sNDA

1
sNDA 20-637GLIADEL WAFER(Polifeprosan 20
with Carmustine Implant)APPLICANT GUILFORD
PHARMACEUTICALS

• ODAC December 6, 2001
• Medical Reviewer Alla Shapiro, M.D.,
  Ph.D.
• Statistical Reviewer Ning Li, M.D., Ph.D.

2
FDA Presentation

• Regulatory history Dr. Alison Martin
• Study T-301 Dr. Alla Shapiro Dr. Ning Li
• Review Issues Dr. Alison Martin

3
Regulatory History

• Approved June 14, 1996
• Indication
• Adjunct to surgery to prolong survival in
  patients with recurrent GBM for whom surgical
  resection is indicated

4
Regulatory History

• Primary Basis of Approval
• North American trial, 8802
• multicenter, randomized, placebo-controlled
• 22 pts with recurrent malignant glioma
• subgroup analysis in GBM ? OS and 6 mo.
  survival
• Supportive
• Scandinavian trial, CL-0190
• early closure (32/100) randomization imbalance
• not sufficient to support a new indication

5
New Indication

• Meeting with FDA 1/30/97
• Single trial can support a new indication if...
• multicenter
• results are robust
• Population of interest GBM
• Placebo wafer accepted
• Standardization of subsequent treatments
• Prospective definition and collection of AEs

6
New Indication

• Sample size (90 power) based on 12 month
  survival rate for Gliadel of 70 vs. 50
• FDA 8/22/97
• overly optimistic
• if 62.5 vs. 50, 53 power
• Amendment 3/18/99
• IDMC review of blinded data overly optimistic
• accrual ? from 200 to 240
• preserve ability to distinguish 68 vs. 50

7
Clinical ReviewStudy T-301
8
GLIADEL Wafer Proposed Indication

• GLIADEL Wafer is indicated for use as a
  treatment to significantly prolong survival and
  maintain overall function (as measured by
  preservation of Karnofsky Performance Status) and
  neurological function in patients with malignant
  glioma undergoing primary and/or recurrent
  surgical resection.

9
Study Objectives

• Primary
• - survival (ITT)
• Secondary
• - survival in the GBM subgroup
• -adjusted analyses for KPS, age and tumor type
• -1-year survival (ITT and GBM)
• - progression-free survival (ITT)
• - survival censoring patients with reoperation
  for
• disease progression (ITT)
• - KPS scores (ITT)
• - QoL (ITT)
• - neuroperformance measures (ITT)

10
Phase 3 Trial Study T-301

11
Study Enrollment

• Patients
• 240
• 14 countries
• 48 patients in France
• 44 patients in Germany
• 38 centers
• 7 centers in France and 5 in Germany
• 5 centers in the US accrued 12 patients

12
Baseline Characteristics
13
Tumor Histology
14
Protocol-Specified Treatments Radiation
Therapy Chemotherapy Other Treatment
for Disease Progression
15
Chemotherapy

• GLIADEL
  Placebo
• N 120 ()
  N120 ()
• Chemotherapy for PD 14
  (11.7) 14 (11.7)
• Within 30 days
• AOD - 11 1
  (0.8) 1 (0.8)
• AOA - 11 2 (1.7) 2 (1.7)
• TOTAL 17 (14.2)
  17 (14.2)

16
Other Treatment for PD
17
Efficacy Results
18
Primary Efficacy Endpoint Overall Survival

• GLIADEL Placebo
• 88 (73.3) 93 (77.5)
• months 13.9 11.6
• 95 CI (12.1 -15.3)
  (10.2-12.6)
• p 0.08 log-rank (protocol-specified)
• p 0.07 log-rank (stratified by center)
• p 0.03 log-rank (stratified by country)

19
Primary Efficacy AnalysisStatistical Issues
20
Primary Efficacy Analysis

• Sponsors Protocol/SAP
• Two groups will be compared with a log-rank test
  for the primary comparison for overall
  survival(OS).
• A log-rank stratified on each prognostic
  covariate (KPS, AGE, GBM/non-GBM, and COUNTRY),
  will be performed as secondary efficacy
  analysis and is considered as supportive..

21
Overall Survival Curves for Study T-301
22
Primary Efficacy Analysis(Protocol Specified
OS ITT Analysis)
23
Secondary Analysis Stratified Log-rank
Tests(Protocol Specified OS ITT Analysis)
24
Survival Analysis Cox Model Adjusting for
Covariates
25
Survival Analysis Summary

• Results of the survival comparison between the
  two arms are not significant except the analysis
  stratified by COUNTRY
• Sponsor presented the log-rank test stratified by
  COUNTRY as the primary analysis and concluded a
  significant survival benefit
• Sponsors argument stratified analysis is
  appropriate because randomization list was
  stratified by country, over-stratification by
  Center

26
Issues

• Stratified vs. non-stratified analysis which one
  is more appropriate?
  FDA Either one is
  acceptable but need to pre-specify one in the
  protocol as the primary analysis. Retrospective
  selection will inflate the Type I error.
• The randomization was stratified by CENTER.
  

27
Randomization List for All U.S. Sites
28
Exploratory Analysis Log-rank Stratified by
CENTER
29
Primary Efficacy AnalysisSummary of Statistical
Issues

• Protocol specified analysis for overall survival
  was not statistically significant (p0.08).
• The log-rank test stratified by CENTER and all
  other stratified analyses demonstrated N.S.
  results.
• Log-rank test stratified by COUNTRY (p0.03) is
  questionable as the primary analysis.

30
Secondary Endpoints
31
Secondary EndpointSurvival in GBM

• GLIADEL Placebo 79 (78.2) 85
  (80.2)
• months 13.5 11.4
• 95 CI (11.4 -14.8)
  (10.2-12.6)
• p 0.20 log-rank (protocol-specified)
• p 0.16 log-rank stratified by center
• p 0.09 log-rank stratified by country

32
Secondary Endpoint 1-Yr Survival

• ITT population
• GLIADEL Placebo
  
• () 59.2 49.6
• 95 CI (50.4 - 68.0) (40.6 -
  58.6)
• Not significant by log-rank, unstratified/stratifi
  ed
• GBM subgroup
• GLIADEL
  Placebo
• () 57.4
  48.6
• 95 CI (47.8 - 67.1) (39.0
  - 58.1)
• Not significant by log-rank, unstratified/stratif
  ied

33
Secondary Endpoint Progression-Free Survival

• Sponsors conclusion median 5.9 months in both
  treatment groups
• p 0.90 Log-rank, stratified by country
• FDA Limitations of the endpoint

34
Secondary EndpointTime to KPS Deterioration

• GLIADEL Placebo
• months 11.9
  10.4
• 95 CI (10.4 - 13.7) (9.5 - 11.9)
  
• p 0.11 log-rank (protocol-specified)
• p 0.27 log-rank (stratified by center)
• p 0.05 log-rank (stratified by country)

35
Secondary Endpoint Time to KPS Deterioration
(Death not account as an event)

• GLIADEL Placebo
• Months 18.3
  17.9
• 95 CI (14.3 - 24.0) (17.2- 24.0)
  
• p 0.61 log-rank
• p 0.47 log-rank (stratified by center)
• p 0.41 log-rank (stratified by country)

36
Secondary Endpoint QoL

• QoL EORTC and Brain Cancer Module
• Global health questions 29 30
• Not powered to show significant differences

37
Time to Neuroperformance Deterioration (ITT
Population)
38
Time to Neuroperformance Deterioration (ITT
Population)
39
Safety Results

• Deaths within 30 days
• Local Complications

40
Cause of Deaths in the First 30 Days of
Randomization
41
Local Complications
42
Review Issues
43
Single Trials Evidence of Effectiveness

• Adequate and well-controlled
• large, multicenter trial
• consistent results across subsets
• multiple endpoints involving different events
• statistically persuasive
• Independent substantiation from related study
  data
• other phases of disease, populations
• strong prior
• FDA Guidance 1998

44
T-301 Single Trial

• Strengths
• multicenter
• placebo wafer (efficacy)
• blinded pathology review
• survival 1o endpoint
• point estimates of median survival
• prior approval

• Weaknesses
• placebo wafer (toxicity)
• 1o endpoint stat persuasive?
• stratified logrank, multiple analyses, interim
  look?
• GBM - bridging population
• not statistically persuasive
• prior approval
• malignant glioma NS

45
Can the trials presented in 1996 be considered
confirmatory?
46
Primary Basis of Approval North American Trial

• Recurrent malignant glioma
• 222 patients entered
• 145 (65) with GBM
• 77 (35) with other histologies - balanced

47
North American Trial
48
Overview of Median Survival in ITT Population
49
Supportive Scandinavian Trial

• Newly diagnosed malignant glioma
• 32 patients of an intended 100
• Overall survival (log-rank)
• Early closure due to lack of drug supply
• ODAC Insufficient to support an indication

50
Scandinavian Trial
51
Summary of Review Issues

• T-301 Adequacy of single trial?
• robustness of survival data
• support from secondary endpoints
• clinical meaning of an analyses significant
• only when stratified by country
• Are other trials confirmatory?
• CL-0190
• 8802
• Favorable riskbenefit ratio?

52
Questions for the Committee

• 1. Is Study T-301 an adequate and
  well-controlled trial?
• 2. Gliadel was considered to have a treatment
  effect only in patients with recurrent GBM and
  not in the overall population with recurrent
  malignant gliomas. If this was a true treatment
  effect, would this pattern be expected to be
  present in newly diagnosed patients?
• 3. Do the data from T-301 provide substantial
  evidence of a survival benefit of Gliadel in
  patients with newly diagnosed malignant glioma?

53
Questions.

• 4. If the answer to 3 is YES, do the North
  American and/or Scandinavian Trials together with
  T-301 provide convincing evidence of a survival
  benefit in patients with newly diagnosed
  malignant glioma?
• 5. Is the toxicity profile of Gliadel acceptable
  for patients with newly diagnosed malignant
  glioma?
• 6. Does the committee believe that Gliadel
  provides clinical benefit with an acceptable
  safety profile in patients with newly diagnosed
  malignant glioma?

54
(No Transcript)