GLIADEL

1
GLIADEL WAFERsNDA 20-637s

• Guilford Pharmaceuticals Inc.
• Baltimore, Maryland

2
GLIADEL Wafer Indication
GLIADEL is indicated for use as an adjunct to
surgery to prolong survival in patients with
recurrent glioblastoma multiforme for whom
surgical resection is indicated.
3
Clinical Trials Recurrent Malignant Glioma
4
Clinical Trials Newly DiagnosedMalignant Glioma
5
GLIADEL Approvals (1996-2000)
Newly Diagnosed and Recurrent Canada Recurrent
France Argentina Austria Brazil Chile Colum
bia Germany Greece Hong Kong Israel
Ireland Luxembourg Malaysia The Netherlands
New Zealand Peru Portugal Singapore South
Africa South Korea Spain United Kingdom
Uruguay
6
Phase III Multicenter Trials of GLIADEL Wafer in
Newly-diagnosed Malignant Glioma

• 0190 Trial Interstitial Chemotherapy for
  Malignant Glioma A Phase III Placebo-controlled
  Study to Examine the Safety and Efficacy of
  GLIADEL Placed at the Time of First Surgery
• T-301 Trial A Phase III, Multicenter, Randomized
  Double-Blind, Placebo-Controlled Trial of
  Polifeprosan 20 with Carmustine 3.85 Implant in
  Patients Undergoing Initial Surgery for
  Newly-Diagnosed Malignant Glioma

7
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.

8
Agenda

• Introductions
• Louise Peltier, Senior Director, Regulatory
  Affairs, Guilford Pharmaceuticals Inc.
• Overview of Primary Malignant Glioma Clinical
  Features and Treatment
• Allan Hamilton, M.D., Professor and Chairman,
  Department of Neurosurgery, University of Arizona
  School of Medicine
• Phase III Trials (T-301and 0190)
• Dana Hilt, M.D., Vice President of Clinical
  Research, Guilford Pharmaceuticals Inc.
• Statistical Analytic Methods
• Steven Piantadosi, M.D., Ph.D., Professor and
  Director, OncologyBiostatistics, Johns Hopkins
  University School of Medicine
• Phase III Trial (T-301) Efficacy and Safety
  Results
• Dana Hilt, M.D., Vice President of Clinical
  Research, Guilford Pharmaceuticals Inc.

9
Guilford Invited Guests

• Henry Brem, M.D.Harvey Cushing Professor of
  Neurosurgery and Oncology Chairman, Department
  of Neurosurgery Johns Hopkins School of Medicine
• Henry Friedman, M.D.Professor and Director,
  Neuro-oncologyDuke University School of Medicine
• Janet Wittes, Ph.D.PresidentStatistics
  Collaborative

10
Overview of Primary Malignant Glioma Clinical
Features and Treatment

• Allan Hamilton, M.D. Professor and Chairman,
  Department of Neurosurgery, University of Arizona
  School of Medicine

11
Primary Malignant Glioma

• Incidence
• Approximately 16,500 new cases annually
• Glioblastoma multiforme accounts for
  approximately 75 of patients
• More than 13,000 deaths annually

Central Brain Tumor Registry US Statistical
Report 1992-1997
12
Primary Malignant Glioma

• Presentation headache, seizure or new
  neurological deficit. Average age at onset 55
  60 years.
• Imaging (CT or MRI) is key in provisional
  diagnosis.
• During surgical resection a provisional or
  tentative diagnosis is made based on
  intra-operative pathology.
• Final pathologic diagnosis requires fixed tissue
  examination.

13
Primary Malignant Glioma

• Treatment
• Maximal surgical resection followed by radiation
  therapy /- chemotherapy1
• Complete surgical resection of high grade tumor
  difficult
• Majority of tumors recur within 2 cm of original
  resection site2
• Carmustine (BCNU) is the most widely studied
  chemotherapeutic agent

1Natl Comprehensive Cancer Network Guidelines
2000 (NCCN) 2Hochberg FH, et.al. Assumptions in
the radiotherapy of glioblastoma. Neurology
198030907-11
14
Glioblastoma Treatment Outcome
McDonald JD, Rosenblum ML In Rengachary SS,
Wilkins RH, eds. Principles of Neurosurgery.
St Louis, MO Mosby-Wolfe 1994 chap 26.
15
Natural History of High Grade Glioma Effects of
PCV Chemotherapy

• Randomized, prospective study of surgical
  resection and radiotherapy (RT) vs. surgical
  resection, radiotherapy, and PCV chemotherapy
  (RT-PCV) in high grade glioma patients
• MRC, 15 centers in the UK
• 674 patients enrolled
• RT (n339)
• RT-PCV (n335)
• GBM Histology 76

MRC Brain Tumour Working Party J Clin Oncol
19(2) 509-518 (2001).
16
Natural History of High Grade Glioma Effects of
PCV Chemotherapy
p 0.50
MRC Brain Tumour Working Party J Clin Oncol
19(2) 509-518 (2001).
17
Prognostic Factors Primary Malignant Glioma

• Prognostic Factors shown to influence survival
• Age (gt60 years)
• Karnofsky Performance Score (lt70)
• Tumor histology
• Proposed Prognostic Factors
• Size of tumor
• Extent of resection

Burger PC, et al., Cancer 591617-1625,
1987 Ammirati M, et al., Neurosurgery 21201-206,
1987
18
Limitations of Systemic BCNU

• Rapidly cleared with t½ 15 minutes1
• Limits exposure of tumor cells to BCNU
• High doses required for adequate CNS levels
• Achievable dose limited by systemic toxicity

1Wang CH, et.al. The delivery of BCNU to brain
tumors. J Control Release 19996121-41
19
The GLIADEL Wafer

• Biodegradable polyanhydride copolymer containing
  7.7 mg BCNU/wafer
• Circumvents limitations of systemic BCNU
• Local delivery of BCNU over 2-3 weeks at high
  tissue levels
• No detectable systemic BCNU levels
• No systemic BCNU toxicity
• No additional surgical intervention required

20
Phase III Trial of GLIADEL Wafer in Newly
Diagnosed Malignant Glioma
21
6 Month Survival Recurrent GBM (8802)
100
90
80
70
GLIADEL
60
Survival Rate ()
50
40
Hazard Ratio .57 95 CI 0.36
0.89 Risk Reduction 43 P 0.02
Placebo
30
20
10
0
0
1
2
3
4
5
6
Months From Implant Surgery
Brem H, Plantations S, Burger PC, et al.
Placebo-controlled trial of safety and efficacy
of intraoperative controlled delivery by
biodegradable polymers of chemotherapy for
recurrent gliomas. Lancet. 19953451008-1012.
22
Overall Survival (ITT) Primary Malignant Glioma
(0190)
Hazard Ratio 0.37 95 CI 0.17
0.82 Risk Reduction 63 P 0.01
23
Clinical Experience To Date

• More than 6000 patients have been treated with
  GLIADEL Wafer to date
• Well tolerated with attention to
• Post-operative management of cerebral edema
• Water tight dural closure
• Post-operative anticonvulsant medication

24
Rationale for Phase III Study T-301

• Confirm safety efficacy results of 0190 Study
• Fully define safety profile in primary surgery
• Determine extent of clinical benefit on
• Survival
• Maintenance of KPS status and neuroperformance
• Time-to-progression

25
Phase III Multicenter Trials of GLIADEL Wafer
in Newly-Diagnosed Malignant Glioma

• Dana C. Hilt, M.D.Vice President of Clinical
  Research, Guilford Pharmaceuticals Inc.

26
Phase III Multicenter Trials of GLIADEL Wafer in
Newly-diagnosed Malignant Glioma

• 0190 Trial Interstitial Chemotherapy for
  Malignant Glioma A Phase III Placebo-controlled
  Study to Examine the Safety and Efficacy of
  GLIADEL Placed at the Time of First Surgery
• T-301 Trial A Phase III, Multicenter, Randomized
  Double-Blind, Placebo-Controlled Trial of
  Polifeprosan 20 with Carmustine 3.85 Implant in
  Patients Undergoing Initial Surgery for
  Newly-Diagnosed Malignant Glioma

27
Study 0190 Trial Design

• Primary malignant glioma patients
• Surgery
• Placebo or Gliadel Wafers
• Radiotherapy
• Study was conducted at four centers in Finland
  and Norway
• Primary efficacy endpoints
• 12-Month survival
• 24-Month survival

28
Study 0190 BaselinePatient Characteristics
29
Overall Survival (ITT) Primary Malignant Glioma
(0190)
Hazard Ratio 0.37 95 CI 0.17
0.82 Risk Reduction 63 P 0.01
30
Study 0190 Overall Survival Adjusted for
Prognostic Factors - GBM Patients1
Hazard Ratio 95 CI P-Value
GLIADEL vs. Placebo 0. 27 0.10 - 0.71 0.008
KPS 0.96 0.93 0.99 0.02
Age 1.08 1.01 - 1.14 0.02
1 Valtonen et al., Neurosurgery 41(1) 44-49,
1997
31
Study 0190 Efficacy Conclusions1

• GLIADEL, in conjunction with surgery and
  radiotherapy, decreases the risk of death by 63
  in patients with newly diagnosed malignant glioma
• Trial was positive in overall (ITT) population
• Trial was positive in GBM patients when
  accounting for all major prognostic factors

1 Valtonen et al., Neurosurgery 41(1) 44-49,
1997
32
Study T-301 Objectives
To determine the efficacy and safety of
(GLIADEL Wafer) implants plus surgery and
limited field radiation therapy compared to
placebo implants plus surgery and limited field
radiation in patients undergoing initial surgery
for newly-diagnosed malignant glioma.
33
Study T-301 Trial Design

• Randomized, double-blind, placebo- controlled
  study
• Primary Efficacy Endpoint
• Overall Survival - All Patients Randomized (ITT)
  by the Kaplan-Meier method 12 months after final
  patient was enrolled
• FDA fully informed prospectively of study design
  and analysis

34
Study T-301 Trial Design

• Secondary Efficacy Endpoints
• Overall Survival - GBM Patients
• Karnofsky Performance Decline, Neuroperformance
  Decline, Progression-Free Survival, and Quality
  of Life Evaluation

35
Study T-301 Clinical Sites
A total of 42 sites in 14 countries participated
in the study
Australia
3 sites
Italy
3 sites
The Netherlands
Austria
1 site
2 sites
Belgium
2 sites
New Zealand
1 site

7 sites
Spain
3 sites
France
Switzerland
5 sites
2 sites
Germany
United Kingdom
Greece
1 site
4 sites
Israel
3 sites
United States
5 sites
36
Study T-301 Inclusion Criteria

• 1. Male or female, aged between 18 and 65 years
• 2. Radiographic evidence on cranial MRI of a
  single contrast-enhancing unilateral
  supratentorial cerebral tumor
• 3. Surgical treatment within two weeks of the
  baseline MRI scan indicated
• 4. Karnofsky Performance Score of 60 or higher
• 5. No previous treatment for suspected primary
  malignant glioma

37
Study T-301 Baseline Characteristics
38
Study T-301 Baseline Characteristics
39
Study T-301 Tumor Size
Comparability at baseline p-value lt 0.05
40
Statistical Methodology

• Steven Piantadosi, M.D., Ph.D.
• Professor and Director, Oncology Biostatistics
• Johns Hopkins University School of Medicine

41
Outline

• Key design features to reduce bias
• Pre-specification of analysis
• Use of stratification
• Control of prognostic factors
• All the analyses are pre-specified per protocol

42
Features to Eliminate Bias in T-301 Study

• Placebo-controlled, double-masked
• Stratified blocked randomization within center
• Study also blocked by country because center is
  nested within country
• Study not blocked by histological type, age, or
  Karnofsky Performance Score (FDA review Page 37)
• Pre-specified analyses

43
Key Pre-specified Features in SAP

• Overall Survival estimated by Kaplan-Meier method
• Treatment differences assessed by logrank test
  and proportional hazards model
• Pre-specified covariates
• Age
• Karnofsky performance score
• Tumor type
• Country of treatment

44
Approach to Analysis

• All analyses were performed by Steven Piantadosi,
  M.D., Ph.D.
• Review of protocol and SAP before acquiring the
  data from the sponsor
• No contact with sponsor before discussion of
  study results
• Initial analysis used stratified (by country) log
  rank test and countries with small numbers of
  patients were pooled together
• No post-hoc analyses conducted

45
Stratified Analysis

• The T-301 study was conducted using stratified
  blocked randomization by clinical center, as is
  typical with such trials. This explicitly
  acknowledges center as a source of variation, and
  requires the use of a statistical test that
  accounts for the stratification, i.e, the
  stratified logrank test.

46
Stratification of Clinical Trials

• Treating known sources of variability as unknown
  sources of noise is to be avoided
• Simon R (1980), Cancer Treat Rep 64 405-410
• Fleiss JL (1986), Controlled Clinical Trials
  7267-275
• Localio AR (2001), Ann Int Med 135112-123
• Over stratification in the extreme becomes
  equivalent to no stratification at all
• Simon R (1980), Cancer Treat Rep 64 405-410
• Simon R (1982), Br J Clin Pharm 14473-482
• Limited stratification is generally desirable to
  increase the sensitivity of the trial,
  over-stratification can be detrimental to a trial
• Simon R (1982), Br J Clin Pharm 14473-482

47
Stratification by Center vs. Country

• Stratified randomization within center also
  creates stratification by country
• Treatment practices are likely to vary more
  between countries than within centers within a
  country
• The protocol and SAP pre-defined country as an
  effect that might need to be controlled
  (covariate or stratification factor)
• Stratification by 38 centers is almost equivalent
  to not controlling for center or country, because
  the sizes of the strata are too small
• Stratification by country (pooling countries with
  a small number of patients) appropriately
  controls this extraneous variation

48
Effect of Country-of-Treatment on Survival
Placebo-Group (T-301)
49
Overall Survival
50
Assessing the Influence of Prognostic Factors on
Survival

• A priori identification of prognostic factors
•  Univariate regression was first used to identify
  the important prognostic factors (defined as
  p-value ? 0.05)
• In order to account for the effects of these
  significant prognostic factors on survival, a
  backward elimination method (stepdown method) of
  multiple regression using the proportional hazard
  model was used
• FDA analysis on Page 39 is misleading

51
Univariable Prognostic Factors1
1 Cox Model stratified by country with single
covariates
52
Multivariable Proportional Hazards Analysis (ITT)
53
Summary of Statistical Methodology

• T-301 provides, by design, unbiased, precise
  estimates of treatment effect. It is adequate
  and well controlled
• All of the analyses presented are rigorously
  prespecified
• The use of stratification by the sponsor is
  correct and consistent with standard statistical
  practice
• The GLIADEL treatment effect (risk reduction of
  30) is clinically significant and convincingly
  independent of prognostic factors

54
Phase III Multicenter Trials of GLIADEL Wafer
in Newly-Diagnosed Malignant Glioma

• Dana C. Hilt, M.D.Vice President of Clinical
  Research, Guilford Pharmaceuticals Inc.

55
Study T-301 Overall Survival Analysis (ITT)
Hazard Ratio 0.71 95 CI
0.52 0.96 Risk Reduction 29 P 0.031
1 Stratified by country
56
Overall Survival Adjusted for Prognostic
Factors1 (ITT)
Hazard Ratio 95 CI P-Value2
GLIADEL vs. Placebo 0.72 0.53 - 0.98 0.03
KPS lt70 vs. KPS gt70 1.93 1.37 - 2.72 0.0002
Age gt60 vs. lt60 1.73 1.24 - 2.42 0.001
1 Adjusted for age, tumor type, and Karnofsky
Score 2 Stratified by country
57
Study T-301 Conclusion
GLIADEL Wafer administration produces a
clinically significant increase in survival
(risk reduction 29) in malignant glioma
patients undergoing primary surgery. Treatment
effect is maintained after accounting for the
effect of prognostic factors (risk reduction
28)
58
Study T-301 Reoperation for Disease Progression

• A higher percentage of patients had reoperation
  for disease progression than originally
  projected.
• Reoperation may have confounded the primary
  endpoint of survival.
• A prespecified sensitivity analysis was performed
  to account for the results of this event by
  censoring patients alive at the time of
  reoperation.

59
Study T-301 Overall Survival Analysis -
Reoperation for Disease Progression (ITT)
1 Stratified by country
60
Study T-301 Secondary Efficacy Endpoints

• Secondary Efficacy Endpoints
• Overall Survival - GBM patients
• Karnofsky Performance Decline, Neuroperformance
  Decline, Progression-Free Survival, and Quality
  of Life Evaluation

61
Study T-301 Overall Survival (GBM Patients)
1 Stratified by country
62
Study T-301 Overall Survival Adjusted for
Prognostic Factors1 (GBM Patients)
Hazard Ratio 95 CI 95 CI P-Value2
GLIADEL vs. Placebo 0.69 0.49 - 0.97 0.49 - 0.97 0.04
KPS lt70 vs. KPS gt70 2.04 1.38 - 3.01 1.38 - 3.01 lt0.001

1 Adjusted for age and Karnofsky Score 2 Cox
Proportional Hazard model stratified by country
and number of wafers (lt8,8) implanted
63
Study T-301 Karnofsky Performance Decline (ITT)
1.0
1 Stratified by country
64
Study T-301 Neuroperformance Decline (ITT)
1 Stratified by Country
65
Study T-301 Neuroperformance Decline in Speech
(ITT)
1 Stratified by country
66
Study T-301 Neuroperformance Decline in Cranial
Nerves III, IV, VI (ITT)
1.0
Hazard Ratio 0.68 95 CI 0.50
0.93 Risk Reduction 32 P 0.021
0.9
0.8
0.7
0.6
Proportion Without Decline
0.5
0.4
0.3
0.2
GLIADEL
Median Deterioration (weeks) Gliadel 54.9 Placeb
o 49.1
0.1
Placebo
0.0
0
10
20
30
40
50
60
70
80
90
100
110
120
Time to Deterioration (weeks)
1 Stratified by country
67
Study T-301 Neuroperformance Decline in Motor
Status (ITT)
1 Stratified by country
68
Study T-301 Neuroperformance Decline in
Cerebellar Status (ITT)
1 Stratified by country
69
GLIADEL Wafer Safety

• Review of safety in newly diagnosed malignant
  glioma

70
Safety Summary GLIADEL Wafer in Primary Surgery

• Intracranial hypertension 9.2 vs. 1.7. However,
  no difference in brain edema.
• Intracranial hypertension was typically observed
  late, at the time of tumor recurrence, and was
  not likely associated with GLIADEL use.
• CSF leak (5 vs. 0.8) was more common in
  GLIADEL - treated patients. However,
  intracranial infections and other healing
  abnormalities were not increased.
• Convulsions are not more common in GLIADEL
  -treated vs. placebo-treated patients.

71
Safety Summary GLIADEL Wafer in Primary Surgery

• Careful monitoring of GLIADEL -treated patients
  for cerebral edema/intracranial hypertension with
  consequent steroid use is warranted.
• CSF leak, though uncommon, may be more frequent
  in GLIADEL -treated patients. Attention to a
  water tight dural closure and local wound care is
  indicated.
• The safety profile of GLIADEL appears more
  benign in the primary surgery setting vs.
  recurrent disease.

72
Study T-301 Neurologic Adverse Events Occurring
in gt5 of Patients
Adverse Event
Abnormal gait
Amnesia
Anxiety
Aphasia
Ataxia
Brain edema
Coma
Confusion
Convulsion
Depression
Dizziness
Facial paralysis
Grand mal convulsion
Hallucinations
73
Study T-301 Neurologic Adverse Events Occurring
in gt5 of Patients
74
Convulsions (T-301)
75
Healing AbnormalityFluid, CSF or Subdural
Collections (T-301)
76
Healing AbnormalityCSF Leak (T-301)
77
Healing Abnormality Wound Dehiscence, Breakdown
or Poor Healing (T-301)
78
Healing Abnormality Subgaleal or Wound Effusion
(T-301)
79
Study T-301 Intracranial Infections
GLIADEL n () Placebo n ()
Abscess 4 (3) 5 (4)
Meningitis 2 (2) 2 (2)
Total 6 (5) 7 (6)
80
Post-operative Surgical ComplicationsComparison
of T-301 to Published Series

• The frequency of post-operative seizures,
  infections and hemorrhage/stroke are similar in
  the GLIADEL and placebo groups in the T-301
  study.
• Is the placebo wafer group a benign control as it
  involves the implantation of a placebo wafer?

81
Comparison of Post-operative Surgical Infections
Author/Study of Patients Disease Infection
Brell et al1 200 Glioma/ Metastasis 5.5
Sawaya et al2 327 Glioma 1.75
Kourinek et al3 2944 Craniotomy 4
Tenney et al4 251 Tumor 6
T-301- GLIADEL 120 Glioma 5
T-301- Placebo 120 Glioma 6
1 Brell et al., (2000) Acta Neurochir (Wien)
142739-50 2 Sawaya et al., (1998) Neurosurgery
421044-56 3 Kourinek et al., (1997) Neurosurgery
411073-81 4 Tenney et al, (1985) J Neurosurg
62243-7
82
Comparison of Post-operative Surgical Seizures
Author/Study of Patients Disease Seizures
Cabantog et al1 207 GBM/AA 1
Brell et al2 200 Glioma/ Metastasis 4
Sawaya et al3 327 Glioma 2.5
Pace et al4 119 Glioma 36-83
Tandon et al5 200 Glioma 51
Moots et al6 65 Glioma 32
T-301- GLIADEL 120 Glioma 33
T-301- Placebo 120 Glioma 37

1 Cabantog et al., (1994) Can J Neurol Sci
32213-18 2 Brell et al., (2000) Arta Neurochir
142739-50 3 Sawaya et al., (1998) Neurosurgery
421044-56
4 Pace et al., (1998) J Exp Clin Canc Rsh
17479-82 5 Tandon et al., (2001) Neurology India
4955-9 6 Moots et al., (1995) Arch Neurol
52717-24
83
Comparison of Post-operative Surgical
Hemorrhage/Stroke
Author/Study of Patients Disease Hemorrhage/Stroke
Cabantog et al1 207 GBM/AA 1
Brell et al2 200 Glioma/ Metastasis 4.5
Sawaya et al3 327 Glioma 2.0
T-301- GLIADEL4 120 Glioma 7.5
T-301- Placebo4 120 Glioma 4.8

1 Cabantog et al., (1994) Can J Neurol Sci
32213-18 2 Brell et al., (2000) Acta Neurochir
(Wien) 142739-50 3 Sawaya et al., (1998)
Neurosurgery 421044-56 4 Events reported within
30 days of surgery
84
Post-Operative Surgical Complications Conclusion
The frequency of seizures, infections, and
hemorrhage/ stroke after GLIADEL Wafer
implantation is similar to that observed after
craniotomy for glioma
85
GLIADEL Wafer in Primary Malignant Glioma
Summary of Benefits and Risks

• No evidence of earlier onset of seizures or
  increased frequency of seizures in primary
  malignant glioma patients
• CSF Leak was more common with GLIADEL treatment
• No evidence for increase in intracranial
  infections or other healing abnormalities

86
GLIADEL Wafer in Primary Malignant Glioma
Summary of Benefits and Risks

• Gliadel Wafer studies expanded to newly
  diagnosed malignant glioma
• Statistically significant and clinically
  meaningful increase in survival compared to
  placebo wafers
• Delayed time to overall function (KPS) and
  neurological decline (10/11 measures)S

87
Consistency of GLIADEL Wafer Phase III Trial
Results
Consistent efficacy and safety profile between
two prospective randomized, placebo-controlled,
double-blind Phase III clinical studies in the
primary surgery for malignant glioma
88
Summary of Efficacy Results from Randomized
Controlled Trials of GLIADEL Wafer (ITT)
89
Summary of Efficacy Results from Randomized
Controlled Trials of GLIADEL Wafer (GBM)
1 Hazard ratio adjusted for prognostic factors. 2
Stratified by country
90
Summary of Benefits and Risks
The benefit to risk ratio for GLIADEL in
patients with primary malignant glioma is
favorable
91
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.