saw by ASCO and CCN guidelines there is some restriction but often we start treatment

before we these biomarker testing uh results come back immun CH blockade has its downsides and some of these uh Adverse Events can be long-term but we need to remember that most of the side

effects actually come from the fibox alasm based therapy and the clinicians really know how to use immunotherapy because as I said most clinicians treat also lung cancer breast cancer renal

cell cancer so they know these agents well next slide so how did we advanced practice for this disease we are we do uh biomarker testing her two is seen in 20%

MSI is a rare but important subset and most of the patients do have some tumor overexpression of pdl1 so in patients uh

we see 80% Plus for CPS testing if the testing is done next slide but is the testing done what are the practical implications so repres having down stain

slides choosing the right essay and also the pathology interpretation you heard this excellent earlier in the talk we do have a difficulty obtaining slides so the subset of patients I'm particularly

worried about are the people who are not getting tested for various reasons because of the sample quality or availability there is a huge variability in which assay is used depending on the

center and there's a learning curve gastric epithelium is tricky some of these biopsies are quite superficial so you don't have the full tumor content and if you're used to looking at lung

cancer and Grading TPS CPS may be a challenge and there's a steep learning curve next slide so how are the real world practices doing well it turns out with all of the research that we're

doing in our gastric cancer world uh it's not translating directly into all our patients a quarter of our patients never even get pan testing or any uh biomarker testing and only about 50% get

immunotherapy suggesting there's really no overuse of these agents in the practice um and there is a variability for example msk doesn't even use any of these biopsy s we use our own lab

developed tests and there is a huge learning curve for CPS scoring so all you know tap and different CPS cut offs

are not interchangeable next slide so the uh Stakes are high here because uh we've changed we've bent the curve the survival curves with these immunotherapies and it's amazing to be

able to sit in the clinic and tell a patient you have about a quarter chance of living longterm with a disease even with metastatic disease but we only have that chance if we use the drugs in first

Clan setting in United States we don't have approval of immune checkpoint blockade in later lines unlike Asia where gastric cancer is very common and you still use immunotherapy in LA later

line next slide and so in conclusion we really need to push the envelope improve long-term survival and choice of firstline therapy really matters for our patients so

clinicians need to have these tools in the clinic askco nccn guidelines do provide really good guidelines as to which agents to use I'm worried about

the patients with unavailable testing or testing that never gets done because statistically uh speaking 80% of patients would benefit uh from getting immune checkpoint bade if we end up uh

restricting so I think it's important to let Physicians take care of the patients and to make decisions individually and not just on population level base um and so uh you know that would be my sort of

two cents on this but I'm happy to take any questions thanks for your attention thank you thank you Dr Janan in summary keynote 859 met success

criteria for all its primary and key secondary points in the intention to treat population the label reflects the study outcome metastatic gastric cancer

is a fatal disease where survival is measured in months holism AB combined with chemotherapy is one of the only treatment options for these patients

while higher pdl1 expression enriches for benefit data show that efficacy can occur across a range of expression

including in those with no or low expression as we heard restricting the indication to pdl1 CPS greater than or

equal to 1 will leave approximately 25% of these patients with no other option besides chemotherapy the current indication allows Physicians to make the best

possible choice for patients with gastric cancer thank you and we look forward to a productive discussion thank you so much uh we will take a brief 10-minute break to allow for the next presentation to set up

panel members please remember that there should be no discussion of the meeting topic during the break amongst yourselves or with any member of the audience we'll resume at 9:20 Eastern Time

e e e e e e e e

e e e e e e e

e e welcome back we will now proceed with our third presentation from Beijing good morning good morning my name is Mark

lenosa and I'm the chief medical officer for solid tumors at Beijing Beijing is a midsize pharmaceutical company developing Innovative medicines for patients with cancer around the globe

beije was founded 14 years ago and has grown into a fully integrated global company with offices around the world we have a broad portfolio of cancer therapies with two internally discovered

globally approved medicines including tisis maab which is the focus of today's presentation additionally we have over 2,000 employees in the United States and recently opened a state-of-the-art

manufacturing facility in Hopewell New Jersey I want to thank the FDA the chair and the members of the committee for the opportunity to share our results with tisel ismb and to provide our

interpretation for this important discussion during this morning session I will briefly provide background information about talism ab and will then spend the bulk of my time reviewing

the results from our pivotal study rationale 305 along with efficacy analyses across pdl1 expression subgroups I will then turn it over to Dr

ioha to provide background on gastric adnoc carcinoma and her perspective on the use of tsilis ab in patients with gastric or gastro sophal Junction adoc carcinoma which we will refer to as

gastro cancer or GJ we also have additional functional area experts with us today to help to address your

questions tiis m is unique anti pd1 designed for potent pd1 binding and robust cd8 positive t- Cell Activation tisis maab is an FC engineered humanized

igg4 antibody tisis binds to the extracellular domain of human pd1 with high specificity and Affinity it binds pd1 at a unique epitope that competitively blocks The Binding of both

pdl1 and pdl2 pilis M does not bind to FC gamma receptors and therefore does not induce antibody dependent cellular cytotoxicity or complement dependent

cytotoxicity these differentiating features enhance the functional activity of te- cells in invitro cell-based

aets our pivotal Global phase 3 rationale 305 study evaluating the efficacy and safety of tisab combined with standard chemotherapy versus

placebo plus chemotherapy and the first line setting in patients with locally Advanced unresectable or metastatic gastric and GJ cancers was initiated in

2018 and the bla was submitted to the FDA on December 28th 2023 and is currently under review please note that the primary endp point of overall

survival was tested hierarchically with the primary analysis of the pdl1 positive group occurring at an interm analysis in October of 2021 and a final

analysis of the it population in February of 2023 overall results from our pivotal study show that first line treatment

with tisab in combination with chemotherapy improved overall survival in patients with locally Advanced unresectable or metastatic GJ adnoc carcinoma and therefore can offer an

important treatment option for these patients the benefit risk of tisab as firstline treatment for locally Advanced unresectable or metastatic gastro cancer

is favor able and is overall consistent with that of currently approved pd1 Inhibitors we do find that pdl1 is a predictive biomarker in gastri cancer

based upon our primary endpoint the benefit risk is most reliably established pardon me in the subgroup with pdl1 expression greater than or equal to

5% however we also observe modest but consistent benefit in patients with expression between 1 and 10% isab plus ke therapy has a tolerable

and acceptable safety profile which is similar to other approved pd1 Inhibitors first I will share with you the design and key efficacy results from study

305 rationale 305 is a randomized double blind Placebo controlled phase 3 study in 997 patients with histologically confirmed gastric cancer overall the

design of the study paralleled those of the other approved products in this indication the study excluded patients with either herto positive tumors or with prior

therapy for unresectable locally Advanced or metastatic gastra cancer stratification factors include a geographic region of enrollment pdl1 expression above or below 5% the

presence of Partin metastasis and the investigator's choice of chemotherapy all patients were required to have at least one lesion one a valuable lesion according to resist

version 1.1 ecog performance status score of Z or one an adequate organ function and nutritional status patients were randomized one to

one to receive either tisis M 200 milligrams or matching Placebo administered by intravenous infusion every 3 weeks until disease progression or unacceptable toxicity both treatment arms were

administered in combination with Physician's choice of standard chemotherapy either oxal platin and cap cabine or CIS Platinum 5 of

few the primary end point was overall survival hierarchically tested first in the pdl1 greater than or equal to 5% group followed by testing in the it

population multiple secondary endpoints including progression free survival objective response rate duration of response and safety were also

evaluated because the primary endpoint was sequentially tested first and the greater than or equal to 5% group followed by the it population I will share the data that informed our

decision to use 5 % to define the pdl1 positive subgroup this cut off was derived from a post Haw analysis of an early phase single arm study of tsilis

map monotherapy administered a second lineer later treatment to patients with Advanced gastric or esophageal adenocarcinoma an initial analysis of 46

patients led to the selection of 5% in the study protocol subsequently a receiver operating characteristic analysis confirmed based upon objective

response rate confirmed the potential predictive value of 5% in this analysis of 77 patients which included the initial 46 patients a pdl1

value of 5% maximized sensitivity and specificity for O next I'd like to briefly explain the assay used for pdl1 determination in

rationale 305 pdl1 status for all analyses was assessed using the sp263 assay and scored following the tuber area positivity or tap algorithm both

the tap and the combined positive score are designed to measure the same biology the tap score was developed by ro tissue Diagnostics our companion diagnostic

partner key differences between tap and CPS are that tap includes tuber Associated immune cells and the tap is visually estimated rather than based on

Cell counting as a result tap can be performed more quickly by the pathologist while retaining interobserver ordance of 95% at the proposed cut off value of

5% importantly although we have confidence in the technical operating characteristics of this validated assay at 5% we also acknowledge the Practical challenge that prescribers face due to

various assays being utilized by different pd1 Inhibitors as well as different assays being used at different clinical sites for this reason we fully support efforts towards assay

harmonization within the class transitioning back to rational 305 study overview I will now share the patient demographics and disease characteristics Baseline demographics

were overall well balanced and representative of the target patient population the median age was 60 years and the majority of enrolled patients were male 75% of the patients were from

East Asia consistent with the global epidemiology of gastric cancer the remaining 25% of patients were enrolled in Europe and the United States our intent was to enroll a larger proportion

of patients in the United States but enrollment in the US became infeasible once the Topline results from the noolab study presented in this session became available similarly Baseline disease

characteristics were balanced and are consistent with the patient population here in the United States the median time from initial diagnosis was less than 2 months almost all patients had

metastatic disease and approximately 27% of patients had prior gastrectomy most patients had an ecog score of one approximately half of all patients had

tubers with a pdl1 score greater than or equal to 5% based on the sp263 tap score the disease characteristics of this group while not shown here were

consistent with the it analysis set now let's move on to the primary endpoint analysis the primary endpoint of overall survival was met in both the pdl1

greater than or equal to 5% population at the interim analysis and subsequently in the it population at the final analysis this is the Capa Meer curve for

the pdl1 positive population at the interim analysis the investigational and control arms show increasing benefit over time during the period of survival

followup and the pdl1 greater that are equal to 5% population tisis plus chemotherapy demonstrated a statistically significant a clinically

meaningful 26% reduction in the risk of death over chemotherapy alone with immediate Improvement in overall survival of 4.6 months in the it population tiis plus

chemotherapy demonstrated a statistically significant 20% reduction in the risk of death over chemotherapy alone with a median Improvement in OS of 2.1

months importantly the overall survival with tisis MMA plus chemotherapy and the pdl1 positive population is supported by improvements in the secondary end points compared to chemotherapy alone patients

treated with tsilis M plus chemother y has statistically significant prolonged progression pre survival with a 33% reduction in the risk of progression or death a higher objective response rate

with an absolute Improvement in response rate of approximately 7% at a 1.9 1.9 month Improvement in median duration of response next and at the fda's request

to better understand the potential Association of pdl1 expression with survival we conducted a number of exploratory analysis evaluating

subgroups across a range of pdl1 expression here we are showing a forest plot of overall survival across various pdl1 subgroups using the final analysis

cutoff date recall that the analysis of patients with pdl1 greater than or equal to 5% was a pre-specified alpha controlled analysis so I am showing that

subgroup first the 5% cut off did predict efficacy with an approximate three-fold increase in overall survival effect among patients with a PDL one score greater than or equal to 5% when

compared to patients with a score less than 5% the additional subgroups presented provide additional support for an association between pdl1 score and

magnitude of overall survival benefit as measured by the Cox proportional hazards model while we propose a pdl1 score of greater than or equal to 5% based on the

treatment effect observed in the pre-specified primary endpoint analysis we also believe that the clinical data across the studies presented today are quite similar and therefore consistency

of labeling within the pd1 class is appropriate finally because tap and CPS are designed to describe the same biology we conducted an exploratory

analysis to evaluate the agreement between tap and CPS for overall survival in this analysis the same tumor section slides that were stained with the sp263

antibody were scored using tap and then rescored using CPS we observed a high level of agreement between tap and CPS particularly among lower pdl1 scores

which are the expression levels most relevant to today's discussion and last I'll briefly review the safety results overall the adverse fed profile

for tisis maab plus chemotherapy is consistent with the known safety profile of tisis maab and other approved checkpoint Inhibitors when administered in combination with chemotherapy for

advanced or metastatic gastro cancer the frequency of treatment emerge and Adverse Events of any grade as well as grade three and greater was similar as expected a higher rate of immune

mediated AES was observed in the tisis ab arm and the majority of these events were skin or thyroid function AES though not shown here there was not an

association between pdl1 score and the frequency or severity of imaes AES leading to treatment modification were similar between the groups which indicates these treatment

modifications were largely driven by the chemotherapy component SES and AES leading the treatment discontinuation were more common in the tisis M plus chemotherapy

group in this tornado plot we are showing the most frequent treatment emergent Adverse Events of any grade in both the it and pdl1 positive subgroup the majority of Adverse Events are

commonly observed in this disease and with the chemotherapy component there is no clear trend of increase of individual AES or of AE severity with the addition

of tisis app in conclusion rationale 305 met the primary endpoint of os in both the pdl1 positive and ITP populations evaluation of our data suggests that the

benefit risk profile is most reliably established at the pdl1 greater than or equal to 5% Group which was assessed using a pre-specified analysis and a validated

assay multiple secondary endpoints support clinically meaningful Improvement within the subgroup additionally we observed modest but consistent benefit in patients with pdl1

expression between 1 and 10% Beijing supports consistency in labeling and in harmonization of pdl1 testing across the across the class of pd1 Agents as it

would help provide Clarity among prescribers and better support treatment decisions in clinical practice overall the totality of data supports tisis AB

in combination with chemotherapy as an effect Ive first line treatment option for patients with locally Advanced or unresectable metastatic gastro cancer thank you and then I'll like to ask Dr

boha to provide her comments regarding Gast regarding gastri cancer and the potential use of tisis maab in this indication thank you Dr lanasa good morning I'm Natalia ooha associate

professor of medical oncology at the University of Wisconsin school of medicine my expertise is is in upperi cancer specifically gastric and

esophageal I am also a member of nccn for gastric and esophageal cancers and co-chair of the NCI esogastric task force I'm here to today to provide some

background on these cancers the challenge in treating these patients and my clinical views on the questions posed before you I have been compensated for

my travel but not for my time preparing for today's meeting gastric cancer is an aggressive solid tumor cancer and carries a poor prognosis because it is asymptomatic in

its early stages initial diagnosis commonly occurs when the tumor is already advanced in the us alone there are approximately 26,000 new cases of

gastric cancer each year and 10,000 deaths overall The Five-Year relative survival is only 7% for patients with Advanced metastatic disease and

unfortunately we are seeing an increase in the incidence of gas and G Junction cancers in young adults similar to the trends observed in colorl cancer effective Frontline treatment is

critical to improve the outcomes of patients with gastric and G Junction cancers and many patients are not able to receive subsequent therapies because of Rapid clinical decline from symptoms

related to their disease progression anti- pd1 antibodies are now part of standard treatment for patients with Advanced gastric and G Junction Cancers and addition I of pd1 Inhibitors

to chemotherapy in the first line setting prolongs overall survival and her to negative gastric cancer patients and result in more frequent and more durable tumor

responses this was confirmed in rational 305 study as well as supported by the data from the other sponsors you have seen today collectively data

demonstrates that addition of immunotherapy to chemotherapy results in significant Improvement overall survival today you've been asked whether

we can use immunotherapy plus chemotherapy for all patients or whether we should limit immunotherapy use to those with certain pdl1 cut offs as a

clinician who treats these patients every day hear my thoughts in rational 305 there was a clear benefit from adding to lism up to

chemotherapy this was greater in patients whose tumors have pdl1 score Higher pedl One score but the benefit was also observed in the intent to treat

population in addition in the subgroup of patients with pdl1 of 1% or greater tumors we also saw Improvement of

overall survival with ha a ratio of 0.78 in this trial at the same time across several studies we see a consistent lack

of benefit from the addition of mapy in patients with tumors who have pdl1 l than one score in clinic we frequently face a question how we should approach patients

whose tumors have low pedal one score for example should patients with pedl 1 cps4 be treated differently than those

that have CPS score six for me in a patient population facing High mortality this is too fine of a line to draw I along with my colleagues want to be able

to offer immunotherapy to any patient who can potentially benefit and importantly from a practicing perspective a pdl1 cut off of one or greater should be unified across pd1

Inhibitors including their use in heru positive API tumors this approach would allow appropriate access to therapy and will preclude offering suboptimal

treatments to our patients thank you for your attention thank you we will now proceed with fda's presentation starting with Dr Viv Kumar

good morning my name is V Kumar I a medical oncologist and clinical reviewer at FDA the purpose of my talk this morning is to consolidate fda's perspectives on the use of pdl1

expression as a predictive biomarker when using immune checkpoint Inhibitors in patients with her2 negative gastric or gastrosoph Junction adoc

carcinoma the members of the FDA review team are listed on this slide my presentation is structured to outline the data from the three pivotal trials that support the use of PDL

expression as a predicted biomarker when deciding to use pd1 Inhibitors in first line her to negative gastric carcinoma importantly for the discussion

and consistent across all three trials despite the use of different biomarker assays and cut offs is that there is uncertainty with regards to efficacy in patients who are biomarker negative

especially at a pdl1 expression of less than one we know that all patients are exposed to the added risks and toxicity from the addition of a pd1 inhibitor

hence the need for a contemporary risk benefit discussion that we're having [Music] today throughout today's discussion the term pdl1 expression or pdl1 status has

been repeated on several occasions this slide does provide a broad overview of the methodologies used for assessing PDL and expression by imun histic chemistry across the three studies being discussed

all three methodologies consider cell membrane staing at any intensity as as positive for scoring purposes tumor proportion score or TPS for short is calculated by taking the

number of pdl1 positive tumor cells divided by the number of viable tumor cells multiplied by 100 combined positive score or CPS is calculated by taking the number of pdl1

positive cells which includes tumor cells lymphocytes and microf phages dividing by the number of viable Tuma cells multiplied by 100 and tumor area positivity or tap is

based on visual estimation of the tumor area which consists of tumor and anymal plastic stroma occupied by the PED positive tumor and immune cells divided

by the tumor area multiplied by 100 when we discuss pdr1 expression as a biomarker it is important to put into context the timeline for the three

studies that are subject of today's odac using data from Clinical trials.gov to provide the trial start date checkm 649 was the first of these studies to be

initiated nearly 8 years ago in October 2016 and the initial approval in a ver map for the indication was in April 2021 keynote 859 and Ral 305 were

initiated in 2018 and the subsequent approval for pism app and the spa submission for talism app occurring last year as the data from these and other Studies have matured us Bas Society and

Global Regulatory Agencies now make recommendations for patient selection based on pdl1 status there is heterogeneity in pdl1 based approaches

between FDA and us-based guidelines which not only poses difficulties for patients and providers but also for future drug development in this disease

area the current FDA label and approved indications for neolab and premisa is agnostic of pdl1 states that is to say the approval was based on broader intent

to treat populations enrolled in the respective studies checkm 649 for the N map approval and keynote 859 for the pisab approval details on efficacy on

biomarket defin subpopulations according to PDL expression were made available to healthcare providers in section 14 of the respective labels bye has submitted their

supplementary bla for the use of tism app in combination with chem y for this indication the verbiage reflects the draft label that was submitted as part of the spa this application is currently

under review and the supportive data here are from the pivotal study rational 305 in 2023 the American Society of clinical oncology have provided recommendations based on histology and

anatomic location for patients with her to negative gastric adoc caroma and pdl1 combined positive score of zero ask your recommendations are for the use of

chemotherapy only without the addition of Neola map they recommend that Neola map be added to chemotherapy with a CPS value of five or greater and for patients who are in between I.E CPS

greater than or equal to one and less than five the recommendations for the additional niola map to chemotherapies to be made on a case-by casee basis just to note that these guidelines were

published prior to the approval of pism app based on keynote 859 the nccm guidelines do not specifically advocate for the use of chemotherapy alone in patients with CPS of zero

however they do provide a category one recommendation for addition of n m to chemotherapy in patients with CPS greater than or equal to five and a category 2B recommendation for patients

with a CPS less than five for pism app the recommendations are that it be added to chemotherapy for patients with CPS grer equal to 1 with a category 1 recommendation for a CPS greater than or

equal to 10 and a category 2 recommendation for patients with CPS values between 1 and 10 another important point to note is that the guidance for patients with mism repair

deficient and micro satellite instability High tumors are independent of pdl1 status the study designs have already been outlined today and the schemas

presented within the FDA briefing document although these Global randomized control studies share many common elements I do want to outline some of the key differences between the studies that are relevant to today's

discussion checkm 649 was the only study to include patients with esophagal adoc carcinoma and did allow patients with undetermined herto status to be

enrolled all three studies included patients irrespective of pdl1 cut off and all three determined pdl1 expression status centrally but the essay and

algorithm differed for each study all three studies stratified randomized patients by pdl1 status however the algorithms TPS CPS and tap and cut

points to differ check PL 649 and keynote 859 used a cut point at one whereas rational 305 used a cut off of five for

stratification OS was the primary end point in all however the hierarchical testing strategy and the primary efficacy populations did differ checkm 649 initially compared over survival in

patients with CPS greater than equal to 5 then one and then it keyote 859 evaluated OS in patients with CPS greater than or equal to 10 and it in

parallel then in CPS greater than or equal to 1 and it sequentially rational 305 prespecified evaluation of os in patients with tap greater than or equal

to 5 and subsequently it before I delve into the study populations the three studies I want to point point out the further details uh that include differences in race ethnicity and region of enrollment have

been outlined in the FDA briefing document here I'll focus on key proportions that are important to note especially when we discuss patients included in the pull analyses that will

be presented later approximately 14% of patients in Checkmate 649 had esophagal adnoc carcinoma the her two stages was unknown

or not reported in approximately 40% of patients in Checkmate 649 as mentioned on the previous Slide the pre-specified pdl1 defined patient populations differed in the three

studies for checkm 649 approximately 60% of patients had a CPS greater than or equal to 5 for keynote 859

35% had a CPS value greater than equal to 10 and rational 305 approximately 55% had PDL expression levels greater than or equal to a tap

of5 approximately 3 to 5% of patients across these studies were known to have Micro satellite instability high or mismatch repair deficient tumors over the course of the conduct of these studies we do know that immun Che Point

Inhibitors are highly efficacious in this patient population and just to reiterate that the discussion from today's odac on PD expression would be focused on patients with microsatellite

stable disease which is a predominant population in evaluated across the three studies the MSI status was undetermined anywhere from 7 to 15% of patients across the three

studies this uh FDA analysis outlines a different composition of the intent to tree populations according to various pdl1 strata for this analysis the raw

CPS or tap score was used to provide patient classification if a particular cut off was not pre-specified in that study which has analytic limitations

patients were all in mutually exclusive strata focusing on the dark blue at the bottom one notes that Checkmate 649 enroll the greatest proportion of patients with pdl1 expression level

greater than or equal to 10 which comprised of 49% of the intent population whereas focusing on the light grate at the top keyote 859 enrolled the greatest proportion of the intent

population where the CPS was less than one at 22% and rational 305 had a greatest proportion of patients at the intermed T values between 1 and

10 now we have a sense of the similarities and differences in study design and populations across the studies I'll provide details of efficacy in overall survival in each study

clearly delineating the efficacy findings that were pre-specified populations and the exploratory subgroups defined by pdl1 States all three studies demonstrated an

improvement in overall survival in the the intent tree population with the corresponding Hazard ratios for overall survival ranging between 0.78 and 0.8 across the three

studies now he focus on what was the pre-specified analysis in the pdl1 high subpopulations at a cut off of five CPS

5 and Tap 5 for checkmate 649 and rational 305 CPS 10 for keynote 859 one can note that it is this population that appear to derive the greatest overall

survival benefit and to note also that the pre-specified analysis at the lower CPS special of one in Checkmate 649 and keynote 859 also demonstrated a

statistically significant overall survival benefit this benefit was attenuated when compared to the pdl1 high analyses specified in that particular

study the next series of slides will focus on overall survival benefits in pdl1 defined populations starting with Checkmate 649 just focusing on patients who would

be biomarker positive at a particular pdl1 threshold the population with a CPS value greater than or equal to 10 appear to derive the greatest benefit from the addition of imun checkpoint inhibitor

with a corresponding Hazard ratio of 0.65 this benefit in overall survival appears to attenuate at lower pdl1 thresholds as we go from five then one

and then it po population now if we look at subpopulations who would be biomar negative at a particular threshold there is neither convincing evidence of

benefit nor detriment in these patient populations with the point estimates for the hazard ratios being over 0.9 now these observations in patients with biomarker negative are similar to

The findings in the populations with CPS between 1 and 10 Again The Point estimates for the hazard ratios being between 0.9 and 1 with broad confidence

intervals visually the capital my curves demonstrate nicely the observations from the forest plots in the biomarker positive outlin in the top row and the biomarker negative populations in the

bottom row we see the separation in Curves in the intend to treat population on the left however the separation is most marked in patients who have a CPS 10 or greater as seen in the top right

with less pronounced separation of the curves as who work down from our CPS thresholds of five than one when looking at the biomarker negative population

there is again no convincing evidence of either benefit or harm in these patient populations now focusing on subpopulations in keynote 859 I want to

reiterate that a CPS thresold of five was not pre-specified in the study and this population was identified using raw CPS values which does have analytic limitations

focusing on the population who would be biomarket positive consistent with Checkmate 649 the greatest benefit appears to be in patients with PDL pdl1 CPS value of 10 or higher with a

corresponding Hazard ratio of 0.64 also consistent is that the benefit appears to attenuate as we go to the lower biomarket positive thresholds and the IT

population in terms of the biomarket negative subgroups patients with a CPS less than one have a hazard ratio of 0.92 with broad confidence intervals once again not providing a strong

argument for either efficacy or detriment the observations for patients with CPS values less than five and less than 10 is a little more uncertain where the observation is of modest benefit

with has a ratio is 0.85 and 0.86 respectively with narrow confidence intervals given the larger sample sizes the sub populations in the

intermediate subgroups for keynote 859 focusing on the 1 to 10 row at the bottom similar to the biomarker negative populations at less than five and less than 10 there is a stronger case for

modest overall survival benefit in this patient population as one would anticipate the Capa curves once again demonstrate that the greatest separation curves are in

those patients with CPS 10 or higher on the top right of the screen and the separation attenuates As you move left across cut offs towards the intended Tre population there does not appear to be

any separation of the Curves in patients with CPS less than one whereas that's not the case when we look at patients with CPS less than five and less than 10 whereas there's some separation

especially at the tail now focusing on rational 305 consistent with the other two studies is that patients with a tap

value 10 or greater appear to derive the greatest benefit with a corresponding Hazard ratio of 0.57 and this benefit attenuates if we look

at patients with lower tap values focusing attention on the biomar negative subgroups at each of the tap 10 5 and one thresholds there is no

convincing argument once more for either efficacy or detriment in these subpopulations similar to the findings that I discussed for checkmate 649 in patients with intermediate pdl1

expression between 1 to 10 there's again no clear evidence of benefit or detriment in these exploratory subgroups the cap plots graphically demonstrate the consistent theme that I

presented with the other two studies where in the top row we see the greatest separation of the Curves in patients with Tab 10 or greater and with less pronounced separations as we get to the it population on the left and in the

bottom row in the biomarker negative population there really does not appear to be any true separation of the curves especially for patients with tap values of less than one

in order to Anchor this risk benefit discussion all patients are exposed to the risks of added toxicity from the pd1 inhibitor and the safety across uh the

safety of the addition of the immune checkpoint Inhibitors is not known to differ across pdl1 strata across the three studies what we note is that the addition of a pdl1 inhibited to

chemotherapy will add anywhere from 3 to 11% increase in the proportion of patients to experience a grade three or four treatment emergent adverse event the incidence of immune mediate Adverse

Events was approximately 30% and we know that majority of these are low grade and endocrine however we know that up to 10% of patients will experience grade three or four immune mediate Adverse Events

and these are predominant non-endocrine events of dermatitis pneumonitis colitis and hepatitis unfortunately there were also fatal immune Medi events and although the incident is thankfully low we would

not want to expose patients to these notable risks if they're not expected to have the benefit gains from an immune checkpoint inhibitor also to help facilitate the

global risk benefit discussion across pdl1 strata we conducted a pool patient level efficacy analysis in patients with known microsatellite stable gastric or

GJ adoc carcinoma excluding patients with MSI High disease and those with esophagal adnoc carcoma this analysis will Strat ified by study now before I go over the findings I would want to

acknowledge the notable caveat of a poed analysis such as this firstly the acceptability of combining data from patients defined using different assays and the interoperability of this

approach has not been determined similarly the studies used different pdl1 cut offs so the analytic validity of presenting this uniform pdl1 starter has also not been determined

additionally the data is limited to the pool populations that FDA has access to and excludes Global Studies that have been conducted which risks the introduction of

bias with those cats in mind our pool population excluded 2111 patients from Checkmate 649 that had esophagal ladin carcinoma eight patients who were her to positive including one patient from

rational 305 155 patients with known MSI high or mismatch repair deficient tumors and 435

patients with unknown MSI status were excluded ultimately this gives us 3,348 patients that were pulled approximately

35% from checkm 649 38% from keynote 859 and 27% from rational 305 this is a forest plus of the PDO on

subpopulations from the pool analysis consistent with the theme when presenting the study results individually we note that patients with pdr1 expression of 10 or greater derive

the greatest benefit with a hazard ratio of 0.64 and that this estimate of benefit attenuates at the lower thresholds of 51 in the overall population when we discuss efficacy

findings in the biomarker negative populations I do want to point out that 177% of patients would be classified as biomarker negative at a pdl1 expression

of one 45% at a pdl1 expression of five and 62% of pd1 cut off of 10 in this pool data in terms of estimates of benefit

the hazard ratio for PDL 1 less than one is 0.91 similar to the observations of PDL 1 less than 5 and 10 where the are confidence intervals are narrower given the larger patient

populations when we discuss uh patient population with pedi Expressions between 1 and 10 comprising 44% of the poor patients the estimates for efficacy in this patient population similar to that

of the biomar negative subgroups with the point estimates for the hazard ratio for the to 10 subgroup being 0.93 so in my presentation I provided an overview of FDA perspectives and

analyses of three pivotal first line studies submitted to FDA that argue for pdr1 expression being a predictive biomarker when deciding to utilize an immune checkpoint inhibitor in patients

with herto negative Advanced gastric adoc of outline concerns at the agency of modest estimates of efficacy especially in patients who have pdl1 expression of less than one these

patients are of course exposed to the ad toxicity of pd1 Inhibitors and it is this patient population that are predominant focus of our risk benefit discussion FDA would like the committee

to discuss whether in patients with herto negative microsatellite stable gastric GG and carcinoma does the cumulative data support the use of pd1 expression as a predictive biomarker

When selecting patients for treatment with pd1 inhibitor and for the voting question is a risk benefit assessment favorable for

the use of pd1 Inhibitors in patients with Advanced herto negative microsatellite stable gastric GG carcinoma in patients with pd1 expression less than one now although

we're specifically asking about the cut off of one we invite you to express your opinions if you believe another cut off would be more important thank you thank you we will take a 15-minute

break panel members please remember that there will be no discussion of the meeting topic during the break amongst yourselves or with any member of the audience we will resume at 10:15 Eastern Time

e e e e e e e e

e e e e e e e e

e e e e e e e

e e we will now take clarifying questions to the presenters uh when acknowledged please remember to state your name for the record before you speak and direct your question to a specific presenter if

you can just some notes for those in the room if you have a question please turn your name placard sideways uh so that we can see that and also because we have three applicants uh in the room please

direct your specific question to a specific applicant if you wish for a specific slide to be displayed please let us know the slide number if possible finally it would be helpful to acknowledge the end of your question

with a thank you and an end of your follow-up question with that is all for my questions so we can move on to the next panel member for our panel member joining us virtually please use the raised hand icon in Zoom to indicate

that you have a question and we'll acknowledge you please remember to lower your hand by clicking the raise High hand icon again after you have asked your question are there any clarifying questions for the

presenters Dr Hawkins thank you very much I have two questions if there's time the first one

has to do with um better testing for um PDL expression and this is directed to Dr Anders and Dr Jen

jingan um so bit concerning about the variable in testing so my question to Dr Anders or both of you is whether we can be

optimistic about better testing um is there an ability to think about artificial intelligence helping us in some way given the the statistics he

showed us about the pathologist and then um I noticed we talked about different testing Dr J Jian

talked about I think uh how they do their own created their own um pedl assay assay question thank

you thank you Robert Andrew Johns Hopkins pathology thanks for your question can I have slide one please so my understanding is you're

asking is there hope in the future for better methodologies well those methodologies are trained by a gold standard and the gold standard is the

pathologist read so it's a little bit of a back and forth where we're only as good as how we can train and as you can see from the graph

there um a patient's eligibility at one of those cut offs at 5 and 10 may be determined by which pathologist looks at it does that mean we should not be

optimistic we need a better gold standard response any studies on artificial intelligence help us they're trying there's a huge push on

it we're working on it thank you hi Elena jenan thanks for that question um I guess we can bring up slide 22

cg22 um and uh I'm glad you picked up on this yes lab develop tests means that uh a clear approved laboratory can validate

in a small Gord of samples um you know another antibod to use and as long as they prove certain level of concordance um most institutions are allowed to

decide what uh I test they will use um some of these uh for example 22 C3 and and so some of these dco Technologies um

are considered relatively outdated I guess and the readers and the stainers most institutions won't in invest in buying these new these older um machines at least that's what the Pathologists

tell me and we have a pathologist here to comment on that so um there is a learning curve and even at an institution as ours like ours um

we use our own lab develop test and I can tell you we expect for example to have a certain rate of pdl1 CPS score five or greater or 10 or greater and one

or greater and there was a steep learning curve in our testing when I did uh the analys review of our clinical samples uh initially when the testing

first began the rate of positivity was significantly lower than we would have expected from several phase three studies some of the studies you saw here but also other studies coming out such

as with other agents pd1 and everybody has a you know each company has a pd1 inhibitor but the um that's my big concern is that there is variability in

testing and in terms of your question about AI um you know AI is is excellent in certain things these type of samples are quite um heterogeneous and I I

honestly don't think in my life as an oncologist lifetime as an oncologist um we will be able to replace experts and Pathologists because they're able with

machines because they're able to tell us how good the quality of the sample is and where to look um and I think that's why if I I you know I don't want to

undermine the quality of the biomarker work that's been done by these companies but we're not we don't have controlled environment with two you know uh trained Pathologists sitting in the same room

looking at a tumor block that was an entry criteria the trials mandate that that tumor quality has to be a certain level for them to even enter into a

clinical trial in clinical practice anyone with one uh slide that says cancer can come in and start therapy um and we do start therapy because it's

urgent so it's a different um different situation thank you both I can Circle back R up another the question but I don't want to hog the mic thank you Dr Hawkins Dr

Sprat thank you so much uh Dan sprad uh Case Western a quick a couple questions I may very direct uh to the FDA um did you perform any interaction tests given

the question at hand is this a predictive biomarker uh thanks uh iming John statistical reviewer of FDA uh yes we did do the um interaction effect test we

added one uh treatment by continuous interaction term to the uh stratified model in the PO analysis and the interaction effect is um interaction

effect is statistically significant but here I would like to acknowledge this is the UN prespecified analysis and we we caution to interpret the statistical

significance here and the effect the direction of effects um showing that the increasing uh treatment benefit in the pdl1 high subg groups appreciate it

thank you so much and a question for BMS this can go to Dr Anders as well um understandable why you showed your study

uh there's a study just came out in June of 2024 first authors uh Dr clemer who leads the MGH program the pathologist

from mskcc is on this paper they show that across all three approved essays um that there is moderate to almost perfect intraassay Capas as well as substantial

to almost perfect intraassay agreement in gastric cancer so can you please explain this especially focusing on outcome is what matters and we are

seeing the outcomes vary by pdl1 expression level so clearly the pathologist can't be that wrong here given the interaction effect is

statistically significant sure again Robert Anders Hopkins pathology so thank you for the question question and indeed one of my slides I showed data from the

clner paper with the staining um so I am aware that there they showed good concordance now we stepped back in our

in our study and and took the people who are scoring this every day and you know you consider those to be experts and and and we just had poor agreement now what

I do think we can agree on is if there's any PDL one expression when we start to move above to thresholds that's when the

um the agreement begins to really fall apart the the agreement in that study for greater than five or greater than 10 was around 75 to8 so just putting out

there but I appreciate it and again there was a concominant study done and my Dr rim and we came to the same conclusions okay the thank you so much

the final question is going to whoever um from Merc feels it's appropriate to answer you put in the briefing documents acute twist analysis that wasn't

presented um here but this basic is an analysis that's trying to weigh toxicity versus progression and benefit and you show the intention to treat and the C offs are greater than one and greater

than 10 staining um however can you have these analysis for less than 10 or less than one given that most of the patients in greater than one in the intention to

treat are enriched for the high pdl1 expressing levels um uh Kathy pansa from Merk uh so

we have uh done uh Q twist analysis in uh all uh CPS cut points um slide one up

please um I'll have uh so slide one up here please this shows the um the entire population and in the greater than one

and the greater than 10 uh where uh so for others that didn't read the background package Q twist combines efficacy safety and quality of life in

single measure um and so uh we performed Q twist analysis to evaluate the quality and the quantity of overall survival classifying time spent for toxicity and

without toxicity before progression as well as time after progression you can see here that in the ITT uh the 20 .9 uh

the relative Q twist uh shows an improvement uh so in in the literature a relative Q twist Improvement of greater than 10 of 10% or greater is clinically

important and a gain of at least 15% is clearly clinically important thank you um but do you have the analysis again given almost all of these are just enriched for high pdl1

expression do you have it for less than 10 or less than one yes so so uh we have it uh yes slide two up please and I will

actually have uh Saka Peters uh explain this a little bit further sonaka Peter MC epidemiology um so I think we had a nice

explanation of what uh what qist is so as you can see here um based off of the

um clinically important cut points we do see that relative Q twist gain in it and cut points that are of interest um of

course in those that are less than um less than one and less than 10 there's a small sample size for these subgroups

and alyses so while there's relative uh gain they are not clinically important thank you so much that cludes

my questions thank you uh Dr Von hi Neil Von this is another question for Dr Anders um I wanted to just push uh a pressure test this idea of the dichotomous variable versus the

continuous variable here and you had said that you think that Pathologists have good concordance as to whether pdl1 is positive or not and so um if you could please pull up slides CG 27 this

again this High inter Observer variability are you able to quantify um what percent of Pathologists let's say would have scored patients as greater

than one based on your data here and also how does that compare to other cancer types uh just just as a gestal yes thank you again Robert Anders

Hopkins pathology um I'm we have a distribution curve of the positivity

rate um I will say that a majority about at least 50% of the patient samp the 100 samples that we've looked at fell um

within the the one or greater so I would estimate maybe 20 or 30% were below um I don't have but I guess we could pull that number for you what

percentage of Pathologists call something positive because it seems from this that there's clearly a spread greater than one just from looking at the entire um individual data points certainly the

number between zero and one for each individual patient I think is a smaller number yeah so yes that's a very important uh observation so I if there's

absolutely no staining there's no brown color that's easy right that's zero what happens is when there's some staining but it's just you know may it

might be just a little bit where it's not nearly enough to count for 1% we're we're left in this limbo saying well we can't say it's nothing right so

we sort of compromise and have that let's call it a half as saying well there is staining but it doesn't meet this one for one

threshold okay thank you you're welcome I'd like to invite Dr Jen jig in as she has another comment I just wanted to comment on the

clumpner uh question um just the purpose of that paper was uh to really look at reected surgical samples so it's 100 surgical samples not what we see in the

clinic and also the purpose of it was to look at 22 C3 288 with sp263 antibody compare those and also to see if CPS and

tap is sort of comparable um so it's very different to what we see in the clinic uh yes you can see tell if something completely zero um but again

often when that happens if I call the pathologist and I clarify with them how many you can cells or any um stroma even in that sample often the answer is well

very few cells so we can't look at the surgical database um and extrapolate that to clinical practice because most of our patients don't have surgeries

they present with small endoscopy samples uh with stage four denova diagnosis uh Dr myart do you still have a question or is your question already answered yeah it was already answered

okay uh Dr DOD yeah this is Lori D my questions for Dr Anders so there's uh been a lot of discussion and data presented about the

inner Observer variability you also mentioned the spatial variability spatial heterogeneity um and I didn't see any data that that really T speaks to the

degree of spatial heterogeneity and the factors driving that heterogeneity are there data that can be can support that and is the heterogenity driven um or

does it result in um differences in the quantity ative assessment of the of the um pdl1 expression or is there also heterogeneity in the evaluation of

presence or no or lack of expression of pdl1 okay great John H uh Bob Anders Johns Hopkins pathology can I have slide

one um so um there was no quantification I have not seen quantification of of heterogen that's part of the uh reason I wanted to show this particular slide um

the expectation for a CPS score on that particular slide would be to count all of the cells whether they're positive or

negative um so the CPS doesn't really have any uh leeway for heterogeneity uh or or way to account for it Dr W and just to clarify though

if there were multiple samples taken multiple biopsies taken do we have any any information about the degree of variability across multiple samples

taken from the same patient right so um most of our biopsy samples are just like the pink area up there would be endoscopic very superficial uh it's

typical and in my reports I will give feedback to the endoscopist and say two of the five endoscopic biopsies contain invasive gastric carcinoma so there

there is tremendous variability and largely under sampling in those samples Ian Waxman BMS um if I could just add we have to answer your question Dr D about

spatial heterogen we do have data for metastatic versus primary sites and there are differences I can call Sarah hery from our Precision Medicine Group just take you through a little bit of

those data to support that Sarah hery BR Meer squid Precision medicine I'll start off first by sharing with you slide two which shows pd1

expression is dynamic and in this publication what you'll see is within the same patient that was sampled they took both primary and metastatic sampling and what they found was that

the agreement was only 61% that was regardless of if it was a cps1 or a CPS 10 cut off in our own

clinical trial we did an exploratory ad hoc analysis of the data and I would caution that the sample numbers were small but we did see differences there

as well between primary metastatic sampling thank you Dr D does I answer your question yes thank you Dr

Hawkins thanks again this is a question about quality of life and it's directed towards uh Dr bajia and Virginian MK and

I apologize a person from biogene who also mentioned quality of life but I I missed that name because I came back in a little late and so my question is folks want to live that's why they're involved in these

trials um and they want to live as long as they can and we saw information about Adverse Events my question is not about Adverse Events my question is about to to do any quality of life objective

assessments what and if you did what tools you Ed can you share some of that with us that answer my question uh Mark lanasa Beijing uh so we

did collect quality of life data in the rationale 305 study I'll share those data uh so we collected a battery of uh General quality of life assessments

through QQ C30 scores as well as disease specific data QQ st22 domains for gastric cancer what I'm showing here are the C30 scores this is the time to

detriment Hazard ratio uh you can see that there's a slight favoring of uh the combination of pilis M with chemotherapy across these

domains so for work we'll have Dr Yan fangan actually respond to this question follow followed by Dr Jen again y patient center endpoints and

strategy from mer so um at mer we for Kino a59 we also use uh the eii TC cure C

measures for uh quality of life assessment and these are well established uh cancas uh specific uh validated P

measurements and um we can show the quality of of Health R quality of life over time um for the ghs

ql um so for our study the patients completed the questions uh during treatment and um at Baseline the quality

of uh the scores of the health really quality of life were are similar to The General cancer patient population and during treatment slide uh three up

please so during treatment the quality of life scores were generally um stable and similar between the treatment arms uh as indicated by the overlapping

confidence intervals no document was observed with the addition of pimpis map to chemotherapy compared to chemotherapy

alone so uh this data supported the overall benefit race profile thank

you slide down slide down please Dr J yes we're we care about quality of life it's and it's very important because this disease is incurable and

most patients need to live lifelong with it um so I think when you think think about quality of life obviously cancer related symptoms is what drives it uh in

my experience the patients typically respond within the first um two to three Cycles so the the quality of life does improve um I think most of the side effects as I mentioned come from the

chemotherapy the addition of immune checkpoint blockade has very minimal uh impact on the quality of life and overall um it remains um adequate and um

perhaps more improved although the confidence interval are overlap I think again I don't want to make this too anecdotal but when you think about quality of life a lot for these patients

it's about the me mental set set of continuing therapy lifelong and to have this hope of having a longterm survival um right with immun checkpoint blockade

um immediately sort of lightens the atmosphere in the room and gives them sort of the strength to keep going so I think that you can't capture on the on

the quality of life questionnaire but I think uh it is a factor with these patients Ian Waxman BMS we we also have collected quality of life data we used a different instrument the fga can I have

slide one please just to uh remind this audience this is a 46 item questionnaire that covers five subscales one of which is specific to G gastric cancer the other four of which are covered by the

more general fact G um if I can now have slide two please what we saw was that quality of life was at least maintained in every subgroup and you can see this includes the CPS less than one

population albe it small numbers for that subgroup thank you very much and Dr

Sprat so this question I'll direct to Dr jigan is can you explain in that we see very clear with

every single one of these drugs a in relative benefit based by the CPS or tab score very clear two of the companies actually in the presentations admitted it's predictive in their talk

statistically it's predictive so clearly these scores are correlating with outcomes so can you please explain when you state and I'm

hearing from Dr Andrew there is just such spatial and heterogeneous outcomes we have poor inter reader um variability

but yet there is consistent by drug by pathologist by assay so C can you explain that because it should just

be random if it's that poor of correlation well um yena jigan uh medical oncology it's we're looking we're making a a distinction between

populationbased biomarker versus a clinical biomarker in the in in the real world um and I I'm a researcher I am all

for biomarker testing I think it's important to be able to translate what you discovered in the clinical trials to the clinic and as I mentioned uh earlier

it's the quality of the sample and the quality of the testing um you know you didn't see the data on uh screening failures from Checkmate 649 and keot 811

I was the global Pi for both of these studies and I can tell you many patients did not make it onto the trial because their tumor quality was insuff efficient and those are the people that uh will

never know if their pdl1 testing was uh conclusive and so you know to understand that uh it's not black and white but it's a spectrum and we're not saying

that everybody should get everything and they should get it at CVS we're just saying let the doctors decide uh What patients actually would benefit and the

data suggests that doctors are actually pretty good at following ASCO and NCC guidelines um and they're not overtreating the patients um it's the

patients who come in as I mentioned with unavailable sample or the sample quality for um there's it's a lot of barriers to getting these patients started on

therapy thank you again this is Dan sprad um it didn't really address the question because you're talking about the real world but we're talking about here that the question is in these

studies it you have the tissue you reviewed it is there strong correlation or not between Pathologists

and if there's not why is there such strong correlation to outcomes a separate whole question is real world applicability sure well I guess let's C

I mean I think we are talking about the real world because we're not in a trial not the question I'm asking but I'll let perhaps and these drugs you're presenting are from clinical trial not

the real world so I'll have Dr Chazy pres uh answer that question Dr Sprat thank you vladislav shki anatomical pathologist diagnostic

Reference Laboratory yes uh very good question and the idea is that in a clinical trial the Pathologists were trained specifically for the score and

the repres reproducibility was great in the clinical trials and as that was mentioned before not patients who did not qualify under three criteria did not make it into the

trial so in the clinical trial it represented very well from the score to the response as you can see it clearly in the real world as we try to point out

you know there's different um issues that come up these issues have been brought up before and they are very specific and very important issues

however I I just wanted to point out these issues are not specific to pdl1 they are not specific to the score as CPS or tpf and they are not specific to

the organ and yet we see in numerous examples we have herto in breast cancers that have been over time we standardized it we were able to show that there is a

reproducible effect and the same thing should apply here what we're lacking right now is a standardization of scoring we have ldt we have different clones we need a standardize the scoring

we need a standardize the practice of doing it for example biopsies in breast are if negative are followed by reection scoring I understand it's not always

possible in clinical trial but some sort of standardization we should improve the overall response of from pathology to

the clinical practice thank you much appreciated thank you thank you H Dr Gibson uh Michael Gibson thank you chair for the time I defer if my question is

for a different part of the session but I'm I'm new to this just a few thoughts um it sounds like we're trying to decide uh as we always do in clinical

medicine a risk a benefit ratio and we're using data which is number one subjective we have a patient advocate here who could maybe comment more on

what quality of life really means in the real world uh but secondly the main conclusion I have is that our assay is extremely flawed uh for many many reasons but unfortunately we have to use

what we have and Dr gbon we'll certainly discuss discuss the question about that Mr do you have a do you have a specific question for no I I did I did just have

a question uh regarding uh if if the if the panel thought that as we have a constant uh adverse event effect across CPS levels uh and your benefit is

inversely proportional to the expression is there a concern that the uh risk benefit ratio shifts As you move closer and closer to a CPS of zero and is that something you consider in your decision

and Dr Gibson is that a question do you have a specific question for the FDA or an applicant I'm sorry I apologize oh no otherwise yeah I'm sure we will definitely get there more to panal discussion for

sure uh sure Dr I mean sure exactly that's what we're getting at is is that you know if there's no benefit and you have toxicity well then that certainly is a different different situation where

if you have you you'll have you know toxicity but that's you know in a we take the CPS greater than 10 where there's clear benefit you know that that trade-off is clearly you would take the

drug um I don't think anyone would sort of um you know go against that here but certainly the trade-off does change as as you go down and so that's what we're asking for you to consider thank you for

your patience with me thank you thank you both uh and Dr mcken so Heidi mcken from the community oncology setting um looking through the data and

hearing the presentation today we can clearly see that there's uh a significant risk for toxicity so if we're looking at treatment related AES um especially immune mediated some of

the studies would say 30% chance that these patients are going to get immune mediated side effects but then looking at the hazard ratio for patients with CPS less than one I mean it seems like

the benefit uh for those particular patients is likely uh less than 10% so we heard from BMS in Beijing about uh potential uh recommendations about

looking at um continuing approval for patients with CPS greater than one my question is for anybody on the Merc team how do you justify treating these

patients with CPS less than one with immunotherapy when it seems they would have a higher risk for side effects than benefit thank you Kathy Panza from Merc so we understand that this is a very

important question that we're here to address today today when um we when we look at our data we look at um the

overall clinical um risk benefit for the entire treatment population as well as subgroups and here we did look at pdl1 less than one and when we see uh when we

see that the hazard ratios are consistent with the intention to treat population we have acknowledgement that the that the entire patient population is uh is is

benefiting we recognize um that the benefit the magnitude of benefit may be less than CPS less than one however the

the hazard ratio was less than one and pdl1 is a Continuum and the score does not predict who will respond there are patients with low to no expression that

responds and there are patients with high pdl1 expression that respond we acknowledge that the safety is an is an

issue uh however the safety of pemis ab across all pdl1 subgroups was me was the same and maintained as was Health rated

quality of life MC really wants to keep the label as it is because it gives patients an options and it gives Physicians an option in clinic when

faced with a patient with a fatal disease to make that decision that the label has the information as does the NC

CCN ASCO and uh and esmo guidelines guidelines can help guide the Physicians having a full a label have a broad it

label will enable the option for all patients that answer your question Dr mcken does can I ask one more question uh sure if it would be brief though uh we're starting to run out of time here

oh sorry so my question is about those nccn guidelines are you all three companies seeing across the country that if nccn categorizes a recommendation as category 2B that insurance will not

cover those medications so um I would have to ask Dr Jen Jan uh because I don't I no longer see patients in clinic sorry can you clarify the

question the question is whether or not the nccn guidelines affect the practice correct so if the medication is categorized as category 2B insurance

companies are starting to not cover that are you seeing that around the country um a bit uh I think it depends on the pair typically a a phone call to the insurance company that clears that up

though so i' if again it's a case-by case basis most of the time if it's a negative case and we think that tumor quality is um sufficient uh we would not

prescribe immunotherapy so um for CPS completely negative uh cases so it's a it's a rare occurrence that um we have to deal with

this Maring i' like to share her we could have her mic Natalia boha University of Wisconsin um I would like to to Echo that it's

very important um what actually makes an nccn guidelines the insurance companies do pay very close attention to what's on the guidelines and we are running into more and more struggles with having

limited access for the patients to medications because of how the medications are ranked on the guidelines thank you thanks your

question Dr mck thank you uh Dr hilard uh yeah this is a question for the most team if you could put up the slide which was the OS and PFS

directionally consistent across all the PO cut off points the can we have the farest plot from the from the key

presentation slide one up please so yeah so this is what I wanted to see if I'm inter in correctly that in

general with all of the studies we've seen there is a a positive overall correlation it would appear between the

um PDL higher cut off points and uh better response on the other hand if I'm reading this correctly even the people

with a score of less than one still most of them did benefit so again if we were just looking at this as a single point then well that's not statistically

significant but then again neither is the uh space between 5 and 10 and so um I guess really from a

patient perspective um I would kind of like to have the option of trying uh these regard just based on these numbers

and also based on the numbers that the level three and four side effects were only three to 11% greater with the immune checkpoint

Inhibitors um so um is that a correct way to think about it uh yes that is the way MC is thinking about it and that there are patients

that do benefit albeit with a lower uh although the magnitude of benefit is less but there are patients that do benefit and we want to maintain that benefit for some patients with this fatal

disease thank you can I add to that if we keep comments short please yeah so we we've actually looked with our longer follow-up data we have fouryear follow-up data now in that specific 1 to

10 population where there's the question and we do see improved overall survival Hazard ratio now um with separation of the curves it was not there at the initial time of the lock and we have

that Kaplan Meyer um and I can pull it up for you in just a second but what it does show uh slide one please just very briefly on the left hand side of the slide you'll see the Kaplan Meyer for the one to 10 population at the time of

the initial lock and then with the four-year followup a clear separation with significantly less censoring has ra your decrease to

088 Dr Hiller does that answer your question yes okay uh Dr mad sure as we uh kind of grapple with

this I'm trying to understand the you know this this kind of binary cut off of one and what it means so we focused a lot on on the toxicity it brings those

patients who are below one I'm trying to I think the more relevant another relevant part is the Mis benefit in those patients below one who could get treated and get and get benefit and so

I'm not sure the best way to understand that but is there response rate data from the the the patients who have the lower scoring I don't know if you guys should resp I know because we've been

looking at this as largely a population and clear C the higher expression the population is going to do better but do you guys have response rate data from these low um expressing

patientsan Waxman BMS we do have response rate data in the CPS less than one population and uh we can pull that up for you as we're getting that slide slide one

please what we see here is a improved response across all pdl1 subgroups including in the CPS less than one it's about a 7 and a half% Improvement and response in that particular population

with all the other C offs listed here as well and sorry Kathy Panza with MC I like to up please um here again as we as

uh as Dr Baga explained um patients in the less than one subgroup have a uh have a improved uh progression free survival the median uh progression free

survival Improvement was about one and a half months the with a difference in objection with uh with the increased objective response rates for

these patients and an an improvement in median duration of response compared to chemotherapy alone so these are also

important uh clinical uh end points for patientsa we're showing sorry sorry Mark lenasa be here we're showing a forest plot of response rate across all the groups being

discussed today uh you can see that the objective response rate is favorable does favor the investigation alarm across all these groups though the confidence interval is very very wide in

the less than 1% group uh and the incremental uh increase in response rate is relatively modest between 1 and 10% uh Dr Kumar did you have a response

or a question just wanted to add to uh Dr hillard's point about um less than one populations and and BMS is respon specifically with the cap M curve for

the 1 to 10 uh if BMS wouldn't mind also just for Mr Dr hel's benefit also showing the capar curves for the less than one uh in the in the with the

prolong followup sure we can pull that up for you as that's coming up in the less than one population the hazard ratio remains close to one just in the

0.9 range uh but let me get the latest version of that up for you here uh slide one please while we saw that improved Hazard ratio in the 1 to 10 we

did not see that same Improvement in the CPS less than one here's the original as well as the four-year follow-up just wanted to clarify that because that was a specific question

that Dr H had thank you drar thank you yeah thank you uh and then Dr myart uh this question for Dr Anders so both during the presentation and question answer you indicated that you feel

confident that Pathologists are pretty good at less than one uh that we've also heard multiple times about the concern that a lot of these patients with Mantic

disease have very small biopsies or just uh endom Cal biopsies so I just want to know if that statement holds for people who have just a a small biopsy versus a

whole tissue reection yeah Robert Anders Hopkins mology thanks so um the requirement for

CPS is that there's 100 uh cancer cells uh to be present that's the minimum for for the score so if if the tumor cells are present I I feel confident that we

can you sort of mediate whether there's positive or negative in fact if there's fewer cells it might actually be a little bit easier but my concern is that

smaller samples or superficial endoscopic samples under represent the tumor so if a patient we look at it we do everything perfectly everybody agrees

that it's zero in that sample it's it's that the the biology really is in the deeper invasive edge of of the cancer does that

answer your question thank you and Dr L did you have a question response thanks Ste L I just wanted to uh just respond a little bit to the the

response rate PFS sort of discussion you know again with PD pd1 Inhibitors we've seen funny things with response rate with not good correlation between different effects and response rate and survival and they go both ways sometimes you'll see benefits and response and no

benefit in survival and sometimes you see the opposite and so I think we have to be careful with sort of reading too much about these response rates especially this includes some patients that may be in there with them say high

tumors as well so I think we have to be a little bit cognizant I think you know we're predominantly looking at survival where we see you know Capital Min Cs on top of each other um you know acknowledge if you're a patient

individual patient better to have a response than not have a response but I think when we look at overall population data we we've seen again funny things with response rate we're looking at you if we're looking at PFS FX with one month well you know what does that what

does that even mean and so I think we just want to be a little bit careful when interpreting some of that data thank you Dr L uh seeing no other questions we'll conclude our clarifying questions

portion uh this meeting uh and we'll move on to the open public hearing uh session we will now begin to open public hearing session both the FDA and the public believe in a transparent process C for information gathering and

decision- Mak to ensure such transparency at the open public hearing session of the advisory committee meeting FDA believes that it is important to understand the context of an individual's presentation for this reason FDA encourages you the open

public hearing speaker at the beginning of your written or oral statement to advise the committee of any Financial relationship that you may have with the applicant for example this financial information may include the applicant's payment of your travel lodging or other

expenses in connection with your participation in the meeting likewise FDA encourages you at the beginning of your statement to advise the committee if you do not have such Financial relationships if you choose not to address this issue of financial

relationships at the beginning of your statement it will not preclude you from speaking the FDA and this committee plays great importance in the open public hearing process the insights and comments provided can help the agency

and this Committee in their consideration of the issues before them that said in many instances and for many topics there will be a variety of opinions one of our goals for today is for this open public hearing to be

conducted in a fair and open way where every participant is listen to carefully and treated with dignity courtesy and respect for those presenting virtually please remember to unmute and turn on

your camera when your op number is called for those presenting in person please step up to the podium when your op number is called as a reminder please speak only when it when recognized by

the chairperson thank you for your cooperation so speaker number one please state your name and any organization you are representing for the record you have

three minutes hello my name is Andrea idelman and I'm the CEO of Debb's Dream Foundation curing stomach cancer we are the largest

International patient advocacy group for stomach cancer I have not received any compensation from any sponsors or speakers for my presentation here today

I am here because this issue goes to the heart of the mission of the Debbie's Dream Foundation co-founder Debbie zelman founded the organization in 2009 for one

year of being diagnosed with stage four incurable stomach cancer Debbie was just 40 years old mother of three young children practicing attorney and the wife and daughter of a prominent

physician Debbie found through her own personal Journey that there had not been a new treatment for gastric cancer in over 30 years her life mission became to

start to start the foundation to fund research and provide as many treatment options for stomach cancer patients as possible as CEO since

2017 I have seen patients struggle through the same Journey as Debbie and I have personally interacted with patients who have benefited from these particular treatments that are being discussed

today and they have made an extraordinary impact dds's position is to maintain access to immunal therapy

for patients with low or negative pdl1 scores uh and that is necessary because it allows more access to treatments there is already a lack of treatment

options for gastric cancer patients and allowing this access we have seen through our own patient Community which you will hear uh from today uh that this

these benefits um have been a seen for these patients patients want and need to be empowered and want to have shared decision making with their Physicians

patients in this situation mostly 80% are late stage stage 4 and there is a sense of urgency in being able to access treatments

immediately this satisfies the patient's desire to take action and take a sense of control over their illness our longtime DDF patient and

Mentor Amy Jacobs has also submitted um a written letter and shared her own survivorship personal Journey with these treatments and the importance of

allowing patients and Physicians to decide what is best for them in their particular situation please don't take these choices away thank you for your

time do you have any questions for me thank you speaker number one speaker number two please state your name and any organization you are representing for the

record my name is oi Smith I'm a caregiver Patient Advocate and the founder and executive director of Hope for stomach cancer while we do receive independent ground funding from a variety of sponsors

including those represented here I am not being compensated for my time travel or expenses to be here I'm here today to share my father's story and express my concerns about the potential impact of

FDA cutoffs on treatment decisions and patient access to immunotherapy my father was diagnosed with Advanced stomach cancer in late 2013 and given 6 months to live by our local hospital a

second opinion saved his life they read stage to Stage 3B and dis and discovered he was hair 2 positive at the time her septum had just been approved and while it was considered experimental in his

Curative treatment it gave him a fighting chance today we typically don't use herceptin in a Curative setting but the flexibility that ex existed back then allowed my father to benefit from a

treatment that possibly cured him one of my main concerns is how FDA cut offs could restrict a doctor's discretion and treatment once these cut offs are in place Insurance per companies will

likely follow suit refusing to pay for treatments outside of these parameters I've seen firsthand how Insurance can influence life-saving decisions for example my father's Power report was

initially denied forcing us through a lengthy appeals process while his a herceptin was approved these kinds of decisions can profoundly affect the quality of life um and Care a patient

receives as the founder of Hope for semic cancer we provide navigational support to patients and caregivers through our programs I've learned learned many things about our healthc care system and the disparities that

restrict patients from um accessing novel treatments not all patients are tested for biomarkers not all patients know their biomarkers through our website stomach cancer biomarkers dorg

we've developed resources including charts and nccn guidelines summaries to help patients navigate biomarker testing and treatment options while I believe these guidelines are crucial we must be

careful not to take the flexibility that can save lives stomach cancer is a deadly disease and for many patients treatments are measured in months not years in cases where doctors are

weighing toxicity against potential benefits we need to remember that many patients are facing fatal outcomes regardless of their treatment restricting access to treatment based

solely on biomarker cut offs May mean that some patients lose the chance for life um extending therapies we must balance the science with the real world

complexities of patient care ensuring that that doctors retain the ability to make decisions tailor to individual patients I want to thank you so much for your time and consideration we did leave

a video in the um open comment that was created by patients and I encourage all of you to watch the video thank you thank you speaker number three please state your name and any

organization you are representing for the record hello my name is Alison cchu and I'm representing um more I'm here on behalf of MK I'm not being compensated for my I'm here today however I did

receive support from my travel from Mark my sponsor I'm A 42-year-old mother of two speaking as a caregiver for my husband Ron I'd like to share my experience advocating for my husband's treatment options and how he has benefited from immuni therapy despite

having a low pdl1 cut off and poor biomarker expression Ron has diffused gastri cancer which is an aggressive underresearched cancer with a pathology that tends to be chemoresistant resulting in patients being subjected to

multiple lines of treatment to keep it stable he was 47 years old and relatively healthy when he was diagnosed via routine endoscopy with Stage 1B gastric cancer in the fall of 2020 we

sought out multiple oncologist opinions from NCI designated facilities across the us all of whom echoed the benefits of mapy however at that time it was not approved in the first line setting therefore we opted for the standard of

care prayer and CCN guidelines which he completed Ron had a partial gastrectomy in March of 2021 wherein we learned he had zero chemo response and will be upstage to 3B he had more chemo and radiation but would ultimately have a

reoccurrence in the fall of 22 it was then that we sought out a imun therapy to be included in his next line of treatment we knew the odds were against us in getting Insurance approval due to ronow pdl1 and biomark with threshold

but our oncologist advocated for us to receive it as we collectively knew that systemic chemotherapy alone would not be enough to fight his cancer as we have feared we were swiftly denied multiple times by our insurance company for lack

of statistical proof but after about two months back and forth which was very timely when you have stage four cancer at this point uh we were able to receive Kuda via Compassionate Care the feeling

of having no other option besides chemo despite seeing the stability of immunotherapy and clinical trial settings as well as other patients took an emotional toll on Ron Ron is a part of a younger population who are seeing a

rise in digestive cancers and who deserve to have access to potentially life extending treatment options our surgical team is part of an NCI designated research hospital and they twoo feel strongly that despite Ron's

treatment history and his low pedia1 score kitura the immunotherapy utilized in his case is doing the heavy lifting and keeping his disease stable he is still on it 2 years later with low uh

systemic chemotherapy as well it has awarded him a decent quality of life and disease stability he is tolerating it well it has given us eth grade graduations vacations and cherish memories of which I am hopeful there

will be more if we take away these options for patients like Ron we're not only losing an opportunity to observe immunotherapy's effects and clinical settings like his but we're also doing a disservice to patients and their

families many of whom are young and have so much to lose so for the sake of gastro cancer patients and medical research please consider continuing to provide immuni therapy regardless of pdl1 indication for which patients like

Ron may continue to receive life extending treatment he is your data and he's the PA of your science thank you thank you speaker number four please

state your name and any organization you are representing for the record hello my name is Kimberly Wilson I'm not being compensated for my time here today however I did receive support for my

travel from mer one of the sponsors may I have my first first slide I'm a Maryland resident and a stage four aagal cancer thrivor but more

importantly I'm a mother wife daughter sister aunt and friend next slide please sorry in April 2022 at the age of 43 I

was diagnosed with stage 4 esophageal IND chroma and my GI Junction the diagnosis hit hard and continues to impact my life and those around me daily

my diagnosis came exactly six weeks after marrying the man of my dreams with the support of my family I received preoperative chemo radiation underwent a 12 and a half

hours refilled mowen esophagectomy that resulted in clear margins but unfortunately I was faced with a reoccurrence 3 months later I personally would like to thank all attendees and

participants who are here today I recognize that everyone in the room is working to create greater awareness surrounding the topic of a esophageal and gastric cancers whether it's working

towards finding a cure uncovering new treatment options exploring the possibilities of conjunctive therapies and more as a patient it brings me great joy to see that there are people here

who are interested in the topic and people who work and understand scientifically what this disease encompasses next slide please today I come before you to make a request please

do not limit my choices and options related to therapies and medications that my fellow esophageal and gastric cancer patients and I have access to I

am proof that stage four esophageal cancer patients can and should be provided with therapies that ensure they're able to live the fullest life possible while none of our stories are

exactly the same we all do wish to overcome the challenges and the trials we are faced with and ultimately say that we survived since my first day of diagnosis I've have had a care team who consists of amazing medical

professionals who have been integral in my care thanks to them I'm here today just this Monday I received my 28th updeep though immunotherapy and fusion

along with my 48th 5 of few and lucor infusions I was disconnected yesterday today I stand before you while Thursdays are generally my most challenging days

each cycle something greater is living within me today to allow me to be here and stand before you despite my challenges during the journey and my low pdl1 threshold I excitedly share with

you that my pet SK SC and circulating DNA tumor markers have shown no evidence of disease since spring of 2023 next slide you can see that I'm living a full life a bit different than

I once pictured but full nonetheless Full Of Love full of Adventure and full of Hope please show compassion in your vote and any future decisions that you make related to the treatment options

for esophageal and gastric cancer patients we all want the best chance of living life and know that means a variety of options to meet all of our unique needs and circumstances please do

not limit the indicators for eligibility or limit the options for treatments I still have a long life to live and to my knowledge no one has yet written a guide for how I should explain this all to my children if my options become limited by

individuals who are not my immediate Medical Care team thank you for your time thank you speaker number five please state your name and any organization you are

representing for the record my name is de Min morai I am the CEO of the esophageal cancer Action Network our organization receives

funding from all of the applicants I am not being paid for my testimony here today or any of my travel costs in 2009 after losing my husband to

esophageal cancer I'm sorry now I'm crying too um I started Ean because I was appalled at how little research and

awareness existed for esophageal cancer the next year 2010 the National Cancer Institute Drew up a list of 20

cancers for its groundbreaking genome mapping project called the cancer genome Atlas or

tcga ES safal cancer wasn't on that list Ean begged the NCI to include esophageal cancer in tcga we even

offered to orchestrate the tissue sample collection and raise a half million dollars to pay for the launch of that project and it

worked the Sagal cancer pilot project of tcga began in 2011 and its findings published in 2017 showed that esophagal adnoc

carcinoma was genomically similar to gaster cancer the result was that our patients were included in clinical trials focused on stomach cancer

including the trials that led to the approval of opo and Kuda for patients with esophagal adinal carcinoma or GE Junction adal carcinoma that approval was just three

years ago and now we're seeing more survivors of stage four esophageal cancer and something that was we could only dream about when my husband was

being treated so I'm here asking you what's going to be gained by making this proposed change initial FDA approvals were based on

sound data they showed promise for our patients and in some cases that was regardless of their pdl1 status we know

patients who are pdl1 negative who are thriving because of their immun therapies take Jim Bennett of Mount

Pleasant South Carolina he's a 77y old Survivor who lost 40 pounds at the time of his stage for a safal adoc carcinoma

diagnosis that was 18 months ago since then he's been on Fall Fox and ABDO and not only has he gained back all of his weight and seen his tumor in Mets shrink

he's now running 5Ks riding his motorcycle and feeling as he describes it as if he doesn't have cancer at all if the FDA decides to restrict his

access to immune checkpoint Inhibitors Jim's Lifeline will be gone and Jim is not the only pdl1 negative patient who's

experiencing these positive responses Dr Kumar repeatedly said no convincing evidence of benefit or harm

for pdl1 negative patients has been found no convincing evidence of benefit or harm when you're looking at possibly

177% of our patients shouldn't that decision be made by the doctor and the patient not an FDA panel especially when we look at the issues with the kind of tissue samples

that we're looking at to come up with these scores and the in the variability in the responses my apologies if you could conclude your statements please yes this is not the time to pull the

plug on this progress I want you to I hope that when you make your decision you will remember Jim Bennett his chances for survival are

very few his doctors should be able to help him make a a good decision he believes that losing access to immunotherapy will cost him his

life and that is too high of a price to pay thank you for the opportunity thank you speaker number six please state your name and any organization you are representing for the

record good morning I'm Betsy Aaron I'm not affiliated or receiving compensation from any organization I'm going to share my story

of delays and restrictions in getting access to immunotherapy during a time of disease progression I'm 70 years old I was diagnosed with stage four gastric

adenocarcinoma in July of 2022 I was told that my treatment would be pallative and that I didn't meet the minimum requirements to receive

immunotherapy I received instead 42 rounds of chemotherapy every other week for two years I was then given a chemo quote

unquote holiday after about 6 weeks I had an endoscopy and learned that the primary tumor had returned I also learned from tests on the fresh tumor

tissue that my PDL score was now 20 my doctor and I discussed treatment options the one I wanted was treatment with two imun therapy drugs I was told

that I would need to obtain Compassionate Care since I did not have the approved biomarkers during this time of waiting

for approval my symptoms continued to increase after waiting five weeks my doctor and I agreed that I had to start

treatment so we start so we opted for a chemotherapy plus one immunotherapy drug this treatment option ALS Al

involved getting approval since third line treatment for anyone over 65 is currently also restricted the that

approval took an additional week after a total of six weeks I received approval for the treatment I wanted in my view

and in my experience access to immunotherapy treatments needs to be made easier for people living with

gastric cancer and not more restrictive thank you for hearing my story and considering my words thank you so much speaker number seven please state

your name in any organization you are representing for the record hi um my name is Ronald capu I'm the husband of speaker number three Allison capu um I

reside in Ringo New Jersey and have access to a lot of doctors but I don't have any affiliation or receiving any compensation from from anyone um I want

to share my story uh regarding my diagnosis uh and treatment of gastric cancer I was diagnosed almost four years ago at the age of 47 I'm married with

two children ages 10 and 12 at a time this was obviously devastating news but I felt confident that I could fight this I wanted to see my children graduate Elementary School and hopefully on to

college my initial diagnosis was stage 1 be stomach cancer in November of 2020 I went ahead and got opinions from my olist and her team I also got second

opinions from many doctors across the entire United States um in in all cases um in all of our

conversations I was told that my gastric cancer is tough and chemor resistant and that getting a clinical trial or including immunotherapy in my treatment

would be the best case for my survival immunotherapy was not yet approved yet for my uh cancer so I moved forward with the standard of care as indicated in the

nccn guidelines I was uh ultimately upstage to Stage 3B uh six months after my diagnosis good news is I continued to remain disease-free for 14 months until

routine EGD in October of 2022 discovered uh a reoccurrence the good news about that is though is there was no tumors and my scans are all clean

um my cancer was just microscopic so what are my treatment options for this reoccurring um getting more chemotherapy for a chemoresistant cancer is is not my best

route my oncologist opinion for my best outcome is to get me on immunotherapy so I a mini battle ensued my oncologist fought the insurance company but I did

not meet the pdl1 requirements and I was past firstline treatment but after a short fight I got good news in December of 20202 almost two years ago now I was

able to get ke truda off label as a second line treatment since then I've experienced very positive results I'm not dealing with any harsh side effects of chemotherapy I'm enjoying a better

quality of life and I'm spending a lot of time with my family um myself and everyone on my team all agree that the imunity therapy has been key in my

current success my oncologist was not just looking at pdl1 score she used her experience with similar patients outcomes my resistance to chemotherapy

the fact that my disease is microscopic topic and I'm in generally good health besides the cancer using stories like mine as well as countless other patients with low

pdl1 score should really be considered in short chemotherapy did not work for me but immunotherapy is and you know my children have went on to graduate Elementary School and both of them are

in high school and I'll be seeing them graduate there soon so thank you for your consideration thank you so much and speaker number eight please state your name and any organization you are

representing for the record speaker number eight yes good morning my name is Pam Hall I'm speaking today as a patient on behalf of myself and others who are

struggling with gastric cancer I've received no compensation for my appearance here today I'm a 68-year-old retired ired banking executive and a devoted yoga

practitioner my husband and I have been married for 31 years we have three children and eight grandchildren six years ago in August of

2018 at the age of 62 I was diagnosed with stage three gastric cancer this diagnosis has forever changed the course of my life and that of my

family the first line of treatment I was given included chemotherapy and it was Then followed by a total gastrectomy since then my cancer has

recurred five times needless to say I've been subjected to every Cancer Treatment Western medicine has to offer this includes participation in two separate

drug trials in all but this last recurrence treatment has worked for me for a short time to eradicate my disease only for it to return time and again

when I was initially diagnosed immunotherapy drugs were not even considered an option for firstline treatment no one understands why some

people respond to certain therapies and others don't likewise no one knows why cancer in some people persistently recurs While others remain cancer free

after only one line of treatment we do know however that cancer is smart and it can morph and change to evade the immune system and render treatments

ineffective my case is a example of this happening after multiple biopsies through the past six years my results

came back this P past May for the very first time as pdl1 positive does this mean that the immunotherapy drugs that didn't work for me in the past will work

for me now I don't think anyone knows the answer to that question what I do know is that I want my doctor to feel

free without reservation to try all the weapons in his or her Arsenal to treat my disease the indications set by the

FDA have an immediate and an outsized impact on What treatments insurance companies will and will not approve frankly I don't have the energy to fight

both this disease and my insurance company who by the way are not doctors I don't have the I I do I don't want to argue with them over whether or

not I should have an immunotherapy drug my ask today is that you consider the first the patient perspective before setting or changing your indications or

guidelines for this class of drugs thank you for your time thank you so much and thank you to all of our open public hearing speakers uh the portion of this meeting has now

concluded and we will no longer take comments from the audience so the committee will turn its attention now to address the task at hand the careful consideration of the data before the committee as well as the

public comments we will now proceed with the questions to the committee and panel discussions I would like to remind public observers that while this meeting is open for public observation public

attendees may not participate except at the specific request of the panel after I read each question we will pause for any questions or comments concerning its wording slide two please

we'll proceed with our first question which is a discussion question in patients with her2 negative micro satellite stable gastric gastrosoph

Junction adenocarcinoma does the cumulative data support the use of pdl1 expression as a pred Ive biomarker When selecting patients for treatment with

pd1 Inhibitors are there any questions or comments uh regarding the wording of the question seeing none uh we will now open the question to discussion and uh you

know please turn your uh name placard sideways uh if you want to make a comment regarding this uh discussion question and I I'll take the opportunity to go first you know I really appreciate

all the data and all the work that's gone on to uh into these presentations I will tell you when treating these patients you know I know our patients don't want incremental benefits in overall survival they want to see that

tale of the curve and and see those durable responses over a prolonged period of time and I think that there's some consistency in all the evidence that we've been presented uh with I

think that you know pdl1 is predictive of uh of response I think we see significant activity at pdl1 greater than 10 I would say we see modest in I

would I use that term you know very seriously I think we see some modest benefit between one and 10 and I see no evidence of benefit uh in pdl1 scores less than less than one I think there

are also some significant challenges here I I see a biomarker that you know is not binary this is you know measured on a gradient and I think that there's no standardization uh as been has been

mentioned I think that there's uh inter uh you know reporter variability which is concerning to me is is a five the same as an 11 it's a 12 before I mean we

I don't think we know the answer to that um and I I think that um you know limiting immunotherapy uh using a somewhat unreliable biomarker is a

little bit concerning But to answer this particular question I will tell you I do believe that pdl1 expression is a predictive biomarker in this disease uh I do see significant activity at uh

levels greater than 10 I do not see any activity in uh scores less than one and I would love to see uh patients have the opportunity to receive these drugs

between a scores of one and 10 but I think that that requires um some discussion between the patient and their provider in terms of the risks because we're asking our patients to undertake

greater grade three and four risks um you know for unclear benefit particularly at lower scores uh Dr

Sprat Dan sprad Case Western yeah I mean the question right is not should we impose OS some restriction in Cut point the question is does the cumulative data

support the use of pedl one expression as a predictive biomarker a predictive biomarker at its core is simply there's a different relative effect size by

biomarker status period end of story there's ways to test this it is a predictive biomarker there's significant interaction Effects by subgroup there's different relative effect sizes every

trial of each of the companies the primary endpoint specifically you know a priority picked these CPS or tap you know thresholds to

be included so it it's acknowledged I don't I hope other than one of the companies it seems to acknowledge there is very significant differences in

relative benefit I think as you just pointed out it's a very challenging when you get from the binary less than one to

greater than one um but a point that I think gets confused a lot is in some of even the amazing open um public hearing comments and touching stories is just

because a patient has an amazing response to chemo and IO doesn't mean they wouldn't have an amazing response to chemo and if has ratio of almost one they probably had a similar probability

of having that benefit so anyways that's all I have to say thank you Dr Madden

yeah I think um you know sorry if you could sorry Robbie Madden National Cancer is too my apologies so I I I think that clearly there is some degree

by which all of this is telling a predictive story but I think the clinical utility and and how the data supports that read between the lines a little bit but the context of the

discussion here is you know I'm just not fully convinced that this is the data set that should be used to to to address this this is hypothesis generating data

that poses the question of is this cut off you know required to to bring about benefit versus risk and again as was

raised by the FDA you know is is the cuto off that's proposed the right cut off we don't know you know we never really went into these studies asking this question and so I think that's

where I have I have some trouble and and that is not sure that this is how we would power this data set we we're trying our best to glean what we can from existing data that was never really

designed to answer this question in my opinion that's kind of where my thoughts come from as a non gii oncologist so I welcome thoughts from

the panel uh in that regard thanks Dr Madden Dr gradishar Bill gr is here Northwestern um I share the sentiments that have been expressed up to this point I think that

this is a predictive biomarker at least I'm you know despite the flaws the caveats that have been discussed about it um certainly for above 10 and one to 10 I think that gets into the realm of

you know letting doctors be doctors where they have an opportunity to talk to their patients and make a discussion in conjunction with their uh patient about whether this is worth doing taking

into account the side effect profile that these drugs have I'm not um in any way um particularly impressed with any of the data that's been presented for

anything less than one and in that group of patients um you know I'm not seeing any uh real effect and I would also Echo

what Dr Sprat just mentioned you know I I I empathize with the compelling stories that were described but we've seen patients in my I'm a breast

oncologist during the era of bone marrow transplant for uh patients with breast cancer you know there was a a lot of enthusiasm for that until there wasn't

and many of those patients might have done just as well with standard therapy and we just don't know so that's my

view thank you Dr grar Dr Von Neil Von um Columbia University um I agree with everything that's been said um I'm I'm a breast oncologist and for me the analogy that I keep coming back

to is actually not the mutation examples that have been brought up like with K and elib um is but it's another continuous variable which is her too and I'm the two things I'm really thinking

about are number one obviously the field optimized and decides th decided thresholds of POS positivity and and these trials large trials were designed to test those questions um but I think

the difference is is that in the her two field the um smaller small subsets were tested and then slightly larger and then slightly larger encompassing that biomarker I think what we have here is

actually the opposite where it got approved in the initial data based on the best available data and we sort of maybe backtracking and refining that biomarker and I think that taken to an

extreme um I I agree that these post talk analyses can of course have biases but you know I'm thinking about other trials

like um uh nsabp 47 where you know actually testing her two antibodies patients who had low levels of her to and so is are we are we thinking about

an extreme possibility like that or are we just making use of the best available data that we have at the time thank you question uh Dr

Hawkins Randy Hawkins Charles University so I I agree with what's been stated th far and particularly Dr Lou's initial

summary um so um pdl1 is helpful but it's not definitive um we've talked about the need to get more tissue to be

able to better get an idea with this particular marker where it exists on this patient do that explain why some people do better uh than others just

because he had more tissue to get included um it appears that we need to work harder if we're going to continue to use

this marker to develop better assays or criteria for getting what is an acceptable uh tissue or assay for pdl1 and of course it means that we need

to continue to search for other markers that may be as be more helpful than pdl1 thank

you thank you Dr Hawkins Dr meart Jeff myart Dan Barber um so I think to this question to me it's it's fairly straightforward the day is clear there's

different levels of dl1 expression are have different levels of overall survival and progression free survival I think the question obviously is is there're cut off and are we going to deny patients uh of a potential for a

therapy if you choose some cut off what is one or or or or another number and the reality is we for good or bad we do that all the time in oncology you know

we don't give gemcitabine for gastric cancer is there a gastric cancer patient who potentially could benefit from gemcitabine I'm sure there is is there's probably multiple but we still have to

use some data to be able to decide who's potentially going to benefit or not and overall as a population is uh is there some benefit is that benefit enough to weigh to the

risk thanks Dr Sprat I don't know if you want sure drat does defer to Dr hilard Dr

hilard okay yeah I mean I do think that uh you know just looking at the data there is predictive value that if you do have like um higher pdl1 expression

you're more likely to benefit but on the other hand um most of the patients in the studies even with the pdl1 less than one have um

on the average had some benefit even though it's not statistically significant and so uh yes I mean cumulative data suggests it's predict a predictive

biomarker but at this point I don't think it's um clinically uh something that uh you know will outweigh all the other factors that might go into the

clinical decision thanks Dr Hillard uh Dr sof um yeah one thing oh sorry Hannah

San off UNC um I was sort of curious to hear the panel's thoughts on the discussion about tissue

inadequacy and um availability of biopsy samples so this is true a cross oncology at this point right we use biomarker testing for every single disease um is

there something unique to the group about gastric and esophageal cancers that would preclude us from RE biopsying someone um you know I think we heard

from Dr Jen jigan that patients respond the best in the first few cycles of treatment I completely agree with that but that's chemotherapy response that's not IO response and gastrosoph agal

cancer which is different than say you know melanoma right um so to me that did not strike me as something we should use as a deciding factor here because to me

I feel like we could you know biopsy is is fairly readily available but I'm curious to see if that sways other panel members at all uh does anybody have a response to

Dr sof's question I just say speaker number seven who spoke can you sayate your name sorry oh Dan Sprat um and I am not a GI medical oncologist but speaker number seven uh I

think spoke very eloquently that was denied and received Kemo first and then later on um received IO and he's doing very well right now so I guess to your

point it seems that the necessity of this being immediate at least in his case we're talking about anecdotes right now but I defer I think someone else had their hand

up Dr Gibson um Michael Gibson I think I'm in the the appropriate session to comment um so sorry guys uh I just wanted to say that uh one of our

speakers it may have been Dr janjigian mentioned that this is a dynamic biomarker which means as was pointed out by one of our speakers it

changes over time I don't think getting another biopsy although I'm not the patient I haven't had patients um not uh agreed to do that if we have an

justification such as retesting for a marker that may have been negative at the first time so I do think this is an appropriate biomarker it's Dynamic and

the question to whether biopsy again I think that's an important consideration that is practically possible thank you Dr Gibson uh Dr Van Lon I see that you had raised your hand I didn't know if you wanted to respond

to Dr s's question as well or had a separate comment um I think I was responding maybe to one speaker earlier and you know from the perspective of a gastrointestinal oncologist I also

wanted to reference um the breast oncologist who had mentioned her too as a biomarker and remind everybody that we use different cut points in different

diseases for different biomarkers so we use her two thresholds differently in upper gastrointestinal cancers than we do

in um breast cancer and I think that's a reference to the fact that we're learning as we go um and unfortunately

we're dealing with a assay that has limitations with pdl1 um but based upon the current preponderance of evidence it

certainly seems to be a a predictive biomarker for this particular disease but I I think we all have to acknowledge that we are still learning about it and

um there is certainly a demand to address the limitations of the biomarker testing um for future decision- making but sitting with the data that we

currently have um is really important thank you Dr vanloon uh Dr Sprat yeah two two things one is and and

I commend uh the applicant for providing that Q twist data which really right is is probably the one method of analyzing this quality of life toxicity or you

know freedom from and then freedom from progression or death and nice harmonize not I mean it's imperfect but I appreciate them putting it in is that you know they showed nicely

that to what I think one of the panel members said is that when you get to these scores less than one I mean essentially not that there's necessarily a uniform agreed upon clinical

significance threshold they site 10 from a old deer sometimes might it might be appropriate to be less than 10 um but it's very clearly different I mean it

was about four to five% versus like over 30% for expression levels over 30 so I think this is when someone said like it is there potential harm of these agents

I mean yes if there's no potential harm of these agents then yes just make it available um ignoring cost in and of itself as a toxicity if you factored in financial toxicity into that Q twist

analysis I think we'd find something uh strikingly different given you know the combined of just Neo and pemro is over $30 billion a year I think for 2024 so who's paying for that patients are

paying a percentage of that out of pocket even if it's not the majority so I I I think we just need to be thoughtful to the potential harms the

point that Dr jigan brought up which was spot on and I think you were just trying to address is the real world aspect of this without tissue I guess what I don't know and I love for the panel if someone

can answer is what is the real world efficacy data in this patient subset that's not enriched for these High um pd1

scores because again you can't talk about trials and the accuracy but then not talk about real world efficacy are these patients that are going to have far poorer response rates because

they're not as enriched and I don't know if anyone knows that uh any responses to Dr sprat's question this is Chris Lou from Colorado and I'm I'm not sure that we necessarily

have all you know that real world evidence data uh and I think that it just kind of speaks to the reality of the problem that right we live in a non-clinical trial world and we're going to have biopsies that are not going to

be able to get a score on uh and that really goes back to Dr San's Point uh and that others have made about rebiopsy uh and I think that uh in reality that's what our patients are you

know may have to undergo if this decision is made to start um you know cutting off at particular CPS scores uh Dr

Madden Robbie Madden NCI so again I'm just kind of stuck a little bit here I mean bleeding into the question a little bit and the CPS score of 1.0 like why is

1.0 the cut off you know is it 08 is it 1.2 you know if we're going with the harms thing maybe we're harming everybody as 1.1 maybe all the the

responders sub one are at 08 and above I I you know this is this is where I struggle with saying that we have enough data you know at least what when it comes to the voting question to to

assign a cut off thank you Dr Madden uh unfortunately uh only if the sponsors are uh directly asked the question uh can they come up uh any other question or any other comments uh

from the panel in regards to this discussion question okay I'll do my best to summarize this discussion so I I you know hearing everybody on the panel I I feel like

there's some consistency right in in kind of thinking that pdl1 expression is a predictive biomarker for immunotherapy right and I think that that's really what the the discuss discussion question

is asking uh I think that there are significant concerns from the panel in regards to the efficacy that we're seeing uh in pdl1 scores that are less than one um and I think that there are

concerns about the overall survival data that we see um I you know to to use Dr Von's point and Dr Van lon's point of refining the population patients that are most likely to benefit from these

therapies as well as kind of learning as we go uh there's some practical issues here about the assay itself about standardization about measuring it about the ability to do this outside of tertiary Care Centers and and major you

know molecular companies that do this type of testing uh the the real life situation of having to rebiopsy patients to determine a CPS score and what cut

offs could mean there's also um Dr Madden had made a point that this may not be the best data set to answer some of these questions about cut offs given that we're really starting to cut up the data into incredibly small subsets and

trying to make treatment decisions based off of those small subsets uh in trials that weren't designed to ask the questions that we're trying to ask less than one one to five 5 to 10 these aren't trials that were designed to do

that but luckily we do have a significant amount of data um any questions or comments in uh regarding question one the discussion

question okay we will now proceed to question two which is a voting question uh we will be using an electronic voting system for this meeting uh once we begin the vote the uh buttons will start

flashing and will continue to flash even after you have entered your vote please press the button firmly that corresponds to your vote if you are unsure of your Vote or you wish to change your vote you may press the corresponding button until

the vote is closed after everyone has completed their vote the vote will be locked in the vote will then be displayed on the screen the DFO will read the vote from the screen into the record next we will go around the room

and each individual who voted will state their name and vote into the record you can also State the reason why you voted as you did if you want to you will uh we will continue in the same manner until

all questions have been answered or discussed um can we have uh the voting question up so the voting question is is a risk benefit assessment favorable for

the use of PD Inhibitors in first line Advance her 2 negative micro satellite stable gastric GJ adenocarcinoma in

patients with pdl1 expression less than one are there any issues or questions in regards to the voting question Dr

Madden Robbie Madden and TI um so on our slide actually has the options for answers is yes and no but is there insain option is there traditionally there is an abstain option so you can

abstain I was asking for a friend just [Laughter] not uh yes uh you can you can vote to abstain any other questions or

comments if there are no further questions or comments concerning the wording of the question we will now begin the voting process please press the button on your microphone that corresponds to your vote you will have approximately 20 seconds to vote please

press the button firmly after you have made your decision your selection the light may continue to flash if you are unsure of your Vote or you wish to change your vote please press the corresponding button again before the vote is

closed e there are two yeses 10 Nos and one abstain now that the vote is complete we'll go around the table and have everyone who voted State their name vote and if you want to you can State the

reason why you voted as you did into the record uh I believe we'll start with Dr Van Lon my vote was no um based upon the if you could state your name sorry oh sorry

this is Katherine vanloon and my vote was no based upon the preponderance of evidence at this time I think the risk benefit ratio is not

favorable thank you Dr gradishar U Bill gradisher my vote was no and as I outlined a few M moments ago for those

reasons thank you Dr sprout Dan sprad Case Western my vote was no and um again the voting questions not for us to decide to change of the comp point but just the risk benefit ratio

favor use of PDL one um in this decision-making process I think that this is when we look at credibility of subgroup analysis

this was part of most primary endpoint analysis was measured a priority significant interaction effects pretty much it was a priority the hypothesis and direction of effect was

correct I I think this is a very good data set so just disagree and I think with Hazard ratios almost approaching one um the other point I just want to

bring up that if let their doctor decide Dr Den jigan who's a world expert she said the average doctor sees five of these a year so I'm just not sure we

want to let their doctor um make this decision when these has ratio Z one and there's uh financial and toxicity impacts for these patients when you look

the last point I'll make is when you look at the tales of where they converge there's less than a 1% absolute difference in this less than one subgroup that's a number needed to treat

over a 100 if not close to a thousand that means you're treating hundreds of these patients to benefit one so thank you thank you Dr Sprat Dr

Madden yeah Robbie Madden National Cancer Institute um so I I voted to abstain I think uh our quest for

biomarkers um has been going on since our quest to develop better Therapeutics and I think unfortunately most biomarkers fall short I think

pdl1 has also fallen short in many many diseases including this one because of issues with acquisition uh you know characterization

variability and sampling error so when it comes to that context it's hard for me to say that this is the data set to make this decision um again I'm not sure

that it should be higher or lower I'm just not sure this is how I would ask the question thank you Dr Madden this is Chris Lou from University of Colorado I voted no uh as I stated before I just

don't see any overall survival benefit uh in this group less than one um and I would love to hear uh others on the panel in terms of where they believe the cut off should be I do think that the

cut off should be one uh just because of the perceived benefit that I see uh in that in that patient population between 1 to 10 um and I do think that that is a conversation as been mentioned before

that needs to happen between a patient and their physician but to give them the opportunity to have that conversation I think is really critical Dr

vasson Neil Von I voted no um I agree based on the totality of the data that um uh there was a favorable risk benefit for this pdl1 low population um I will

say in to address Dr Lou's point you know for me it was clear that there was uh across these data sets just a a clear benefit in the greater than 10 uh no

benefit in L less than one and it's that intermediate range that this is where we need clinical trials to help answer questions where we have levels of equipo this is just with any continuous variable the the important questions in

the field thank you thank you Dr DOD yeah Lori DOD um I voted no um because of the preponderance of evidence

presented with a meta analysis in those who were PDL uh one less than one the question was I think very carefully worded to say those that were less than

one um because we don't have enough data for those who we don't have a result from as well as those who are between one and

10 thank you Dr Hillard yeah James Hillard um patient uh I voted yes just in that there's it's clear that

there's some variability in terms of how this is assessed in different settings that clearly um having a high

pdl1 uh liend measurement is associated with um greater efficacy um I don't think there's clear

evidence for the null hypothesis that there's no chance that the uh less than one is going to be

valuable thank you Dr Hiller Dr Hawkins yeah so difficult question I voted um yes um there some resolation I think there was enough

responders who are less than one to make me say it's possible um I felt that the side effect

profile was good when once you got past chemotherapy but one thing I would emphasize would be the really importance of educating GI Specialists uh G oncologists and those that are in

private practice because they're the ones that see the patients first I believe and really need to emphasize the importance of tissue size we need enough tissue for this imperfect uh assay and

we need to work on this assay but we also need to look really hard for additional markers that may help us do a better job at this group of patients

thank you thank you Dr Hopkins Dr Gibson uh thank you Michael Gibson I would start out by saying this is a bit of a wrenching uh question for me uh I

made my decision objectively on the data that I saw today and uh I have reviewed before however I might add I am also a clinician I do appreciate uh the

considerations from the group thank you Dr Gibson Dr mcken hiding mcken my vote was no based

on the uh Hazard ratios for overall survival pd1 or CPS pd1 less than one just want to comment though as a community oncologist I too saw 30

patients a day earlier this week but that meant 15 different cancers and so it is often overwhelming for Community oncologist to keep all of this straight

and so some um great effort from FDA nccn to put in guidance does help the community oncologist

thank you Dr meart uh Jeff meart I voted no um so in addition to the comments that I think it's telling that multiple guidelines and CCA and ASCO and others

have all actually chosen a cut off despite a broad indication right now so while the F should have an independent decision on this I think multiple experts including some that have have

spoken today sit on those guidelines and and uh those are the agreement of of that there should be some cut off in terms of your question regarding should it be less than one or something higher

I think the one concern I have with the 1 to 10 patients is when you look at the pisma breakdown data the 5 to 10 who actually also had a hazard ratio of like

0.92 and then the one to five so there's clearly some variability there but I think the testimony where there was more confidence and less than one being truly

negative uh is is uh helpful thank you Dr sof all right Hannah F off um I also voted no um and as the last person probably a little bit

repetitive I think a couple pieces of evidence are really important here first as Dr Sprat explained this is really high quality evidence right we had a

priori cut points we have repeatedly demonstrated right um evidence here um I think one thing that's really difficult here is this question of are there

people in this less than one subgroup who do benefit right what do we make of these responders um and I think there may be people who respond but we're not seeing that tail

of the curve right I think that's really important the question of can we provide people hope offering them long-term survival with a g b gastric cancer who

have pdl1 less than one to me that really looks like the answer is no now it may evolve over time for those patients which may mean repeat biopsy and subsequent availability of these

drugs is important but that's not what was asked here the other thing is and even though I cannot even tell you how moving it is to hear everyone come up and speak the

folks we don't have at the microphone or the folks who have passed away from getting pd1 Inhibitors and everyone around this table has probably seen one of those patients these are not just

grade three and four toxicities these are Al also grade five fatal toxicities and that is very moving to me when you look at these curves that do

not show long-term survivors from these drugs in a pdl1 negative population so I really hope we can see how this evolves

and how we can get immunotherapies to be effective in this pdl1 negative population but until we do that I just did not see enough evidence that we're

helping people and not harming them thank you Dr s so to summarize uh a majority of this panel did vote no I

think those that voted no spoke to uh the really essentially negative data that we see in the CPS or pdl1 less than one uh cohort um that that cut off

appeared to be at least reasonable there's some uh variability in terms of where people believe that that cut off should lie I think the greater than 10 is pretty obvious and then the 1 to 10 really has a decent amount of

variability in terms of overall survival benefits so there's some concern there uh as well and then the to the point that you know the guideline committees have also

uh instituted uh these cut offs as well for those that voted yes or abstained uh I think there's a really understandable concern about missing patients that may truly get benefit from these agents and I think that we heard from the open

public hearing speakers uh how meaningful it has been to them as well as well as some concerns in regards to the data sets that we have available uh that we're trying to answer questions that those trials weren't necessarily

designed to answer uh but overall there is fairly uh good consistency uh across the vote for the panel um I do uh want to say um thank you so

much to our our applicants BMS Merk Beijing the FDA the incredible amount of work that's gone on to producing uh wonderful presentations a wonderful summary of all the data that's available

as well as to our open open public hearing speakers who really their stories have been truly moving uh and thank you so much for adding to our meeting uh before we adjourn the morning session are there any last comments from

the FDA Dr pasor I I it's a rare opport opportunity that I get three drug companies in front of me simultaneously okay so question number

one okay uh when these drugs were being developed okay we spent a great deal of time in having conferences trying to coordinate with the drug company's

uniform marker development pd1 drug development marker obviously those efforts failed could you address this each one of the companies and express your willingness as we go forward in the

field of imun technology really to harmonize efforts between companies or among amongst companies to have standard

pd1 or whatever biomarker development so Merc since you have the leading drug here what is your position on standardization as we move

forward thank you I will actually have Dr Scott Pro respond to this question uh Scott puit Merc translation oncology we'd be very interested in

working to try to to see if we can make these assays interchangeable uh it would be great if they were interchangeable but I think the data to date suggests that they're actually not we would have

to do cut Point mapping studies analytical and bridging studies which we may or may not have sufficient this boat has sails so to speak what I'm talking about is our ship has s but I'm talking

about as we move forward because there will be further developments in biomarker development in this area obviously but this is not the end of the story The pd1 essays that we have today

well absolutely we we we always try we would try to you're on record you'll collaborate with anybody okay uh let's hear from BMS on this are you going to

collaborate with everybody put away your own commercial concerns here and like come to a Kumbaya with everybody that's developing a similar type of drug we do

welcome efforts for harmonization I think our goal here is to simplify the process for patients and Physicians so the process by which we do that is up for discussion I think we've learned from this experience this is not this

has been U not a great experience obviously having all of these different tests here uh and here again I want to emphasize we did bring people together we made a concerted effort the FDA and

trying to harmonize these test uh with several conferences and telephone calls with friends of cancer research and other external organizations so you're on board right

okay bying as I mentioned in my presentation yes by absolutely supportive of harmonization okay so here again this ship has sailed I don't think we could

do anything more about this but as we move forward I think and looking at new buyer markers we really have to develop platforms across the the commercial

concerns of the companies okay second question okay if if and I underline if we restrict the labels to less than one

uh you are concerned Merc that some patients who may potentially benefit will not receive this drug would you be willing to offer the drug

on an expanded commercial uh I mean expanded use program or a compassionate use program for those people that are less than one free of

charge Merc has um Merc understands the financial toxicities of patients um with these uh with these diseases and yes we

do uh already have programs in place for for patients with financial hardships and we actually also provide drug free of charge for patients who have because

here again if if it's not an approved indication obviously insurance companies not may not approve it so would you have an expanded use protocol for example so-called compassionate use protocol

providing the drug free of charge for these individuals we would um we would have um we we have provided drug free of charge to eligible individuals who

cannot financially pay so you would consider a a expanded use protocol in this situation we'll have to take it back and and think about it okay we'll

be in contact with you how about Ving um very seldom do I have this opportunity that's why I want to make full use of it I guess I would say I I don't exactly understand the context of

the question the if the drug is not be if the drug will not be reimbursed indication is lab the committee just voted the benefit risk is not favorable I know but here again so you would not

but here again other companies have stated that there might be people because of the ambiguities of this ass certainly we have a expanded Access

program that's available globally uh and we would those requests actually come to me so certainly I'd be happy to review if a physician felt that a patient would benefit okay Bristol Myers I just want

to get this on because there's other avenues for use of the drug or access to these drugs yeah so if a physician and their patient deem that there may be benefit we would look for mechanisms by

which we could help them achieve access so you would consider that yeah there's a lot of steps that need to be discussed I I do want to address the concerns of patients we realize the issues here with

the biopsy Etc and here again if we do restrict it and if somebody wants the drug it probably would not be paid for uh so here again we want to have make

make our views that we're patient Centric here that there might be other avenues that patients may have access to this drug okay thank you for the

opportunity thank you Dr Pastor um we will now adjourn the morning session and break for lunch we will convene at 1:15 p.m. eastern time that's 1:15 p.m.

1:15 p.m. eastern time that's 1:15 p.m.

eastern time panel members please remember that there will be no chatting or discussion of the meeting topics during the break amongst yourselves or with any member of the audience additionally for the panel you should

plan to reconvene at 10:5 p.m. eastern

time to ensure you're seated before we reconvene at 1:15 thank you e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e e e good afternoon and welcome I would first like to remind everyone to please mute your line when you're not speaking also

a reminder to everyone to please silence your cell phones smartphones and any other devices if you have not already done so slide two please slide

two for media and press the FDA press contact is Lauren J McCarthy her email is currently displayed slide three please my name is Dr Christopher L and I'll be chairing this meeting I will now

call the afternoon session of the September 26 2024 oncologic drugs advisory committee meeting to order we'll start by going around the table and introducing ourselves by stating our names and affiliations we will start

with the FDA to my left and go around the table uh Richard Pastor director oncology center of excellence Paul clut deputy director oncology center of excellence Sten

memory director do3 kasak division team leader do3 G clinical reviewer

do3 Jo statistical reviewer division of batric 5 Dr vanin we can come back to Dr Van Dr gradishar Bill gradisher Northwestern

Unity City Dan sprad uh Sidman case Weston Reserve University Dr mad Robbie Madden medical oncologist

National Cancer Institute Chris Lou GI medical oncologist University of Colorado Joyce ver pong designated Federal Officer FDA Neil vson Columbia

University Lorie DOD clinical trials research and statistics Branch nyad Dana Dayton patient representative Randy Hawkins Internal

Medicine P medicine Charles University consumer rep uh Michael Gibson air Digestive and GI oncology Vanderbilt Ingram Cancer

Center Heidi mck Community Medical oncologist Vera Cancer Institute sou Fall South Dakota Jeff meart GI medical oncologist Dana Farber

Boston Hannah sanof J medical oncologist University of North Carolina and Dr Van Lon hi I'm Katherine vanloon I'm a

gastrointestinal oncologist and professor of medicine at UCSF thank you for topics such as those being discussed at this meeting there are often a variety of opinions some of which are quite strongly held our goal

is that this meeting will be a fair and open forum for discussion of these issues and that individuals can express their views without interruption thus as a gentle reminder individuals will be allowed to speak into the record only if

recognized by the chairperson we look forward to a productive meeting in the spirit of the federal advisory committee Act and the government in the sunshine act we ask that the advisory committee members take care that their conversations about the topic at hand

take place in the open for of the meeting we are aware that members of the media are anxious to speak with the FDA about these proceedings however FDA will refrain from discussing the details of

this meeting with the media until its conclusion also the committee is reminded to please refrain from discussing the meeting topic during breaks thank you Dr frong will read the

conflict of interest statement for the meeting the Food and Drug Administration is convening today's meeting of the oncologic drugs advisory committee under authority of the federal advisory

committee Act of 1972 all members and temporary voting members of the committee are special government employees sges or regular federal employees from other agencies and are subject to Federal conflict of

interest laws and regulations the following information on the status of this committee's compliance with Federal ethics in conflict of interest laws covered by but

not limited to those found at 18 USC section 208 is being provided to participants in today's meeting and to the public FDA has determined that members and temporary voting members of

this committee are in compliance with Federal ethics and conflict of interest laws under 18 USC section 208 Congress has authorized FDA to Grant waivers to special government employees and regular

federal employees who have potential Financial conflicts when it's determined that the agency's need for a special government employee services outweighs their potential Financial conflict of interest or when the interest of a

regular federal employee is not so substantial as to be deemed likely to affect the Integrity of the services which the government may expect from the employee related to the discussion of

today's meeting members and temporary voting members of this committee have been screened for potential Financial conflicts of interests of their own as well as those imputed to them including Villers of their spouses or minor children

and for purposes of 18 USC section 208 their employers these interests may include Investments Consulting expert witness testimony contracts grants cdas

teaching speaking writing patents and royalties and primary employment today's agenda involves a discussion on the use of immune checkpoint Inhibitors in patients with

metastatic or unresectable esophagal Sous Sal carcinoma the current labeling for Approved checkpoint Inhibitors in this indication reflects broad approvals in the intent to treat population

agnostic of pdl1 expression cumino data has shown that pdl1 expression appears to be predictive by a marker of treatment efficacy in this patient population however

trials however clinical trials have used different approaches to assess pdl1 expression and different thresholds to Define pdl1 positivity FDA would like the

committee's opinion on the following adequacy of pdl1 expression as a predictive biomaker for patient selection in this patient population differing risk benefit assessments and

different subpopulations defined by ped1 expression and adequacy of cumulative data to restrict the approvals of immune checkpoint Inhibitors based on pdl1

expression the committee will discuss the existing sbas which were approved for patients with previously untreated unresectable or metastatic esophagal

sisol carcinoma spa1 125 514 s-96 for katuda from bralab

injection submitted by Merk sharp and doome LLC a subsidary of Merc and Merk and Company Incorporated Spa

12554 s-105 and s106 for pedo NOAB injections submitted by Bristol Myers SB

company and S1 125 377 s- 122 for youro epip injection submitted by Bristol Meers squib company the committee will also discuss

new bla 761 380 for Tisa maab submitted by Beijing USA Incorporated for the same proposed indication this is a particular matters meeting with specific matters

related to Bristol Myers squibs Spa's MC's s and Bean's NDA will be discussed based on the agenda for today's meeting and all Financial interests reported by the committee

members and temporary voting members no conflict of interest waivers have been issued in connection with this meeting to ensure transparency we encourage all standing committee members and temporary voting members to disclose

any public statements that they have made concerning the product at issue we would like to remind members and temporary voting members that if the discussion involve any other product or firms not already on the agenda for

which an FDA participant has a personal imputed financial interest the participants need to exclude themselves from such involvement and their exclusion will be noted for the record FDA encourages all other

participants to advise the committee of any Financial relationships that they may have have that they may have with the firm at issue thank you thank you Dr frong we will now proceed with FDA

introductory remarks starting with Dr Sandra kazak good afternoon and welcome back my name is Sandra kasak and I am a team leader

in the division of oncology 3 at dfda I will provide a brief introduction to the afternoon session similar to this morning session we will discuss the predictive value of pdl1 tumor

expression and the potential for optimization of treatment using pd1 inhibitors for patients with esophagal SEL carcinoma the first line trials sorry the first

line Trials of immune checkpoint Inhibitors in combination with chemotherapy that will be discussed today have demonstrated overall survival benefit for patients with metastatic or

unresectable esophagal sosal carcinoma the FDA approvals of pism app based on keynote 590 an evolab based on

Checkmate 648 in combination with chemotherapy or epilim are agnostic of pdl1 expression status similarly the

trial of TS lisima currently under review Ral 306 has shown a survival advantage in patients treated with disisa in combination with

chemotherapy later today the applicants will summarize the design of each study and Dr Shasta will highlight similarities and differences between

them for the purposes of this discussion fda's analysis will be centered only on the comparison of pd1 Inhibitors as addons to chemotherapy versus

chemotherapy in patients with esophagal squamous cell carcinoma in other words we will not discuss the data in patients with esophagal adenocarcinoma or the

results of a comparison of neolab in combination with epilim versus chemotherapy alone in all three trials overall survival results were statistically

significant in all prpf subgroup analysis of pdl1 cut offs as shown in this table the hazard ratios for the comparison of chemo therapy in

combination with pd1 monocan antibodies in the it population were 0.73 0.74 and 0.66 in the pisum neolab and

tum trials respectively please note that there's an error in the slide and the results in the Bottom Road reflect all patients with the SAG SC cell carcinoma similarly

to the it the overall Hazard ratios for pd1 monoclonal antibodies were 0.72

0.73 and 0.68 in the pemis of neol andisa Trials respectively this table summarizes the results of the trials based on PDL one

cut offs each trial use different diagnostic test methodology to assess p dl1 and different cut offs for the pre-specified statistical analysis which

are highlighted in colors yellow for keynote 590 blue for checkmate 648 and green for rational 306 the analysis of Checkmate 648

presented in this slide is based on CPS pdl1 scoring instead of TPS which was the original pdl1 scoring algorithm used in the trial as this information is now

available although the overall survival results were statistically significant for the anti pd1 containing arm in all three trials highlighted in the red box in

subgroup analysis the point estimates for the treatment effects appear marginally or not favorable in patients with pdl1 less than one tumors and

intermediate in patients with pdl1 less than 10 tumors as shown in the right column although these results are exploratory an uncertainty exists for

each trial the data does not appear to support the use of anti pd1 drugs in patients with pdl1 less than one tumors and benefit appears to be of a higher

magnitude in patients with PL 1110 or higher expressing tumors FDA granted approvals for pisola regardless of pdl1 status

reflecting the it patient populations however results of the pre-specified cut offs as well as exploratory analysis of additional pdl1

cut offs as shown in this table were included in FDA product label to provide data on differential efficacy seen in patients with lower pdl1 expression to

in treatment is decision making although both keynote 590 and Checkmate 648 were positive studies in the overall population professional

guideline recommendations for the first treatment for patients with unresectable or metastatic esophagal musel carcinoma are generally based on subgroup analysis

of the pdl1 cut offs of each individual study as you can see in the slide the guidelines recommendations may result an inconsistent approach regarding who

under goes testing and which drug might be used at a given period one cut off the ASCO and NCC and guidelines recommend uh recommendations were based

on pdl1 scoring algorithm and statistical designs using the individual studies Al as can be seen in the right upper corner ASCO recommendations for

niola is also accommodating to the use of a different scoring algorithm than one used in the clinical study supporting that particular recommendation none of these

recommendations specifically describe or require the use of the individual pdl1 test used in each one of these trials as I have previously mentioned FDA

labels do not restrict indication for pralis or noolab and currently include information on the efficacy of both drugs for pdl1 status in the product

label so why discuss pdl1 in the ESOP musel carcinoma population now although subgroup analysis in single trials can be misleading we now have

results across three trials suggesting lack of efficy in pdl1 negative or less than one patients to summarize Improvement in survival with the

addition of a checkpoint inhibitor was greatest in patients with higher pdl1 expression 10 or more in all three

trials although sample sizes were limited the point estimates for treatment effects one for pism 0.93 for neolab and 1.34

forab did not appear consistent with a beneficial effect of immune checkpoint individuals in patients with tumors that were pdl1 less than one based on exploratory analysis of

each trial pointing to the lack of clinically meaningful benefiting patients with pdl1 low the issue whether pdl1 testing is needed to select patients for immune checkpoint inhibitor

therapy for treatment of esophagal or Gast esophagal cancers have been extensively debated in a published met analysis by Dr Yun and colleagues of randomized clinical trials including

gastric and esophagal carcinomas that was conducted to evaluate overall survival benefit from immune checkpoint Inhibitors based on

High versus absent or low P1 expression 5,067 patients with esophagal s musel carcinoma were included the metanalysis was based on

published trial level data and per the report among patients with a Sagal squ SEL carcinoma across all lines beel one

tumor proportion score or TPS was the strongest predictor of immune checkpoint Inhibitors benefit and TPS hi was defined as TPS of one or greater except

in one trial that you said 10 cut off in the TPS hike subgroup the overall survival has ratio was

0.60 while in the TPS nonik sub group The ratio was 0.84 the second strongest predictor of benefit of treatment with immune

checkpoint Inhibitors was CPS High defined as CPS of 10 or higher in all trials except one trial which Ed a one

cut off in the CPS hik sub group the overall survival Hazard ratio was 0.62 while in the CPS nonh High the

Survivor Hazard ratio was 0.82 while FDA did not independently review this study results are consistent with pdl1 status being predictive of benefit

in our patient level evaluation of three pivotal trials although typically drugs approved by the FDA are indicated for use in the total patient population studied when

there are consistent treatment effects across important studies subgroups consideration should be given to rise indication to better informed use of a

drug as Dr Lem previously presented when considering subgroup analysis replication of results across multiple trials sample ascertainment biological

possibility and study design considerations including stratification and pre-specification of analysis are all factors important to the strength of the

evidence in each of the three Trials of anti pd1 Inhibitors there was no prespecification for the pdl1 low groups although for each pdl1 cut off subgroups

the analysis of the studies were underpowered we now have the results of three studies with generally consistent effects in addition to the data provided

in drun CS met analysis when talking about lack of effects or uncertainty regarding treatment efficacy safety should also be

carefully considered exposure to treatment May potentially result in life altering toxicity which may be a risk patients are willing to take when a benefit is expected the table at the

left is a summary of the incidence of Select immune mediated Adverse Events across four trials with single agent fisim and Evola and though many of these

events are treatable immune mediated Adverse Events can become chronic these Adverse Events or even the steroids used to treat them may compromise the

patients's quality of life which is fundamental to patients in addition although infrequent there are deaths related to immune reactions associated

with the use of immune checkpoint Inhibitors in short PD pd1 monocl antibody treatments have toxicity and

benefit to Patient relies on efficacy at waking that risk in pedal one low populations efficacy has come into question and with it whether a favorable

risk benefit remaining for those patients with tumors that are pdl1 less than one the table on the left is a snapshot

of data for each trial results in theop agal SEL carcinoma patients are displayed in the second column and the pedal one less than one and less than 10

subgroups are displaying in the third and fourth columns later today Dr Shasta will present an extensive review including fda's exploratory pool

analysis data to provide additional context regarding the results of each trial kapan maor curves median overall survivals and Hazard ratios in the pdl1

low populations highlighted in the red boxes appear to show marginal benefit across the class or even potential detriment one cannot ascertain whether

any minor differences are related to sample size testing methodology or chance again it is important to consider that although minority of patients will

develop severe or life-threatening toxicity benefit to patients relies on efficacy outweighing that risk imp perial one low populations efficacy has

come into question and with it whether a favorable risk benefit remains for those patients with tumors that are pdl1 less than one going back to subgroup analysis

considerations based on the data just presented it appears that differential EIC ay based on pdl1 status is replicated across independent clinical

trials this repetition was seen both in the first line of agal SEL carcinoma trials in combination with chemotherapy as well in addition in additional trials

when assessed head-to-head against either chemotherapy or placeo as described by Dr Yun and colleagues some pleaser tment for pdl1

expression was above 90% in the pisma vanola trials and 84% in the talisma trials with results available from the vast majority of

patients with respect to biological possibility although pdl1 has variable utility as a biomarker in different tumor types it does appear to be useful

in select disease settings finally a limitation of these findings is that all three studies use different pdl1 testing methodology pdl1

High populations were selected at two different thresholds and none of the studies were specifically designed to

test for pdl1 negative or low subgroups in summary the current US FDA approvals of immune checkpoint Inhibitors in combination with

chemotherapy for the first line treatment of esophagal s Cel carcinoma are agnostic of pdl1 expression status

however three independent trials and FDA exploratory level pool um patient level pool analysis as well as the published trial level metanalysis support a

predictive role of pdl1 expression for treatment efficacy patients with tumors with PDL 1110 or higher appear to have the greatest magnitude of benefits

patients with tumors with intermediate P1 expression between 1 and 10 have lesser magnitude of benefit patients with pdl1 negative disease although

there may be some residual uncertainty based on small numbers of patients irrespective of the assay used appears to uh have no evidence of benefits and

patients may actually be at risk for harm selection of a pedial one cut off of one will results in approximately 90% of patients being legible for checkpoint

Inhibitors and would allowed a consistent approach to treatment in the clinic following all the presentations we would like the committee to discuss the risk

and benefit of the use of anti pd1 antibodies for the first line treatment of patients with metastatic or unresectable esophagal musel caroma with

pdl1 status less than one following the discussion we would like the committee to vote on the risk benefit assessment for the use of anti

pd1 antibodies in first line and seable or metastatic esophagal carcinoma with a pdl1 expression less than one thank

you thank you Dr kasac both the Food and Drug Administration and the public believe in a transparent process for information gathering and decision making to ensure such transparency at the advisory committee meeting FDA

believes that it is important to understand the context of an individual's presentation for this reason FDA encourages all participants including industry's non-employee presenters to advise the committee of any Financial relationships that they

may have with the industry such as Consulting fees travel expenses honoraria and interest in a sponsor including Equity interest and those based upon the outcome of the meeting likewise FDA encourages you at the

beginning of your presentation to advise the committee if you do not have such Financial relationships if you choose not to address this issue of financial relationships at the beginning of your presentation it will not preclude you from

speaking we will now proceed with our first presentation from Merc good afternoon I am Kathy Panza vice president of clinical research in late

stage oncology I'm a medical oncologist and prior to joining MK I was an attending physician at Memorial phone ketan Cancer Center we will share evidence supporting the positive benefit

risk of Kuda in patients with esophageal cancer which comprise both esophageal suus cell carcinoma and adenocarcinoma these may be referred to

as esophageal cancer in this presentation this morning we presented the biological rationale for combining pisab and chemotherapy as well as the

methodology for testing and validating pdl1 CPS cff points used in merc's pivotal trials these pertain to our Sagal trials as well I will describe

huda's meaningful place in the treatment of aof cancer Dr Pua bagya will then share data from keynote 590 finally Dr Peter enzinger will provide

his clinical perspective metastatic esophagal cancer is a rare disease with patients having a poor

prognosis surviving only five 5% surviving 5 years before immunotherapy the only treatment for firstline metastatic disease was

chemotherapy this malignancy has no biomarkers or targetable molecular aberration and as such we Face a dir of therapeutic options rigorous study

design and conduct gives confidence in the positive results of keynote 590 which met success criteria for all primary and key secondary endpoints in

the intention to treat population the Kuda label includes information about pdl1 subgroups empowering Physicians to work with patients to make the best

choice for therapy the indication for Kuda and sopal cancer should be retained based on the safety and efficacy data in this patient

population kyote 590 was the first Global phase 3 study to assess pd1 Inhibitors in combination with chemotherapy in advanced esophagal squa

cell carcinoma and esophagal adenocarcinoma based on the best evidence at the time multiple interactions occurred with the FDA where

key design ele elements were jointly agreed upon keynote 590 was approved in March 2021 and set a new standard of care for

patients with locally Advanced and metastatic esophageal cancer as you heard this morning MC determines the pdl1 CPS Cup points used

in its randomized trial through the process outlined here Pathologists were trained to use these pre-specified Cup points during patient screening for

0590 the validation set for assessing pdl1 expression in the study all pdl1 evaluation was performed in a central

laboratory high quality pdl1 data informed meaningful efficacy subgroup analyses and support the allomer indication for keynote

590 while higher pdl1 expression enriches for pemis M monotherapy efficacy in esophageal cancer we cannot predict who will benefit especially when

chemotherapy is added to pema's map finally acknowledging that pdl1 testing outside of clinical trials is variable We Stand by the meticulous process and

data in keynote 590 Merc phase 3 trials are designed with strict statistical methods and these methods should be followed for

labeling postt talk subgroup analyses without type 1 error control may lead to chance findings that could potentially be misleading and thus should be

interpreted with caution as discussed this morning the FDA pooled analysis has inherent limitations postt talk subgroup and

pooled analyses should not supersede patient specific data of a phase three trial with a diagnostic specifically developed for use with pism

app since the approval of keynote 590 in 2021 there have not been any new efficacy or safety data that changed the

benefit risk profile for pmis MB in this patient population the pdl1 assay is specific to pem's M there are key

differences in determining a restriction of this indication by pdl1 Cup point compared to those for camab or panitumumab and elpar molecular

alterations such as K and B M excuse me molecular alterations such as K and braam mutations strongly predict response whereas pdl1 expression is a

Continuum can be modulated by other therapies like chemotherapy and is not always predictive of immunotherapy response nccn ASCO and esmo guidelines

provide detailed recommendations based on different pdl1 Cup points Physicians use these guidelines the label and P patient specific characteristics to

choose the right treatment now Dr Pua Baga will share data from the pivotal phase three study thank you Dr banza my name is Puja

bhagya I am the upper GI cancer clinical lead at MK and I will present efficacy and safety data from keynote 590 keynote 590 supported the approval

of Kuda for the first line treatment of adults with metastatic esophagal cancer in keynote 590 patients had metastatic or locally Advanced

unresectable esophageal cancer the stratification factors in this study were regions histology and Eco status this study had dual primary

endpoints of os and PFS and secondary endpoint of Orr Alpha for statistical testing was initially allocated to the overall

survival and progression free Survival Dual primary end points and then passed to the key secondary Endo of objective response

rate keynote 590 was initiated based on earlier studies showing that fism AB alone could trigger anti-tumor responses in esophagal

cancer chemotherapy forms an essential backbone for treatment of esophagal cancer and combining it with pisab May benefit patients with a wide range of

pdl1 expression as seen in other cancers this led to the combination being used in this trial the study was originally designed

to test hypothesis in both the it group and biomarker positive patient after keynote 590 began results

from the phase 2 keynote 180 trial showed that patients with CPS greater than equal to 10 responded better to pism app this led to CPS greater than

equal to 10 being selected as subgroup for analysis in keynote 590 the statistical plan was then adjusted to test hypotheses in both the

CPS greater than equal to 10 group and the IT population Baseline characteristics were well balanced between the two arms

approximately 73% patients enrolled had squamous cell carcinoma pdl1 CPS greater than or equal to 10 comprised a approximately 50% of

the population and pdl1 CPS greater than or equal to 1 included about 85% of the population although CPS greater than or

equal to 1 was not a pre-specified cut point in keynote 590 MC has confidence in the accuracy of these data for two

reasons first this cut Point has been used in previous esophagal studies and continues to be utilized in our ongoing esophagal

studies second Pathologists at our testing Labs were trained and certified at this cut point and the testing lab had validated this cut Point

establishing reproducibility and repeatability keynote 590 met success criteria for all endpoints in the intent to treat population which was all

patients regardless of pdl1 status and histology the overall survival curve favors pisab with a 27% reduction in the risk of

death progression free survival curve also favors pisab reducing the risk of progression or death by

35% at 2 years 28% of patients receiving pisma plus chemotherapy remain alive versus 16% of those who received

chemotherapy notice the tail of the curve which is characteristic of pism the safety profile of the investigational arm is consistent with

the established safety profiles of fmab and chemotherapy the addition of fmab adds immune mediated AES and infusion reactions which were mostly low grade

and manageable it is known that some immune mediated AES such as endocrinopathies will require long-term hormone replacement these data highlight the

favorable benefit risk profile of falisa plus chemotherapy for all patients to address fda's questions we

will now look at different pdl1 cut points although a higher magnitude of benefit is seen with increasing pdl1 expression all subgroups are

directionally consistent with the it population with Point estimates of the hazard ratio being less than one in the CPS greater than equal to one

subgroup the point estimate of the hazard ratio for OS and PFS is 0.7 and 0.63 respectively demonstrating a

clinically meaningful benefit the CPS greater than equal to 10 subgroup also shows a clinically meaningful benefit the CPS less than one subgroup

was not pre-specified with formal statistical testing given the small number of patients the OS has aard ratio confidence intervals are very wide and

overlap with it what this Forest plot does not show is that there were four patients that

had a complete response in the chemotherapy plus pisab arm versus none in the chemotherapy alone arm of the four patients three patients were still

alive at 5 years in patients with CPS between 1 and less than 10 we see a benefit with an OS Hazard ratio of

0.84 and confidence intervals that overlap with the it indicating that the benefit is not driven by CPS greater than

10 at a 5-year follow-up assessment the benefit of pism app is consistent with the primary analysis with an improvement in the hazard ratio underscoring a tenet

of immunotherapy there is no biological rationale to suggest that the safety profile of fism app would change based on pdl1

expression the safety profile is in general similar across CPS cut points even when CPS less than one is compared

with other CPS cut points of note any death due to AES is already accounted for in the km curves and the hazard

ratio for the CPS less than one subgroup as we see here is less than one for the escc patients in keynote 590 the

results were similar to the it population with statistically significant and clinically meaningful effect regardless of pdl1 status at a

five-year follow-up assessment the benefit of femal isab in the escc is consistent with the primary analysis and maintained suggesting long-term efficacy

for this patient population as requested by the the FDA overall survival and progression free survival data by additional pdl1 CPS cut

points are shown here about 90% of patients with escc have tumors expressing CPS at a score of one or more and this subgroup demonstrates

meaningful benefit the escc population with CPS less than one represents a subgroup of a subgroup and therefore meaningful conclusions cannot be made

overall the results in the ESC population are consistent with the it population in summary there is a high unmet need in firstline metastatic

esophagal cancer pism AB added to chemotherapy significantly improved overall survival progression-free survival and response rates in the it

population with greater benefits at higher pdl1 levels patients with lower CPS course also benefited showing that CPS alone cannot predict who will

respond to pisab and chemotherapy health related quality of life remains stable during treatment was similar between arms and consistent

across CPS subgroups the manageable safety profile reflects the known safety profiles of the components and is generally similar across CPS

subgroups the totality of evidence supports that P nalism app should be available to all patients with esophageal cancer as a treatment option consistent with the approved label thank

you and I will now invite Dr enzinger to the podium good afternoon uh my name is Peter enzinger I am a GI oncologist at the Dana Farber Cancer Institute and an

associate professor at Harvard Medical School I'm happy to discuss my clinical perspective on the data shared today here are my disclosures unfortunately esophageal

cancer remains underst studied and underserved leaving patients with limited options for treatment most patients are diagnosed at stage four and as we can see on the right they have a

dismal prognosis before approval of immunotherapy the only option was chemotherapy this is a difficult to treat disease and most patients do not

live to get second line current treatments are largely paleo of which emphasizes the urgent need for new treatment options to improve patient

outcomes and quality of life for more than three decades treatment has been a combination of platinum and flu urisol only by borrowing the FDA indication

from gastric adenocarcinoma have we been able to introduce tusab and ramam ab for some of our adenocarcinoma patients pmab plus chemotherapy is

practice changing for firstline metastatic esophageal cancer and a valuable treatment option that should remain a choice Physicians should

consider urgency of treatment timing of biomarker testing adverse event profile and Prospect of long-term survival

patient P1 expression level can assist with individual patient management decisions although pdl1 testing AIDS in understanding who may have increased

benefit from immunotherapy it presents some challenges in clinical practice some of the reasons are as follows there can be different assays and antibody clones used by various organizations

that are not FDA approved which may result in staining variability and unlike a clinical trial in the real world patients may not have a sample of

sufficient quality for pdl1 testing and finally there is significant inconsistency in pathologist training and cut Point interpretation to investigate pdl1 testing and

treatment patterns among Advanced or metastatic esophageal cancer patients a retrospective observational study was conducted using the flat iron database

of adult patients who received firstline systemic treatment of the 670 patients treated in the first line setting 66%

were evaluated for pdl1 in this group 41% were treated with chemotherapy plus immunotherapy and 58% received chemotherapy alone these data suggest

that Physicians and patients carefully weigh the risks and benefits of available treatment options importantly if the indication is restricted to patients with CPS greater than equal to

one this will deprive approximately 11% of patients of potentially effective therapy I would like to highlight a patient treated on keynote 590 this is a

western patient in her 40s with esophageal Squam carcinoma with lung metastasis she had a CPS score of less than one she was randomized to the

treatment arm and after 15 cycles of chemotherapy and pism ab Imaging showed a complete response I want to acknowledge that such responses may be seen with chemotherapy

alone however what is remarkable is the durability of response which lasted for about 50 months further this patient was alive at 5 years this patient

illustrates that despite a CPS score of less than one durable responses and long-term survival is possible which is not typical of chemotherapy of note this may not be

such an unusual result the final results of keynote 590 showed fiveyear survival in approximately 11% of all randomized patients treated with pism and

chemotherapy compared to only 3% with chemotherapy alone in conclusion treatment options are severely limited for this disease checkpoint Inhibitors have

revolutionized the care of patients with esophageal cancer improving survival and maintaining quality of life the choice to add a checkpoint inhibitor must be

individualized and depends on many factors variability in real world pdl1 biomarker testing May hinder treatment decisions further the scientific

Community informs decision-making through clinical guidelines the allomer indication allows patients to have immunotherapy as a firstline treatment option at the discretion of their

treating physician thank you and Dr Panza will now make closing remarks thank you Dr enzinger in summary Kino 590 established pisab and

chemotherapy as a standard of care in esophageal carcinoma it met all its primary and key secondary end points in the intention to treat population the

label reflects the study outcome metastatic esophagal cancer is a fatal disease where survival is measured in months pisab combined with

chemotherapy is one of the only treatment options for these patients data show that efficacy can occur across a range of pdl1 expression

including in those with low or no expression as we heard from Dr Enz singer restricting the indication would leave these patients no choice but

chemotherapy the current indication allows Physicians to make the best possible choice for patients with esophageal cancer thank you for your

attention thank you we will take a quick 10-minute break to allow for the next presentation to set up panel members please remember that there will be no discussion of the meeting topic during the break amongst yourselves or with any

member of the audience we will resume at 2 210 p.m. eastern

time e e e e e e e

e e e e e e e e

hey welcome back everybody we will now proceed with our second presentation from bris Meers qub good afternoon my name is Ian Waxman and I'm part of the late development oncology organization at brist Meers

squib I'd once again like to thank the advisory committee members and the FDA staff for this opportunity to discuss the data for ABDO this time in combination with chemotherapy or

ipilumumab in first line esophagal SEL carcinoma also known as escc these data come from the Checkmate

648 study and resulted in FDA approval for this indication in May of 2022 the indication statement for this approval is shown on the slide and it's

important to highlight two things here first the approval was granted regardless of pdl1 status and second since the initial approval our interpretation of the study results has

not changed with longer followup although the indication statement is not limited to pdl1 positive population clinical data by

pdl1 expression level are included in section 14 of the uspi these data are included to ensure that treating Physicians have sufficient information

regarding the impact of pdl1 positivity when discussing treatment options with their patients since the time of this approval results from additionally soual cancer

Studies have been reported as was the case for gastric cancer different sponsors in Incorporated different methods for measurement of pdl1 and selected different cut offs to determine

positivity in these studies for esophageal sis cell carcinoma nccn recommendations for the ABDO combinations are not based on pdl1

expression level a decision influenced by the overall High rate of pdl1 positivity in Checkmate 648 since information regarding the

impact of pdl1 expression on outcomes is readily available including in the uspi it's helpful to understand how often Physicians are testing their patients and whether or not their treatment

decisions are influenced by those test results what we see based on us flat iron data is that about 60% of advanced escc patients who receive firstline

treatment are tested for pdl1 expression even without a requirement to do so demonstrating that many Physicians see value in pdl1 testing today

and when we move from testing patterns to treatment patterns we see that Physicians are often times incorporating the test result into their treatment decisions with the presence of a

positive test result leading to Greater likelihood of treatment with an iio regimen on the left hand side of the slide we see that among patients who test positive for pdl1 expression three4

are receiving IO plus chemotherapy in blue and about 8% receive neop plus Epal umap in red in the middle you can see that among the small subset of patients

tested negative for pdl1 only about one quar receive an IO regimen another way to think about this is that among all among all treated

patients less than 5% are treated with IO and known to be pdl1 negative on the far right hand side we are reminded that some treatment

decisions continue to be made in the absence of a test result we consider use of IO to be approp in this patient segment since an untested patient is

more likely to be positive than negative with approximately 90% of escc patients considered pdl1 positive when using the cps1 cut

off with this in mind we're here to discuss whether any label changes for obdo and firstline escc are needed also using this as an opportunity to consider

harmonization of product labels based on pdl1 expression our goal is to ensure that each firstline escc patient has every appropriate therapy available to

them along with clear guidance to inform choice of treatment a review of subgroup analyses by ped1 expression level from the Checkmate 648 study and the unique

clinical considerations for this patient population are also critical parts of the discussion and we will turn to these topics next once we've considered these

additional points I'll return to summarize potential options for labeling also briefly described here one option is to modify the indication to only

include patients testing pdl1 positive using any FDA approved test this would limit treatment to patients more likely to benefit based on the clinical trial data but could leave some patients

without a potentially important treatment Choice the proposal to use any approved test would minimize the impact on each institution's current testing

practices since CPS is used much more widely than TPS today the second option is to leave the indication as is so that Physicians can

continue to make treatment decisions informed by the data as currently described in the uspi and consistent with nccn guidelines given that escc is a rare

disease with very high prevalence of pdl1 expression we consider this to be the preferred option additional considerations for each of these approaches are shown here

and will be discussed in more detail in the next parts of the presentation here is the agenda for the remainder of our time first Dr Dana Walker from the drug development

organization at BMS will review the relevant efficacy and safety data from Checkmate 648 then Dr Ronan Kelly from Balor University will provide his clinical

perspective on the value of pdl1 testing for patients with ESC and finally I'll return to review options for labeling thank you and I'll now turn it over to Dr walk

thank you my name is Danna Walker and I'm the global program lead for opdo and yvo for GI and gu cancers at BMS today I will present data from the check meet

648 study demonstrating the benefit risk profile across pdl1 subgroups in esophageal cancer Checkmate 648 is a randomized

open label phase 3 study that enrolled previously untreated patients with unresectable Advanced recurrent or metastatic esophageal squa cell

carcinoma regardless of pdl1 expression a total of 970 patients were randomized to receive noolab plus chemotherapy

noolab Plus iolab or chemotherapy alone stratification factors included tumor cell pedl 1 the primary end points of the study were overall survival and

progression-free survival per bicor for patients with tumor cell pdl1 of 1% or higher referred to as the pdl1 positive population Checkmate 648 demonstrated

both a statistically significant and clinically meaningful Improvement in the primary Endo of overall survival in the pdl1 positive population with a hazard ratio of

0.54 and a 6.3 month Improvement in median overall survival compared with chemotherapy alone the secondary end point of overall survival in the all randomized

population was also met with a hazard ratio of 0.74 shown here is overall Survival by TPS subgroups the data in the blue boxes highlights the pre-specified primary and

secondary analysis populations the other TPS subgroup analyses were exploratory overall survival benefit was observed across all pdl1 positive subgroups there was a higher likelihood

of overall survival benefit observed in patients whose tumors expressed pdl1 in patients with TPS less than one the overall survival Hazard ratio was

0.98 suggesting there is no overall survival benefit during the study data began to emerge suggesting the potential predictive value of CPS in upper GI

tumors therefore we conducted exploratory overall survival analyses in pdl1 CPS subgroups similar to the TPS

less than 1% subgroup the hazard ratio in the CPS less than one subgroup was 0.98 please note approximately 90% of

patients in the trial with known pdl1 status were CPS equal to uh CPS greater than or equal to one therefore the CPS less than one analyses should be

interpreted with caution additionally and similar to the TPS subgroups patients were more likely to derive an overall survival benefit at any level of

pdl1 positivity as measured by CPS in contrast to what we saw in gastric cancer there is no evidence of increased benefit at higher CPS

scores here we present the exploratory subgroup data for the noolab Plus iuma versus chemotherapy comparison looking at overall survival across CPS subgroups

a similar Trend was observed to the Neo Plus chemo versus chemo comparison with a higher likelihood of overall survival benefit with Neo Plus ippi and all CPS positive patients

results of long-term overall surviv survival followup across pdl1 subgroups were generally consistent with those reported at the primary analysis for both the Neo Plus chemo and Neo Plus

ippy regimens and is discussed in more detail in our briefing document the safety profile of Neo Plus chemo and Neo Plus II observed in

Checkmate 648 was consistent with the known safety profile of the individual drug components as expected the addition of noolab to standard chemother therapy

was associated with added toxicity grade three4 treatment related Adverse Events and those leading to discontinuation of any treatment component were numerically higher in patients receiving Neo Plus

chemo of note in both Neo containing arms the majority of immune mediated events were low grade manageable with established treatment algorithms and

reversible importantly the safety profile of both neoc containing regimens did not differ based on pdl1 expression and was consistent across all pdl1

subgroups evaluated in summary check meet 648 demonstrated statistically significant and clinically meaningful overall

survival benefit in the TPS greater than or equal to 1% and all randomized population exploratory analyses suggests similar overall survival benefit across

all pdl1 positivity and long-term overall survival follow-up data are consistent with the data available at the time of approval the safety profile of Neo Plus

Kimo and Neo Plus ippi was consistent with the known safety profile of the individual drug components and did not differ based on pdl1 status overall there's a positive

benefit risk profile in all pdl1 positive subgroups thank you I will now turn it over to Dr Kelly for his clinical perspective thank you very much Dr

Walker it's it's a real pleasure to be here my name is Dr Ronan Kelly I'm the director of The Charles A Salmons Cancer Center at Baylor University Medical

Center in Dallas Texas and I'm the chief of oncology for the Baylor Scot white health system which is the largest not for-profit health system in Texas I'm a

paid consultant for BMS the Bor Scott White health system has 51 hospitals throughout the state of Texas and 13 dedicated Cancer Centers which

represents one of the largest Commission on cancer n network of cancer hospitals in the United States as such I have exposure to treatment patterns in both

an academic and in a community setting in both urban and rural areas alike therefore I can see the challenges that exist for both medical oncologists and

Pathologists with regards to pdl1 testing and pdl1 interpretation for esophagal gastro cancers in real world treatment

settings esophageal squamous cell Caron is truly an orphan disease in the United States if you look at the sear data

approximately 14,000 patients were diagnosed with escc in the US over an 11-year period so just over a thousand

patients per year which translates into very few patients being seen by doctors across the country unfortunately the majority of these patients are diagnosed

with Advanced disease historically the breakdown between esophageal adenocarcinoma and esophagal squel carcinoma was 70% and 30% respectively

but recent epidemiological data indicates that escc is decreasing in the United States because of falling smoking rates throughout this

country treatment recommendations therefore in my opinion should be kept as simple as possible for this orphan disease and I agree with the nccn

guidelines that continue to recommend treatment for this disease regardless of pdl1 expression unfortunately very few patients with this disease go beyond

firstline treatment the data shows that approximately 70% of escc patients receive firstline treatment but less

than a quarter of our patients only 23% make it to the second line setting that's an enormous drop off and it

indicates we should not be waiting to give our best treatment options in the second line setting or third line setting in fact less than 8% of these

patients make it to the third line setting if you look at the clinical trial data from Checkmate 648 which includes patients who historically would have a better

performance status than real world patients less than 50% of those make it to the second line setting it is my

opinion that a pd1 inhibitor plus chemotherapy or the combination of neolab Plus iaab for patients who may

decline chemotherapy or who may be considered by their oncologist to be too frail for chemotherapy represents breakthrough treatment options for patients with escc

it's very important for the panel to understand that escc is biologically different from the disease we spoke about this

morning this is not gastric cancer and it's not gastro esophageal Junction adenocarcinoma these are two very different diseases notably escc is also

regarded as more immunogenic than the adenocarcinoma histology you can see the cancer gen data on the on the right there showing

that escc which is highlighted in red is genomically similar to Sous cell head and neck cancers and the location as you

can see by the figure also indicates this tumor occurs much more proximately in the esophagus which leads to significant more dysphasia and and

significant problems for our patients recent data from other large phase 3 trials which we won't discuss today around the world have now shown

the uh the continue to demonstrate the efficacy of iio regimens in ESC even in patients with low pdl1

expression pdl1 in esophagal SEL carcinoma is likely not the only Factor influencing response here escc develops

in a chronically inflamed tumor micro environment dominated by exhausted te cells and suppressive cell populations another challenge that's

different here than what we talked about this morning is in these patients to relieve their discomfort to improve their calorific intake we often often uh

give them paliative radiation to improve their ability to eat many of these patients struggled to even swallow their own saliva when we offer them radiation what

we do is we upregulate pdl1 we talked this morning about the dynamic nature of this biomarker so when we offer paliative radiation we are changing the

pdl1 status of that patient therefore spatial and temporal heterogenity may be even more problematic in this disease where radiation is the norm and it's

also not safe to continue to repeat endoscopic biopsies post radiation so the ability to do longitudinal biopsies as discussed this morning is not feasible in this situ

situation in terms of pdl1 testing we know the majority of patients are tested by 60% and the vast majority of centers

are utilizing CPS with less than 5% of centers utilizing TPS alone in this setting we heard this morning again the reality pdl1 is imperfect and I'm not

going to get into that all again all those reasons were discussed previously furthermore escc C has very high pdl1

expression at Baseline with about 90% of tumors expressing measurable pdl1 by CPS so in conclusion it's my opinion

that maintaining the current indication in this disease setting is appropriate the biology of the disease is different from gastric and gastrosoph Junction

adenocarcinomas pdl1 may be not as important in this disease setting which is dominant ated by other immunosuppressive phenotypes unfortunately only about 25%

of our patients even make it to the second line treatment setting so it's important to give our best treatment options up front and this would include noola M with

chemotherapy or a chemo free option with noola Plus iuma at the present time we do not require testing for

escc as for the nccn guidelines for many of the reasons that I've explained and the high prevalence of pdl1 expression in this disease setting

if a restriction is required then pdl1 positivity by any FDA approved measure makes the most sense in clinical practice thank you very much and I'll now turn it back to Dr Waxman to

conclude you Dr Kelly as I turn to review our proposed options for labeling I'd first like to reiterate that this is an important issue with more than one potential solution the FDA has asked you to consider

whether the benefit risk assessment is favorable in ESC patients with pdl1 of less than one regardless of how you answer that question there are still two important options to

consider the first option is to modify the indication based on pdl1 positivity by any approved test reserving treatment for those more likely to benefit and allowing for potential harmonization

across the class the second option is to keep the current indication with details regarding the impact of pdl1 expression remaining in section 14 of the label

this provides Clarity regarding the impact of pdl1 on outcomes while providing an opportunity for all patients to receive immunotherapy especially important given

that most escc patients will be pdl1 positive we believe that both of the proposed options are reasonable although leaving the indication as it's written today provides the most flexibility for

patients thank you once again for your time and attention thank you we'll take another 10-minute break to allow for the next presentation to set up panel members please remember there should be no discussion of the meeting topic during

the break amongst yourselves or with any other member of the audience we will return at 2:40 p.m.

p.m.

e e e e e e e e

e e e e e e e e

e e proed with our third presentation from Beijing good after everyone my name is Mark lanasa and I'm the chief medical officer for solid tumors at Beijing I

again want to thank the FDA the chair and the members of the committee for the opportunity to share our tiis results in this important discussion of squis cell esophageal

cancer this afternoon I will review the results from our pivotal study rational 306 in squamous hystology esophageal cancer and share additional subgroup

analyses to explore a potential relationship between pdl1 expression in survival I will then ask Dr uboa to provide her clinical perspective on the

use of pd1 Inhibitors in patients with the we also have additional experts with us today to help to address your questions in 2018 we initiated the

global pivotal phase 3 rationale 306 study evaluating the efficacy and safety of tisis combined with chemotherapy versus placebo and chemotherapy as

firstline treatment for patients locally Advanced unresectable or metastatic Sous syy esophageal cancer which I will refer to

ASC study 306 met the primary endpoint of overall survival in the it population at a pre-specified interim analysis in February of

2022 our bla for this indication was submitted on July 18th 20123 and is currently under review on March 14th 2024 talism AB was

approved by the FDA to treat patients with unresectable or metastatic escc after prior systemic therapy that did not include a pd1 inhibitor based on the

results of the global phase 3 rationale 302 study in this study tisis maab prolonged overall survival as monotherapy when compared to

investigator choice of available chemotherapies overall results from our pivotal study show that firstline treatment with tisab combination with

chemotherapy offers substantial benefit and overall survival tisis maab in combination with chemotherapy produced statistically significant and clinically meaningful Improvement in OS as well as

improvements in progression free survival objective response rate and duration of response in the overall population TI's map also showed an acceptable safety profile across a broad

population of patients with unresectable Advanced or metastatic esophageal sisel carcinoma finally analyses across pdl1 expression

levels show that the benefit of tsilis ab plus chemotherapy in patients with locally Advanced or metastatic ESC is most favorable among patients with a

pdl1 score greater than or equal to 1% now I would like to review our study design and key

results rationale 306 is a global randomized double blind Placebo controlled study in 649 patients with a histologically confirmed diagnosis of

esophageal sell carcinoma with either metastatic or locally Advanced disease that was not amenable to Curative intense surgery or chemor radiation patients with adoc carcinoma

were not eligible for the study stratification factors include a geographic region whether the patient received prior Curative intent therapy and the investigator's choice of

chemotherapy all patients were required to have at least one valuable lesion per resist version 1.1 and ecog performance status of Z or one as well as adquate

organ function and nutritional status patients were randomized one to one to receive talism M 200 milligrams administered intravenously every 3 weeks

for matching Placebo until disease progression or unacceptable toxicity both treatment arms were administered in combination with the investigator's choice of standard chemotherapy

including a platinum agent combined with either a Flor perimidine or pet Axel the primary endpoint of rationale

306 was overall survival in the it analysis set additional secondary end points such as progression free survival objective response rate duration of

response and safety were also evaluated overall survival was tested hierarchically in the pdl1 greater than or equal to 10% subgroup defined as a

pdl1 score as assessed using the sp263 assay following the tap scoring algorithm note that OS testing in the pdl1 positive subgroup was a secondary

endpoint and was tested after it analyses of both PFS and O because the requirement for pdl1 testing was added via a protocol

Amendment and because Central testing was retrospective a small proportion of patients do not have an available pdl1 score next I will share Baseline

demographics and disease characteristics overall Baseline demographics were generally balanced between treatment arms the median age was 64 years and 87% of the participants

were male consistent with the epidemiology of ESC the majority of patients were enrolled in East Asia which is also consistent with the global epidemiology

of escc with the remaining 25% of patients enrolled in Europe Australia and the United States as was the case with our study in gastri cancer that I presented this morning enrollment in the

United States became infeasible after Top Line results from the pisab study presented today became available similarly Baseline disease characteristics were also generally

balanced and representative of the target patient population the median time from initial diagnosis was approximately 2 months most patients had metastatic disease at study entry and

44% of patients had prior Curative intent therapy approximately 2third of patients had an ecog performance status of one in total 34% of patients had a

baseline pdl1 score greater than or equal to 10% 49% of patients had a pdl1 score less than 10% and approximately

16% had an unknown pdl1 status rationale 306 met the primary endpoint of overall survival in the it population tisab plus chemotherapy was

Superior to Placebo plus chemotherapy with a statistically significant 34% reduction in the risk of death and a clinically meaningful Improvement in

median OS of 6.6 months upon visual inspection you can observe that the Capa Meer curves separated early and maintain separation throughout the period of

followup the benefit observed in overall survival is supported by the secondary end points patients treated with tsilis maab and chemotherapy had longer PFS

with a statistically significant and clinically relevant 38% reduction in the risk of progression or death median PFS was extended by 1.7 months objective

response rate also showed a statistically significant and clinically relevant B benefit favoring tisis the absolute difference in response rate was

20.2% with an odds ratio of 2.31 the duration of response was also extended in the tisis ab plus chemotherapy arm

to further assess the clinical benefit in patients with pdl1 low expression we conducted several additional subgroup analyses across a range of pdl1 expression

levels here we show a forest plot of overall survival across various pdl1 subgroups at the interim primary analysis data cutoff date although a

pdl1 score of 10% was the prespecified cut off for FC analysis in escc we do not observe meaningful differ iation and treatment effect on overall survival

above or below 10% therefore we next evaluated additional lower cfff scores per for predictive treatment effect at a cut off

of 5% a differential effect is observed but this apparent effect is potentially driven by the underperformance in the pdl1 less than 1%

group to further explore the potential relationship between pdl1 score and overall survival we are now now showing a forest plot of the overall survival

Hazard ratio within the specific pdl1 subgroups please note again that the pdl1 score is unknown in 16% of the it population and that some of the

subgroups presented are quite small first we observe a particularly strong treatment effect in the 5 to 10% Group which further supports our position that above below 10% is not an

appropriate cut off for patient selection while we acknowledge that the hazard ratio in the greater than is greater than one in the pdl1 less than

1% group the control arm in this subgroup is very small only 25 patients I will show in a subsequent slide that median overall survival in this subgroup

is 16.1 months and this result appears to be random High bias that said we acknowledge that a favorable benefit risk is not established in this

subgroup turning to the 1 to 5% group this group compris approximately 20% of the total population and therefore merits careful consideration at the time of the primary

analysis the hazard ratio in this subgroup was 0.93 here we are showing the same Forest plot with more mature data these data are from the three-year followup and

have a data cut off date approximately 20 months after the interm analysis we believe that these longer term data provide a more robust assessment of effect sizes in small subgroups given

the greater data maturity the hazard ratio in the it population increased slightly from 0.66 to 0.70 but the trends are essentially

identical and the median Improvement in overall survival of 6.6 months is maintained with longer followup with additional followup we

observe that the median overall survival in the 1 to 5% group improved to 0.86 at the three-year followup in the subgroup we observed immediate

Improvement in overall survival of 3.4 months supported by favorable Trends in progression free survival and objective response rate we believe outcomes in the one to less than 5% group to be

clinically meaningful and thus propos at a cutoff value of 1% is most appropriate for patient selection in this indication next I would like to review

Medan overall survival across all of the subgroups presented with 3 years of followup at the proposed cut off of greater than or equal to 1% the median benefit in overall survival conveyed by

the addition of tisab to standard of care chemotherapy is 7.2 months now I will briefly review

safety overall the AE profiles observed for tisis maab plus chemotherapy in the first line setting were similar to the known safety profile of chemotherapy Ed of talism ab and the expected symptoms

of ESC the overall Trends in the safety data set are consistent in the overall safety data set and then the pdl1

greater than or equal to 1% group as expected immun mediated Adverse Events are more frequent in patients receiving tisel ISM AB than in those receiving

chemotherapy alone grade three or greater AES and AES leading to dose modification were similar between groups similar to the data presented this morning for gastric cancer there is an

increase of the rate of AE leading to discontinuation and SES in the tisis M containing arm in this tornado plot we are showing the treatment emerge in Adverse Events

of any grade in pdl1 greater than or equal to 1% subgroup occurring in 20% or greater of patients the majority of Adverse Events are commonly observed in

this disease with the chemotherapy component as was the case with gastric cancer there is no clear trend of increase of individual AES for e

severity with the addition of tisab to summarize the primary analysis of rationale 306 showed that the addition of tisis to chemotherapy

provided substantial Improvement in overall survival with a hazard ratio of 0.66 and 6.6 months of incremental OS benefit at the median this benefit was

observed across all pdl1 sub groups with the exception of the less than 1% group secondary endpoints also showed clinically mean meaningful benefit the safety profile was manageable and

generally consistent across the range of pdl1 expression based on the overall data set benefit risk is most favorable in the pdl1 greater than or equal to 1% group

overall we conclude the totality of data supports tisis AB for the Frontline treatment of patients with unresectable Advanced or metastatic escc thank you

and i' now like to invite Dr boha to provide her clinical perspective thank you good afternoon um and thank you for the opportunity to address the panel again I've been compensated for my

travel but not for my time in preparing for today's meeting let me start with a brief background on a Sagal cancer patients diagnosed with Advanced

esophageal scol carcinoma have poor prognosis and limited treatment options in the US this is a rare tumor with declining incidents and comprising

roughly less than 1% of all cancers patients with Advanced esophagal scol carcinoma have a 5year survival rate of only 6% which is one of the

lowest rates among all cancer types these patients are generally older and have other cor morbidities additionally many patients have extensive disease related symptoms

they have trouble eating struggle with weight loss and have significant pain over the last few years we've seen that treatment with anti pd1 antibodies in combination with chemotherapy can

significantly prolong overall survival for these patients in my practice which is not different from most from most other experts pedal one testing is done for all patients with Advanced

gastroesophagal cancers regardless of histology unless tissue is unavailable about 90% of patients with a sopal scal carcinom I have pedial one

positive tumor as a clinician who treats these patients here's my interpretation of the results Tes liab improved overall

survival in the rationale 306 study across the intent to treat population with a hazard ratio of 0.66 which was maintained at 3 years

there were very few patients who had tumors with pedl one tap score of less than one only about 9% of the population enrolled in the 306 trial the outcomes

of these patients are based on the post Hawk expl laboratory analysis in the study and the numbers are too small uh to be statistically

reliable never the last in the pulled analysis shared by the FDA the hazard ratio in this subgroup across studies is

1.1 so in my clinical practice I offer treatment with anti- pd1 agents in combination with chemotherapy to patients whose tumors have pdl1 score one or greater I feel this is a

clinically relevant pdl1 expression question threshold and I would urge the committee to recommend a unified pdl1 C of of one or greater across pd1

Inhibitors thank you for your time thank you so much we will now proceed with fda's presentations starting with Dr gas trasta

thank you good afternoon everyone uh my name is gaka shast I'm a hematologist medical oncologist and a clinical reviewer in division of oncology 3 at

the FDA the FDA is convening this odac meeting to discuss the risk benefit of the use of immune checkpoint inhibitors for the first line treatment of patients with unresectable or metastatic

esophagal Sile carcinoma at different levels of pdl1 expression the members of the FDA team are listed on the slide as discussed in the presentation

by Dr cassak the current label of immune checkpoint Inhibitors approved in first line ESC are agnostic of pdl1 testing results these were based on the

intention to treat population enrolled in the pival studies at the time of approval of bolism ab in 2021 and noolab in

2022 less was known about pdl1 as a predictive biomarker in escc and given the exploratory nature and small sample sizes of the subgroups

FDA did not restrict labeling based on pdl1 status based on the results of each individual trial results are now available across multiple trials that make inferences

based on these subgroups more reliable and provides a framework to discuss the adequacy of pdl1 expression as a predictive biomarker for patient selection I will review the data from

the individual studies that led to the approval of pisab and neolab in this setting the data submitted to support potential approval of disis AAP for the

same indication and the FDA patient level pooled analysis from these studies in escc there now appears to be a consistent pattern across three

available trial data sets that the overall efficacy of immune checkpoint Inhibitors in this setting is driven predominantly by pdl1 high subgroups

with a concern for lack of benefit in tumors with low pdl1 expression particularly those with pdl1 less than one FDA believes a contemporary risk

benefit discussion evaluating the available data is required to further Define the indication of anti- pd1 antibodies based on pdl1 expression in

patients with escc the three individual study designs have already been outlined today and the schema is presented within the FDA briefing

document before we review the efficacy results from each individual study this table provides the key features of the three pivotal studies keynote 590 for

pisab checkmat 648 for noolab and rational 306 for talism app keyote 519 Ro patient with esophagal and G Junction

carcinoma irrespective of hology overall 73% of the enrolled patients in keynote 590 had squamous esophageal sosal

carcinoma both Checkmate and rational only enrolled patients with scous cell hystology escc all three trials allowed for

patients to enroll regardless of the tumor p L1 expression the tests used for determining pdl1 expression were different for each

trial keynote 590 use combined positive score CPS Checkmate used DPS with a pre-planned retrospective analysis for

CPS and rational 306 use a visually estimated CPS also known as tumor area positivity dap only Checkmate 648 had pdl1

expression as a pre-specified stratification Factor the primary endo and the pre-specified pdl1 cut off you use in

each study are listed here in summary each trial used different assay for the assessment of pdl1 expression different scoring algorithm and different primary

endpoint based on pdl1 cut off the results of the three studies were statistically significant for overall survival analyses as pre-specified in the statistical plan

the hazard ratio for the it population across all three trials range from 0.66 to 0.74 in the pre-specified pdl1 positive

subgroups which was CPS 10 or greater for keynote TPS one or greater for checkmate and tap 10 or greater for rationale the magnitude of benefit was

higher with Hazard ratio ranging from 0.54 to 0 62 to study the adequacy of pdl1 expression as a predictive biomarker for

use of immune checkpoint Inhibitors in the setting FDA conducted patient level analysis of the three randomized study in the relevant population of

escc for uniformity of comparison between the three studies the fdcc modified population consistent of only patients with escc histology who

received immunotherapy in combination with chemo versus chemo alone patients with adenocarcinoma or those on the Neo I arm in Checkmate were

excluded this population therefore differs from the it population for keynote 590 548 patients were identified excluding 2011 with

adenocarcinoma for checkmate there were 629 patients identified excluding 16 with nonam hology and 325 on the Neo I

arm for rationale there were 648 patients and 1 patient was excluded for nonology please note for the remainder of the FDA presentation we will focus on

this escc modified population this table summarizes the key demographic and disease characteristics of these patients in general the characteristics of patient across trials

was comparable 2/3 to 3/4 of the enrolled patients were Asian not shown here however relevant to the discussion today there was one known MSI High

patient enrolled in rationale and none in keynote however the MSI status of tumors for both for most patients in these trials was unknown the proportion of patients at

pdl1 cut offs across the three studies is depicted in the bar graph here this distribution is based on cpf for keynote 590 and Checkmate 648 and tap for

rational 306 the proportion of tumors that were pdl1 less than one was similar across studies ranging from six uh ranging from 8 to

10% we will now review the efficacy results of the escc modified population again those are patients with escc histology who received chemotherapy in

combination with immunotherapy these are the overall survival results the overall survival appears to be favorable across all three trials and this appears to be generally

consistent with the results of the it population but has ratio around 0.7 in the following slides I will present the FDA analysis of overall

survival data for each study for the escc modified population based on pdl1 cut offs FD acknowledges the limitations of small sample size and exploratory

nature of these these subgroup analyses this is the forest plot for overall survival for escc for keynote 590 analysis using the pre specified

pdl1 cut off of CPS 10 or higher same as the it was performed this represented 52% of the escc Improvement in survival with

addition of bolism app to chemo appears to be of Greater magnitude for CPS 10 or greater with Hazard ratio

0.57 whereas in patients with CPS less than 10 the hazard ratio is 0.95 and in patients at CPS less than one the hazard ratio

is one which appears to suggest no benefit for checkmate 648 analysis using the pre-specified pdl1 cut off of one or

higher same as the it was performed this represented 87% of the escc population when using CPS the hazard ratio for CPS

one or greater was 0.69 while again in the subgroup of CPS less than one there appears to be marginal or no benefit meod with Hazard ratio at

0.93 not shown here but results of analysis in TPS less than one were concordant as well and the hazard ratio was 0.96 for rational 306 analysis using the

pre-specified tap cut off of 10 or higher same as the it was performed this represented 34% of the population Improvement in survival with addition of

this Liz map to chemo appear to be of Greater magnitude for dap 10 or greater with Hazard ratio of 0.66 however in patients with t less

than 10 the point estimate is 0.76 as the point estimate for the subgroup of tap 1 to 10 is favorable at

0.65 the attenuation and benefit in tap less than 10 subgroup is likely driven by patients with TP less than one where there's a potential for detriment that the hazard ratio

1.34 FD acknowledges the limitations of small sample size and expiratory nature of the subgroups in spite of the heterogeneity amongst different trials use of

different pdl1 assays and testing algorithm as well as different pre-specified pdl1 cut off for overall survival in this FDA patient level

analysis over three independently conducted trials it appears that the overall efficacy of anti pd1 in this setting is driven predominantly by pdl1

high subgroup as defined by each study and regardless of the method used to determine pdl1 expression there is replication of results over three trials

and a consistent pattern suggesting lack of benefit in pdl1 less than one tumors see in another way the lack of benefit in pdl1 less than one is further

exemplified by these Kaplan Meer survival curves the figure on the top is for pdl1 1 or greater showing separation of Curves however in the bottom figure

in PDL 1 less than one the curves overlap for keynote 590 and Checkmate 648 and in fact a reverse for Ral

306 in addition to looking a trial separately using the modified escc population FDA conducted an exploratory pulled analysis of patients from all

three studies in the escc population stratified by study this schol analysis included patient level data and was therefore limited to studies that was

submitted to the FDA for review this does not include data from other pivotal studies either positive or negative and can thus introduce bias for pooling of

data pdl1 expression was based on cpf for keynote and Checkmate and tap for rational for uniformity of comparison in escc of immunotherapy plus chemo versus

chemo as shown in the concert diagram patients with nonology and those on Neo I arm in Checkmate were excluded a total of 1825

patients were pulled 9910 received immunotherapy in combination with chemo and 915 only chemo in spite of limitations of pooling of data FDA believes that a pooled analysis of

patient level data May provide the advisory committee with additional context to discuss the risk benefit of anti- pd1 antibodies in relationship

with pdl1 expression in this patient population this is the forest plot for overall survival results there appears to be a benefit in the overall population with addition of

immunotherapy to chemo as compared to chemo alone with Hazard ratio of 0.71 indicated in the Green Arrow the magnitude of benefit appears to increase with increasing pdl1

expression as seen by improving Hazard ratios pdl1 of one or greater 0.68 pdl1 of of 10 or greater Hazard ratio of

0.61 the magnitude of benefit appears to attenuate at lower pdl1 expressions with Hazard ratio for pdl1 less than 10 at

0.82 the hazard ratio in patients with pdl1 between 1 and 10 is 0.77 consistent with the potential for benefi in these patients however again there does not

appear to be a benefit in patients with pdl1 less than one with hait ratio 1.1 like the independent trial results the FDA pooled analyses shows that patients

with higher pdl1 expression benefit most while those that are pdl1 less than one appear not to be benefiting in summary the current

approvals of immune checkpoint Inhibitors in escc are agnostic of pdl1 status as previously stated at the time of approval of pisab in 2021 and noolab

in 2022 less was known about pdl1 as an predictive biomarker in ESC and given the exploratory nature and small sample

size of the subr FDA did not restrict labeling based on pdl1 status based on the results of each individual study however consistently across three

applications e efficacy appears to be predicted by pdl1 expression patients with tumor PDL 1110 or greater appear to have the greatest magnitude of benefit

in patients with intermediate pdl1 expression between 1 and 10 there may be a potential for benefit but those with pdl1 less than one do not seem to

benefit importantly we know immune checkpoint Inhibitors can have added toxicity and treating patients with pdl1 less than one with immune checkpoint

Inhibitors May expose them to toxicity without a clear benefit FDA has provided both pre-specified and exploratory analyses of the efficacy across a range

of L1 expression levels and stated the notable caveats FD is concerned with a lack of benefit observed in patients with escc and bdl1 less than one in

light of these findings FDA feels that this topic warrants a contemporary discussion on the risk benefit profile of immune checkpoint Inhibitors in a biomarker selected patient

population FDA would like the committee to discuss the risk and benefit of the treatment with anti pd1 antibodies for the firstline treatment of patients with

metastatic or unresectable esophagal SOS cell carcinoma for pdl1 less than one following the discussion we would like the committee to vote is the risk

benefit assessment favorable for the use of anti pd1 antibodies in the first line unresectable metastatic esophagal scamell carcinoma at pdl1 less than one

thank you thank you we will now take clarifying questions to the presenters when acknowledged please remember to state your name for the record before you speak and direct your question to a

specific presenter if you can again because we have three applicants here please uh direct your question to a specific applicant if possible if you wish for a specific slide to be displayed please let us know the slide number if possible and finally it would

be helpful to acknowledge the end of your question with a thank you and end of your follow-up question with that is all for my questions so we can move on to the next panel member for our panel member joining us virtually please use

the raise hand icon in Zoom to indicate that you have a question and we will acknowledge you please remember to look Your Hand by clicking the raise hand icon again after you have asked your question are there any clarifying

questions for the presenters that meart so this morning we spent a fair oh Jeff meart Dan haror um this morning we spent a fair amount of time talking about pdl1 ass say I'm not going to

totally bring that up again but my only question to and I'm not it's probably to any of the companies with pathology expertise um is there any differences how we think about the assay for squeam

and cell carcinoma thank you Ian Waxman BMS I'll have Dr Andrew speak to that hi Robert Anders Johns Hopkins

pathology I'm I'm compensated for my uh travel but not my time so um these are rare there are no uh St systematic

studies of interobserver agreement because they they're rare um I the caveats that I mentioned earlier uh

would all apply here um I would just make the one comment that um squamous cancers um and I'm thinking more headand

neck type squamous cancers are generally a little bit simpler to score thank you I'll also have Dr chisesi answer

this question thank you vladis cheski iic pas ologist I am a pay consultant to mer uh the sell carcinoma tends to the com the

scoring component of scell carcinoma tends to be T tpf more than cpf more tumor cell sustaining so it is simpler to score but the same variables apply

near the cut off as they are in gastric adenocarcinoma but being a TPS score is a little simpler it's more consistent in scoring thank you that's all I have

thank you this is uh Chris Le from University of Colorado and this question is to Dr Kelly but welcome to hear from any of the other uh experts uh that have

already presented uh in regards to the data that BMS showed showing that uh only 60% of patients are receiving some type of pdl1 scoring uh my question is

is the are the 40% not being scored because the approval doesn't require uh testing for pdl1 or if there's something fundamental about biopsying these patients which makes it therefore then

infeasible understanding that that 40% that's not being tested that number is always somewhat increased by the fact that patients are not getting tested sometimes because they're not eligible to receive therapy um but I would just

like to hear what the practicality is of of the biopsy uh because if this indication changes it will therefore then require biopsies for our patient population thank you very much Ron and

Kelly from Baylor University Medical Center that data is from the flat iron database which shows 40% across the country now I think it's important for us to step back and realize the proportion of patients we're talking

about it's very hard to get exact numbers for escc in the United States but I showed the steer database of 14,000 over 11 years which is about

1,200 right so we're and and the majority of these patients are seen on the coast very few are seen in the center of the country so when a patient walks into a community oncologist with

this disease they have no real prior experience of treating these patients the patients then are incredibly sick they have as you know you know more

significant dysphasia than the distal esophageal or the gastric cancers so the patients are walking in sick we know only 70% even get firstline treatment so

the doctor has an important decision to make do I start trying to biopsy a patient to get a result or do I have to start treatment straight away and I think it's clear in this instance

they're they're choosing to go for the treatment straight away and I think they're foregoing the biopsy and I think that's an important thing because as we talk about the risk benefit here for the

panel I don't want us to overlook and be so focused on biomarkers when the reality on the ground is if we mandate testing where we've never done it before we may actually decrease the number of

patients that get treated and I don't think that's in our interest thank you Natalia B University of Wisconsin um I would like to add to this I absolutely

agree with Dr Kelly that these patients are very sick and that urgent indication of treatment is indicated most of the patients all of the patients we see in clinic do have biopsies this is have the

diagnosis of cancers made all these patients have endoscopy or biopsy of metastatic site I suspect that in part that we have biomarket testing is because our treatment decisions were not

based on biomarkers so there was no reason to do the testing um I think if we actually demonstrate that we need pedon positivity in the testing and that treatment selection

would be based on biomarkers we will see a lot more testing done in the community I also think that for our Community Partners both medical oncologist and pathologist if you have unified

guidelines regarding of hology or regarding of stage and this is more for reflex testing the way we do when we look for mism repair protein expression it is rare to see it but yet we look in

all GI tumors regarding of stage it will make it easier so that neither our clinicians nor our Pathologists have to think about it look at the cut offs and figure out which antibody to use we need

to make it simple another point is all these patients can be started on chapy and then when pdl1 testing is ongoing either noolum up talism up or pism up

can be added with cycle 2 day one thank you comment from sure uh Peter and her um medical oncologist Dana Farber you

know I agree with Dr Kelly I think many of these patients are extremely sick uh their esophagus is basically failing them they're unable to eat they're losing weight they're getting very weak

we often need to move very quickly and I think the problem is that we often uh need to make a decision very quickly to start therapy because if we don't they

get weaker and then we can't they can't tolerate Platinum 5fu therapy anymore so I think that for excluding 10% of

patients from treatment uh we're really putting 90% of the patients at risk for failing and I think by uh removing this

10% group we're actually uh significantly harming the other 90% of the patients that benefit from this

immunotherapy thank you thank you Dr stanoff um a a question in a comment I'll start with a comment I'd like us all to take a step back and um

understand what flat iron data tell us right we're seeing patients not getting tested but we have to be careful not to say why they're not getting tested and if you look at biomarker testing for

lots of diseases for really important indications including I think as we heard her to testing in gastric cancer this morning it's not getting done that's a quality of care issue that's

not necessarily related to what oncologists are testing for based on their treatment decision- making so let's all be cautious about our interpretation of that um I think to

sort of respond and ask Dr enzinger actually to come back about this I think um in Dr Kelly you could comment as well uh both of you are incredibly experienced and expert clinical

trialists and very good scientists if you were writing a grant with the pdl1 negative popul

would you expect your Grant to be funded if it were investigating ongoing immunotherapy in that population and because of that should we

actually be giving it to these people we'll start with Dr anzinger well um Peter enzinger uh medical oncologist Dana har I think that's an uh an

interesting hypothetical question uh I mean I think that uh we clearly need to find better therapies for these patients who have low CPS scoring patients and

frankly I do think you should fund me if I have something that I can improve upon so I mean I think we're all we all realize that there are limitations to

checkpoint Inhibitors and we're all now trying to improve upon the outcomes that we have seen here today and if I have a treatment that theoretically is going to

be better and specifically May address these low CPS scoring patients I do think that this should be funded and I think that it actually should have a priority funding because this is a

significant unmet need thank you thank you Dr Kelly yeah Ronan Kelly Baylor University Medical Center thank you for the comments um if you actually look at

the Checkmate 648 data the four-year data is better now the hazard ratio in the less expressors is

085 so we always talk about the tail on the curve and as the data matures it seems to get better and I think that's important for us to be aware of the other thing is there's been newer trials

done we're not talking about all the trials here as you know there's been a lot in Asia in the last couple of years and they are showing in the low expressors the CPS less than once in a

sopal sth responses there was one study that Jupiter 06 with Tor palab less than one Hazard ratio 61 so we're beginning to see this data now and I think it it's

it's playing into the fact that we're seeing benefit in some of the those patients not everyone but some of those in the lower group thank you good uh Dr

Sprat thank you uh can you pull up it was one of the FDA slides number 11 got a six K Meer curves and a table

next to it uh the question will be actually for Dr eninger we can get the slide up is it from the first presentation by

drak Yep this is the first presentation slide 11 from that one not the first FDA presentation please the intro

and you know and Dr Kelly you're free to respond as well so when I look at this you know I I I hear the anecdotes of you know a patient who gets the combination

therapy and has this long response I I also see pretty much overlapping curves here with the chemotherapy arm here and I I try to think about you know the

comment you made about 90% of patients are benefiting in even in for 10% that that may not but even in the ones that are

benefiting just to be clear it's not 90% of patients are benefiting there's a small percent that are benefiting over

chemotherapy and so even in the PDL uh one less than 10 when I look at 2ye let's say we look at threee depending on which trial you look at you're talking about a couple percentage points so

number needed to treat of you know 40 to 50 and when you look at the median here of let's say pdl1 less than one in terms of

the median survival is another metric you're talking about days right and the keynote study it's identical you look at the Checkmate

study it's actually appears to be worse um and same on the last one so I I guess I'm just trying to understand what do

you mean when you say that 90% of patients are benefit from receiving um immunotherapy so I think that one of the

most exciting things about keynote 590 um was when we saw the long-term results of that study um and uh where we see

that about 11% of the patients who receive chemo and immunotherapy are alive at 5 years whereas only 3% with ch

therapy alone so I ask you as as individuals if you had esophageal sisol carcinoma would a CH a one in 10 chance

of being alive at 5 Years be worthy of consideration I completely agree however that's the overall trial and when you look here yeah you see functionally no

difference so I guess I would say would you rather take a more expensive agent that increases toxicity for the exact same probability of long-term survival

and median survival yeah I know what my answer would be yeah well I just want to add that among those long-term survivors there were patients several patients with Squam cell carcinoma and adoc

carcinoma who had CP CPS scores of less than one so thank you Dr thank you I appreciate it Dr BR uh Neil vson um this is a question

for Dr enzinger and Dr Kelly um just going back to what you were saying earlier about these patients being uh first of all this being a rare disease and these patients being quite sick and infirmed that the biopsy that even

getting a biopsy sometimes can be challenging are you referring to a second biopsy after the cancer diagnosis has been established or um uh meaning that like the first biopsy was just

inadequate to to obtain pdl1 can can you just clarify what what you mean by that yeah it's look it's I think it's it's clear we can get biopsies and patients the question is is the right

thing to do when someone comes in that's so sick who's failing who's losing weight and if we offer radiation to that patient is it safe to try to do a biopsy

post radiation that wasn't part of the trial but that's what happens in the community because the patients so sick they come in they have to give them something to Del relieve the discomfort

to open up the esophagus it's not safe to be just biopsying post radiation you know to put another end scope in so there's real world challenges to getting a biopsy in this setting which may not

have existed in gastric because it's more distal and it's not the same presentation and the same level of um morbidity that the patients are having presumably in that scenario they would

have gotten some biopsy initially to confirm diagnosis from an A mucosal reection or something like that or are you talking about people who are just being treated based on their clinical

phenotype that it's consistent they're a smoker it's assistant with S I just want to be those patients would have had an initial biopsy of some sort is that correct yeah I think you know as I said

it's Poss you can get a biopsy but then is the is the biopsy enough to make a diagnosis was it the right you know biopsy all of the issues that play into real world you know challenges that we

see every day in clinic do we have any sense oh sorry Dr en I just wanted to bring up I mean Dr Kelly brings up the Frontline patient the the patients who

present with with this disease we often have locally Advanced patients who unfortunately have recurrences so here I am 2 years later um and the only biopsy

I have is from the original diagnosis two years earlier it's an alligator clamp biopsy now two years later they

have lung Mets um and so my choice is currently is I can go to that biopsy that alligator clamp biopsy if it's positive great if it's negative I still

have approv if you remove this approval I basically now have to show that the current disease is actually CPS positive and I may actually have to do a lung biopsy

and lung biopsies as many of you know are risky right I mean people die from lung biopsies so you know I think it's worthwhile if you want to prove

recurrence or if there's an important reason but to prove that yes there's CPS positivity may not reach that level in

my mind for a lot of patients if I could I'd like to show you a slide on radiation and what it can do in this setting if we can pull up slide number one please this is from Checkmate 577

which I had the privilege of of being the global Pi for it's in a different disease setting but it was it it it shows what we think radiation is doing in this setting what we showed here was

we had matched biopsy samples pre- radiation and post radiation and you can see on the left the significant change in pdl1 CPS score post radiation this is

the only data I'm aware of in a phase 3 setting this was a retrospective analysis but it still is interesting to look at you can see

51% of the patients pdl1 score changed post chemo radiation so this is important because we talked about the dynamic nature of this this morning but we didn't look at

the biology of what we do and does it impact here you can see the TTS score is not changing we think the TTS score is driven by internal oncogenic signaling

but the CTS score is driven by cyto kind release interfer on gamma from the radiation upregulating pdl1 and if you look at slide number three which is a

busy slide but you'll see in in this particular study we showed that if you upregulate pdl1 which you can see at the

top this is CTS change the hazard ratio went to3 if for disease free survival so I'm just making the point that when we

introduce chemo radiation into these patients we altered a TDL one score and I'm just not sure it's the right thing to do to be mandating everyone have a

TDL one score for a couple of hundred patients not minimizing the number of patients but the delay in treatment is the bigger risk factor for me than getting a

biopsy yeah I I I understand what you're saying I guess what I'm trying to say is that it's I mean there's two two groups of patients there's one group of patients that had localized disease that then recur then there's a second group of patients that have denovo metastatic

disease patients with denova metastatic disease would have a biopsy to establish their diagnosis and presumably that could be tested for pdl1 of course there might be issues with that but I think I think those are two different groups of

people we don't necessarily have statistics going into that that that I'm aware of so it's uh I guess what I'm trying to say is this question about the morbidity of a biopsy doesn't that issue

doesn't apply to all patients thank you uh Dr myart Dr Kelly could you just CL if we could pull that slide up again could you clarify what is the postbiopsy because

really for most of these years met we're talking about their metastatic disease so is that a biopsy of their metastases because as you know we've all hoped and prayed that there's anos scopal effect

from radiation and we really don't see that so I'm just trying to understand what the postbiopsy is yeah great question check Mage 577 was not a metastatic study it was a it was an Aden

study so we had the sample before chemo radiation and then we had the surgical sample at the time of reection so they were the two time points we were looking at it's the only data that I'm aware of

in a phase three setting showing the impact of radiation in a softage gal cancer thank you any other questions okay we'll take a 15-minute

break panel members please remember that there will be no discussion of the meeting topic during the break amongst yourselves or with any member of the audience we will resume at 3:50 p.m. uh

3:50 p.m. eastern time thank you e e e e e

e e e e e e e e

e e e e e e e e

e e e e e e e e

e e we'll now begin the open public hearing session both the FDA and the public believe in a transparent process for information gathering and decision- making to ensure such transparency at the open public hearing session of the

advisory committee meeting FDA believes that it is important to understand the context of an individual's presentation for this reason FDA encourages you the open public hearing speaker at the beginning of your written or oral statement to advise the committee of any

Financial relationship you may have with the applicant for example this financial information may include applicants payment of your travel lodging or other expenses in connection with your participation in the meeting likewise

FDA encourages you at the beginning of your statement to advise the committee if you do not have such Financial relationships if you choose not to address this issue of financial relationships at the beginning of your statement it will not preclude you from

speaking the FDA and this committee plays great importance in the open public hearing process the insights and comments provided can help the agency and this Committee in their consideration of the issues before them that said in many instances and for many

topics there will be a variety of opinions one of our goals for today is for this open public hearing to be conducted in a fair and open way where every participant is listened to carefully and treated with dignity

courtesy and respect uh for any presenting virtually please remember to unmute and turn on your camera when your op number is called as a reminder please speak only when recognized by the chairperson thank

you speaker number one please state your name in any organization you are representing for the record you have five minutes thank you I'm Mindy Min morai I am the president and CEO and

founder of the esophageal cancer Action Network Ean um Ean does receive funding from all three of the applicants my travel and my presentation here today

was not funded by any of them um I don't envy any of you the decisions you have to make here um you're looking at statistics and numbers and that's your

job but I come to you today to represent patients and families who are struggling to stay alive in the land of very little

hope esophageal cancer accounts for 1% of cancer diagnoses in the US but 2.6% of the cancer deaths fewer than one

out of 17 stage 4 patients Will Survive 5 years that's less than 6% my husband was diagnosed with the sappal cancer when our kids were six and

11 he went through punish chemo radiation and a radical esophagectomy and then Less Than 3 months later he started having pain when he was walking pet scan showed that he

had a 6 cm tumor in his liver Mets to his pelvis ribs and lungs we tried some experimental targeted therapies but we lost him exactly six

months to the day of his esophagectomy that was in 2008 and I know what it is to stay up all night searching for a study that

might save the life of a person you love more than anything in the world you're always looking for that glimmer of hope but what I learned very quickly is that esophageal cancer patients have been

woefully neglected especially in the area of research funding and focus and that's why I started Ean in 2009 that very next

year the National Cancer Institute Drew up a list of 20 cancers for its groundbreaking genome mapping project and saage cancer wasn't on the

list Ean Ean made a very strong push and we were successful in getting esophageal cancer included in the tcga but the reason this was so important to us was

that we believ that if esophagal cancer was not included in tcga it would be like every other opportunity for Progress for our patients the train would leave the station and Es safal

cancer would be left behind and we wanted to change that Dynamic but our community had very little faith that progress was possible we started

campaigns to increase research funding and I had calls from people who said you're an idiot this is never going to happen nobody cares about as safo cancer patients and they sure aren't going to

fund any research for us that's what happens when you're up against a cancer that's often thought of as a death sentence and you live in an environment

that provides No Reason for Hope the approval of immunotherapy for our patients has been our best reason for optimism yet and I have to say that

today it's demoralizing that after waiting so long for progress and finally getting it the powers that be are considering cutting back on access to these

therapies for some of our patients approval of the proposal before you will take away hope that we've fought really hard to find

we believe that the original FDA approval is the one that's appropriate we've heard that there's no new data that shows a change in the risk

benefit equation and then Dr Kelly talked to us about more mature data that's actually showing much higher benefits and G given the significant

questions that have been raised about testing and scoring of pdl1 expression it doesn't seem to me like this question is ready for prime time I

don't know why it has to be done now I think there are a lot of reasons to wait until you have better information until we have better tests that can actually tell us what is that pdl1 score that

we're looking at is it reliable because unreliable tests should not form the basis of whether our patients have hope for the

future so I'm asking you that when you make your decision you think about whether it's worth giving up the opportunity for

somebody to save their lives possibly even for a short enough period for that next Discovery to extend their life

further um that you don't vote to extinguish their hope thank you thank you so much this concludes the open public hearing portion of the

meeting and we will no longer take any further comments from the audience the committee will now turn its attention to address the task at hand the careful consideration of the data before the committee as well as the

public comments we'll now proceed with the questions to the committee and panel discussions I would like to remind public observers that while this meeting is open for public observation public attendees may not participate except at

the specific request of the panel after I read each question we will pause for any questions or comments concerning its wording may we have slide two

please we proceed with our first question which is at discussion question FDA would like the committee to discuss the risks and benefits of the treatment with anti- pd1 antibodies for the firstline treatment of patients with

metastatic or unresectable esophageal SEL carcinoma with pdl1 expression less than one are there any issues or questions with the wording of our discussion

question seeing none uh we'll go ahead and open the question to discussion and I'll go ahead and start again and I think that many of our comments are likely going to mirror what we heard this morning uh certainly you

know what we're looking for these durable responses and the consistency that we see in the evidence of course is that in the high expressing pdl1 uh tumors we see significant benefit at 1 to 10 there's some modest

activity but plausible activity and that's certainly less than one we have the same story of lack of overall survival benefit and I think that we can take a little bit uh I guess of safety knowing that whatever decision that we

make here regarding CPS score less than one is only going to impact somewhere between 8 to 10% of the patient population so that makes me feel a little bit better the challenges we discussed significantly uh before the

pdl1 scoring but I really got to point to the fact that we're trying to assess data from a group that is CPS less than one or pdl1 less than one that has 77 in

the treatment arm and 87 patients in the in the control arm this is an an unbelievably small data sets and I have trouble determining whether or not patients should should or should not

receive therapy based off of this few patients but if you know to answer the question again you know what are the risks and benefits we just don't see any overall survival benefits so I don't

think we can make any conclusive evidence uh that it's helping patients and to the open public hearing speaker you know we hear you we want to offer these therapies to as many patients as

possible but not if there's no survival benefit uh and I'd really like Dr ioa's point of standardizing this testing making it simpler uh for practitioners

but also ensuring that the right testing gets done on the right patients and right now I believe that that's in patients whose pdl1 score is greater than one um but I really would love to

hear other people's um perspectives on this Dr sof so Chris I uh sorry Hannah stof UNC I I think I agree with everything you

said and I really want us to think about that small number so that's my biggest concern right we've seen repr reproducible evidence across a couple of Trials um but it starts to become a lot

less certain when you're talking about such small numbers um and that makes me a little more uncomfortable about these subgroup analyses I think like the prior one I don't think you can make much of

the one to 10 subgroup analyses um they're sort of all over the place and not consistent or reproducible but um we see no benefit at all we see nothing to

offer hope to our patients with a pdl1 that Zero from these data um but is that enough if you're talking about

120 130 people I I don't know the answer to that uh Dr Madden uh Robbie mad National Cancer

Institute so yeah I think I think the presentation by the FDA was appropriately cautious there was a lot of conditional language ex you know exploratory analysis May there very

very clear that we're working with very small numbers here and you know I like to flip the the equation on its head and say all right so this was a paper a grant or a

company coming to you with a very interesting subset analysis with 8 to 10% of the patients reading out in a different way you'd say that's a good hypothesis go explore that in a

subsequent trial and so that's where I'm left with I'm again left with you know I'm not see seeing clear benefit but I'm not seeing robust enough numbers that I

feel I can make an honest determin not honest honest is not the right word but a I'm not comfortable making a determination based on 8 to 10% subset

analysis thank you Dr Sprat Dan sprad uh s western yeah I mean this is I I think is

more challenging I two points just to respond to Dr Madden you know one is is these are small sample sizes but let's not forget the event rates are I mean

almost everyone is having an event so in trials that a 500 patients having a 100 events is you know many of them are powered uh for things such as that so I

think that's one point but the second is I mean to push a bit on the FDA here and this isn't the first time these primary end points for at least two of the trials right these

companies decided and designed and powered their trials based on these subsets they wanted these CPI or these various scores to be enriched for a very

specific reason a priority it was shown and I actually commend I think it's BMS who showed on you know the label literally says the has ratio I think you

highlighted was 1.0 in um the the low pdl1 expression not dissimilar to the olaparib data not

to you know rebr that but someone else brought up is that you know approval across 15 mutations when we knew a priority the investigators made a cohort

with the biomarkers that were likely to be enriched and there was no evidence of response later it seems far more practical to be giving approvals for the

primary endpoint population especially when it was known at that time when uh there's a large potential benefit and to then walk back and encourage the trial

to be done in these unknown populations because we're now in in a a stance that yes could there be some subset of some subset that benefits but

if we saw all this data initially if this was a pulled trial of all of these and the and it was a stratified analysis and you saw in these 150 or 100 patients

these kapan Meyer curves these median survival curves I I mean there's no evidence of any benefit in this population so I I

think the broad indic the broad approvals creates a very challenging time um I don't know how we'll get better data because as one of the companies

said it B became impossible to run these trials in the US because once these approvals were given than BR does the FDA have a response sure I mean we we hear you um

you know when these studies were done you know we we you know we had discussions with the companies about um you know both the biomarker positive and the IT populations I think you know learning over time we are seeing that

you know maybe we do need to push more on the on the biomar negative population to get enough patience in events in in those to get better data um you know for these studies this is what you know this

is this is the way they were done and these are the results we have it's not optimal uh we hear you and optimally we would had a better powered study for including more of the uh pdl1 low

patient um you know we are here in the situation now so we have to you know make a decision yeah this is Dan brat uh s yeah I mean it's still though it's the

primary Endo it's powered for this subset but the approval goes to the entire trial and so that's a decision that again this isn't for example the

the third company uh beige Gene I think it was it was not their primary end point was based on that but I think going to kind of the Erb jitb his

history here is if the artinb trial was initially designed in egfr mutant patients for powering this and you approved it for everyone people would be like what the heck are you doing it

wasn't and so I I I think just that that's I guess at least my uh hindsight 2020 advice which is easy to give and I I hear you and you know in this case you

know P1 is a little bit more challenging than than the you know egfr situation and we do hear you Dr Dog yeah this is a question to

the FDA statisticians did you do an interaction test looking at the effect of PDL uh the the two groups less than one and greater than one and and what

was the P value for that uh Jo fi reviewer at FDA yes we did a um analysis of the interaction in between treatment and the the pdl1

status using Cox model I can quickly take a look of the P value I think absolutely it is less than 05 only use the pdl1 less than one as

the indicator P Val is17 thank you thank you that answerers my question Sprat uh Dan sprad I mean to that point talking about sample size a lot of times people will say oh it's

it's underpowered well if you have a significant interaction test it's powered that these are significantly different effect sizes any other comments do can I just add to that

though often and I mean just to add further strength to that when we do interaction tests sometimes because we know they're historically underpowered we sometimes allow a P value point1 for

for an interaction test so a 0.01 I think to me is pretty convincing Dr V yeah just addressing um one of Dr sprat's points and I think what Dr Van L said in the last session as well is that

because this is just such a dynamic field I think you know I think the one difference between this today and you know let's say olaparib is that here we

have three different drugs that are you know um have a lot of similarities on a similar patient population and it's when we do the pooled analysis that we see these higher order effects and I I think

in some of these other drugs um uh you know we have at least here I think we as a field I think have to be prepared for the fact that if there are other drugs approved in the same setting for a

specific cancer type that do show these higher order effects where a specific um biomarker cut off does not meet the Mark um I think we just have to sort of be prepared for that that if if we have

these high quality analyses that that that may change the way we treat our patients for example there are other pd1 antibodies that are approved regardless

of pdl1 status and maybe there's going to be more data that comes out to that effect uh Miss Dayton okay Dana Dayton patient

representative so this is kind of a lay question but before the approval is pulled back from that um population that

is pdl1 less than one is there anything else that can be looked at in that population like MSI status or anything like that because there usually is a

small set a subset of people that you know do respond differently yeah it's a wonderful Point um you know the analyses

that were presented did exclude the MSI High population uh but as also was discussed uh with some of the experts that I've already presented uh I think that really is the call to action uh in

this in this subgroup that there you know certainly it's small there not a lot of patients that are pdl1 negative uh but at the same time there is a certain amount of you

know there are patients that clearly have a benefit we have no idea who those patients are and I think that really is incumbent upon us and it goes to Dr s's question about the grant that you would write you know to further investigate

this this population but it really does speak to the fact that you know we don't know everything that there is to know about biology but we should and I think that that's where the science will will certainly lead us to to better Therapies in a patient population that obviously

desperately needs better therapy thank you Dr Kakak thank you just to clarify the analysis actually included all patients

not only MSS uh but because we have very few patients that wear MSI high and we have a very high rate of unknowns we we didn't separate those having said so

this approval does not affect patients with that that are MSI High thanks I appreciate that uh that will be treated

anyway yeah thank you Dr kazak uh Dr panon um I have a I guess a comment and a question I think unlike this morning's

conversation where really the decision was um to uh align with existing guidelines

from nccn if you look at the existing guidelines for use of Neo and pemro for this particular indication they are all

over the map um currently Neo doesn't have a biomarker Neo iy doesn't have a biomarker requirement and pemro does and

so I'm guess I'm asking is the goal now that there's three drugs to have some consensus around a biomarker um indication that would

unify guidelines and actually help clinicians who very rarely see this disease you have a response your decision should be based on the data that has been

presented there's no attempt here to unify guidelines whatsoever we that is not part of our uh procedure so to speak

or our objectives here it's it's really on the data that is presented here okay uh we do not know what went in or who made the decisions in the guidelines so

there is absolutely no attempt here to provide some unification of the gu guidelines with an FDA label I I will say though when we you know when we did started doing some of

these analyses we weren't you know we we knew about like the 10 you know with one of the drugs and you know Cate category one recommendations and we we didn't want to be set on like okay that's what we have to focus on we want to look at

the totality of the data because really if if patients you know when we look at the pool analysis between one and 10 well the the you know again it's it's exploratory and you take it with um some

caveats but the the uh Force plots were you know clearly to the to the left of one and so you know we wanted to make sure that if if there was going to be you know a limitation we wanted not to

like um you know we wanted not to you know wanted to be on as few patients as possible that you know if we have an effect we want to make sure those patients get that so we didn't want to

be limited to um what was in the you know in guidelines uh thank you uh Dr Madden looking back through my slides

here so your interaction score which I'm not said so it's good to try to understand it better but that's with all three trials pulled together correct review at FD yes we the DAT to

do that analis if I remember correctly maybe D what was that slide you had pulled up before Dr sprad uh maybe that's it was 11 of the fda's talk the first one the intro trying to remember

because it look like I mean not all the curves are exactly similar if we look at that and so I mean our question here is focused on three trials so you know some of them

cross some of them don't and so you could understand where the interaction score tells part of the story but maybe not all the story for all the trials

that would just be my kind of naive perspective as a non-statistician good uh

Dr I mean just to be clear obviously you did the analysis but what the inter action is is looking at The Columns of uh well actually this isn't the slide

that would even have it but it's basically the pdl1 less than one and I don't know there's probably a slide that has a pdl1 greater than one it's comparing those two um Hazard ratio so

it's not comparing these curves really that you're seeing here so it's haard ratios of basically 6ish versus Hazard ratios of about one is there a significant difference in those relative

benefits is the intera treatment biomarker interaction yeah slide 28 from my slide deck that was

presented can we have slide 28 from Dr shasta's presentation please and for rational those are reversed uh Dr Dodd yeah I was also I

think for me slide 30 is a little more um instructive just showing you know it's basically testing the the rows for the yeah the PDL lesson one and the PDL

greater than one is that has you know is there evidence to suggest that there's a treatment effect difference between those two groups and according to the statistical criteria yes there is so

it's the 1.1 versus 68 roughly speaking am I correct FDA that oh yes that's correct thank you

uh Dr meart Jeff meart Dana Farber um you may have not done this but um this the interaction term that we've been talking about how much of it's driven by

the telis um because that abviously the Haz a ratio I know the numbers will get much smaller and the power you'll have much less power and it's only like 50 per each arm if you just take the Neo and the pemro but the chisl map given

that the hazard ratio is 1.34 is probably driving some of that interaction more than others so I don't know if you just did sort of Just The Two uh groups that uh and see

what that interaction looked like uh Jo sta review at FDA for that analysis we use the pool data including the data from three tril oing but here

uh we uh observed the consistent pattern across the three studies with the pdl1 less than one subgroup thank you any additional comments or

discussion questions or points I'll try to summarize this uh this discussion so I think unlike the discussion that we had this morning um

while I believe that there's consensus in regards to some concern about how the CPS less than one or pdl1 less than one group is doing uh with the treatment uh

I think the entire group is really concerned about the sample sizes that were being discussed I think that statistically that makes things very difficult so we had a lot of conversation about uh the analyses and

the interactions uh performed on the poed analysis having said that I think that maybe the differences in our discussion here is that you know now that we have a pulled analysis which is

something that we haven't had at least on a trial level uh with the individual trials um we can start to make some conclusions but I hear a little bit of

uh discomfort uh in regards to making uh a determination based off of the current data that we have uh in hand any additional comments or questions to that

okay so moving on uh we will now proceed to question two which is a voting question it'll be using an electronic voting system for this meeting once we begin the vote the buttons will start flashing and will continue to flash even after you have entered your vote please

press the button firmly that corresponds to your vote if you are unsure of your Vote or you wish to change your vote you may press the corresponding button until the vote is closed after everyone has completed their vote the vote will be locked in the vote will then be on the

screen the DFO will read the vote from the screen into the record next we will go around the room and each individual who voted will state their name and vote into the record you can also State the reason why you voted as you did if you

want to we will continue in the same manner until all questions have been answered or discussed so for question two which is a voting question is the risk benefit

assessment favorable for the use of anti- pd1 antibodies in in first line unresectable or metastatic esophageal

sell carcinoma with pdl1 expression less than one are there any questions or comments concerning the wording of the

question seeing none uh we will now begin the voting process [Applause] for there is one yes 11 Nos and one

abin now that the vote is complete we'll go around the table and have everyone who voted State their name and their vote and if you want to you can State the reason why you voted as you did into

the record and we will start with Dr Van L this is Katherine vanloon and my vote was no uh I think we'll go ahead and go

around the table from the other side so Dr s off Hannah s off my vote was no the reason being that though we heard a lot about how inflamed these tumors are in

their micro environment um there seems there must be something different about the pdl1 negative patients given the completely overlapping survival curves

and even some suggestion of uh you know no benefit at all even potential harm so I just did not see enough evidence to suggest and despite the small sample

size um I think uh the the fact that the effect of treatment is statistically Modified by pdl1 is pretty compelling as well as the event

rates thank you Dr meart uh Jeff myart Ana Farber um my vote was no um pretty much the similar reasons I mean I think the consistency across the three trials where the hazard ratio

really is one or even higher um despite the small sample size uh May me aboute no Dr

mcken hiy M Keen bar Cancer Institute my vote was no U my honest answer to that question is the risk benefit assessment favorable no despite the small numbers I

think this is the best uh data set we're going to get Dr Gibson I'm Michael Gibson I also answered no I would

like um just for the record to state that the data as I see it now is uh and described and commented on by my

colleagues uh stands alone but I I know that I will also have to take this decision into consideration when I'm in clinic uh and I appreciate the

opportunity to um to be a part of the decision thank you thank you Dr Hawkins uh yes uh voted no and with some apprehension about removing something that was available

but could not ignore the numbers I suspect uh as my colleague D may be mentioning some doctors will probably still attempt to offer this to patients

who feel that I'm willing to take a get take a chance give the opportunity to continue to use the drug irrespective of their low uh pdl1 number thank you Dr

Hawkins Miss Dayton Dana Dayton patient representative um I understand the numbers and I understand the concerns the numbers are low but as

a patient I also had no options and had to fight for you know anything I could find and I do believe in the Hail Mary

passes and sometimes things do work and you don't know why so that's why I voted yes thank you Dr DOD yeah so uh this is Lori Dodd and I

Ved no um the question stated is a risk benefit profile favor you know the use of this and PDL uh one less than one there was no evidence of benefit in this

group um and in particular there was um evidence of a clear treatment by pdl1 status interaction um and so you know in spite of that I I mean I I think that

that there needs to be more done um and I also tried to approach this as if we were looking if we were presented with these data to noo today would we approve this in this group and there's just

simply not evidence to suggest that it would provide a benefit in this uh subgroup thank you thank you Dr DD Dr basson Neil voson Columbia um I voted no

based on the the totality of the data that despite these small numbers that this was a statistically sound analysis um I I will say that I think today's discussion both in the morning and the afternoon has really been quite

instructive and I think also shows us the importance of well some people may accuse um oncology of having so-called me to drugs these these multiple pooled analyses actually allow us to refine a

very complex biomarker that really could um help us better identify who's really responding to these drugs thank you thank you says Chris Lou I voted yeah no

I voted no uh I I think um I agree with the statement that this is probably the best data that we're going to have some of this is related to how the question was asked and you know it really asked is the risk benefit profile in favor of

and I'm not really sure you can look at this data and say that the answer is yes having said that I do share everybody's concern that this data set is still quite small um but I am thankful that

again this is a a minority of the patients that are going to be treated mad excuse me Robbie Madden National Cancer Institute I I voted to abstain um

again it's actually not that different than Dr Lou's comments it's just from my perspective I don't think that I feel comfortable with 8 to 10% of the patients from these trials that didn't

all have like Universal readouts in terms of the curves and things that were pulled together you know I know there's unique opportunities here to look at this but to to you know to use a legal

term I'm not sure there was enough to convict uh so to speak um so that's why I abstained I just don't think the question could be answered in a

comfortable way based on the data we saw thank you Dr Madden Dr Sprat Dan Sprat you and Case Western um I voted

no it's really challenging especially from you know some of the comments we've heard is you know does this which again we're not voting for regulatory change

that's not what ODC does is justes this takeaway hope and I think often we forget and it's been brought up before that pretty much every single therapy

causes toxicity every single therapy especially when we're talking about you know these agents can cause substantial toxicity or financial toxicity if you

have it and so even the numbers will just take when you say it's going to benefit one in 10 which is a phenomenal and many first you know we haven't really said these drugs are amazing

drugs these drugs have changed many patients lives so we're talking about a subset is potentially harming nine patients does that always justify

helping one and in this case it might be harming 99 patients to maybe help one and so if you're that one that's helped that's great but if you potentially are harming many

more I think that's something that I'm not sure um this to the question this risk benefit ratio is we've seen evidence

that there's going to even be that one that we could clearly reliably uh identify so that's why I voted no but again I don't think any of us want to

take hope away we don't want to harm patients we want to help them thank you Dr Sprat Dr gratar uh succinctly Bill grad is sure

Northwestern succinctly uh despite the concerns about small numbers there was no evidence of benefit

period and just for the record you voted no okay to summarize the discussion uh certainly I think that there is fairly

uh widespread consensus across the panel uh that the risk benefit profile does not favor to use in uh tumors and and patients that have pdl1 score less than

one um I think that there's fairly good consensus that um the available data do not support its use uh having said that I think that we've heard from the panel

members that there's some discomfort with the small amount of patients that were trying to make decisions uh from and and understanding that the data is never always you know going to be perfect and this is a very good example

of that but we have to make decisions based off of the data uh that we have um I think that we uh you know heard from a state and this desire to offer as many therapies as possible to our patients

and and I think that we all feel that uh at the same time uh to Dr sprat's Point uh there's certainly a number needed to treat to help people but there's also a number needed to harm uh where when you

treat enough patients you are going to deliver harm from some of these therapies and therefore I think that that's really what kind of sums up the risk benefit profile and the reason the vote kind of came out the way it did any

other questions or comments okay I know that this was an incredibly long day and I will tell you having uh three applicants in a room uh the presentations from the applicants

and the FDA were outstanding the information uh that was provided was outstanding and so really you know from the panel to the applicants and the FDA thank you so much uh also for all the

incredible work that has gone into all of this uh to our open public hearing speakers thank you for sharing your stories please know that those impact us as well uh and encourage us to do even

better in clinic and for our patients and the research that goes on uh in this room and then obviously beond uh to the FDA do you have any final

comments okay and sorry just thank for you know reiter your your thanks and appreciation for everyone that you know their hard work and the travel and and time to come out and and participate in

this meeting wonderful and with that we'll adjourn and thank you guys so much appreciate it safe travels e