1 ODAC September 13 th , 2005 Tarceva ® (erlotinib) Adrian Senderowicz, M.D. DDODP, CDER FDA
Jan 13, 2016
1
ODACSeptember 13th, 2005
Tarceva® (erlotinib)Adrian Senderowicz, M.D.
DDODP, CDERFDA
2
sNDA 21,743Tarceva® (erlotinib)
Applicant’s Proposed Indication:“TARCEVA® in combination with gemcitabine is indicated for
the first-line treatment of patients with locally advanced, unresectable or metastatic pancreatic cancer”
3
Overview of FDA Presentation
• Available therapies for pancreatic carcinoma
• Single pivotal study (PA. 3): Design
• Study results
• Conclusions
4
Pivotal study Gemcitabine vs 5-FUAdvanced Pancreatic carcinoma (N=123)
Gem: median OS: 5.7 mos (95% CI: 4.7-6.9)
5-FU: median OS: 4.25 mos (95% CI: 3.1-5.1)
p= 0.0009
Gem: clinical benefit response: 22%
p= 0.004
5-FU: clinical benefit response: 5%
5
Advanced Pancreatic carcinomaGemcitabine randomized trials
Gem vs Gem Ox (2)
Gem Gem/oxali N = 156 N = 157
7.1
9.0
HR = 0.83
p = 0.13**
26.9 38.2
p = 0.03
27.8 34.7
3.7
5.8
p = 0.04
16.7 28.7
p = 0.02
Gem vs Gem/CEF (3)
Gem Gem/ CEF N = 47 N = 52
NA
NA
HR = 0.65
p = 0.047
25 65
p = 0.01
21.3 38.5
3.3
5.4
p = 0.003
8.5 38.5
p <0.001 NS
Gem Vs FU (1)
5-FU Gem N = 63 N = 63
4.41
5.65
HR = NA
p = 0.0025
4.8 22.2
p = 0.004
2 18
0.92
2.33
p = 0.0002*
0 5.4
Median
Survival (mos)
Clinical benefit response (%)
% 12-month
survival
PFS (months)
% Response
Note: (1)Gemcitabine (1000 mg/m2) versus 5-FU (600 mg/m2). Burris, H et al J Clin Oncol, 15: 2403-2413, 1997. (2) Gem(1000 mg/m2) versus Gemcitabine (1000 mg/m2) + oxaliplatin (100 mg/m2). J Clin Oncol, 23: 3509-3516, 2005. (3) Gem (1000 mg/m2) versus Gemcitabine (600 mg/m2) + CDDP (40 mg/m2) + Epirubicin (40 mg/m2) + 5-FU 200mg/m2, CIV. Reni M et al Lancet Oncol, 6: 369-376, 2005. (4) Gemcitabine (1000 mg/m2) + 100 mg erlotinib versus Gemcitabine (1000 mg/m2) + placebo (sNDA). NA = not available, *unadjusted cox proportional ratios. ** cox proportional adjusted for PS and disease status. .
(4)
Plac/Gem Gem/Erlo N = 260 N = 261
5.95
6.38
HR = 0.81*
p = 0.02*
NA NA
17 24
3.55
3.81
p = 0.004**
8.0 8.6
p = 0.875
Gem vs Gem Erlo
6
PA. 3 TrialAdministrative History
• Protocol Final Version August 10th, 2001
• First Patient recruited November 29th, 2001
• Last Patient recruited January 31st, 2003
• Data Cutoff January 15th, 2004
– (484 deaths/85 censored)
• Statistical Analyses Plan (FDA) August 2004
• sNDA Submission April 29th, 2005
• Update sNDA data June 20th, 2005
– (551 deaths/18 censored)
7
PA.3 Study Design • Single, randomized, double-blinded, placebo-controlled, Phase
3, multinational• 2 Arms:
– Erlotinib/gemcitabine (E/G) vs – Placebo/gemcitabine (P/G)
• Planned sample size to obtain 381 deaths:
470 patients*• Stratified by:
– ECOG (PS ≤ 1 vs PS 2)
– Extent of disease (local vs metastatic)
– Center**
• Designed to show superiority
* sample was changed twice. **Center was removed as stratification factor at the time of the SAP, 9/2004
8
Study schema
RANDOMIZE
Gemcitabine 1000 mg/m2 weekly x 7 of 8 wks, then weekly q 3 wks of 4
Erlotinib 100/150 mg PO daily
N=285 (100 mg, N=261)
+ Placebo (Double-Blind)
N=284 (100 mg, N=260)
Metastatic or locally advanced Dz
ECOG
No prior chemoRx**
EGFR Status not required
Gemcitabine 1000 mg/m2 weekly x 7 of 8 wks, then weekly q 3 wks of 4.
Center*
*Center was removed as stratification factor as of 9/04. ** concomitant chemo/radiation as radiation sensitizer was allowed
Extent Dz
9
PA.3 TrialEndpoints
• Primary: Overall Survival– based on median PG survival of 6.6 months, expected increase by 33% to 8.8 months in
the EG, 80% power, hazard ratio of ~ 0.75. – Covariates:
• ECOG PS 0/1 vs PS2
• Extent of disease: locally advanced vs metastatic
• Pain intensity *: > 20 vs ≤ 20 vs not measured.
• Secondary: – PFS
– Response Rate and duration of Response
– Correlation of EGFR status and response rate
– Quality of life
– Toxicity
* SAP 8/2004: exclude pain score adjustment only if majority of patients have missing values
10
PA. 3 Study Results100 mg cohort
11
PA.3 TrialBaseline Characteristics
E/G, N= 261
%
P/G, N= 260
%
Extent of disease: LA vs Mets 23/77 24/76
ECOG PS: ≤ 1 vs 2 83/17 82/18
Gender: Female/Male 51/49 44/56
Age (median)* 64 yrs* 63 yrs*
Pain intensity**: ≤ 20 vs > 20 46/ 51** 46/52**
Race: white, black and other 86/ 3/ 11 89/ 2/ 9
EGFR status***:
positive/negative 16/13*** 11/12***
LA: locally-advanced. Mets: metastatic at baseline. *ages range: EG: range: 37-84 and PG: 36-92. **Missing values: 3% and 2 %, respectively. ***Missing values: 71% and 77%, respectively. .
12
Protocol Violations
• Major violations: pathological confirmation
– FDA assessment: other tumor types or
unable to confirm the diagnosis of
adenocarcinoma pancreas: N=18
13
Major Protocol violationsLack of pathological confirmation
Sponsor/FDA assessment
(N =18)
No pathological report 2
Lack of confirmation of malignancy 3
Other primary tumor* 10
Metastatic without proof of pancreatic origin 3
*Tumors discovered: ampula vater (N=5), acinar cell carcinoma (N=2), colon cancer (N=1), gastric cancer (N=1) and adenocarcinoma of non-pancreatic origin (N=1).
14
PA3. TrialPopulations for Analysis
Patients E/G Arm
N (%)
P/G Arm
N (%)
All Pts
N (%)
Analysis Population (patients randomized)
261
(100)
260
(100)
521
(100)
Analysis population minus major protocol violations* Sponsor/FDA discussion (N=18)
251 (96)
252 (97)
503 (97)
Safety Population (patients who received study drug)
259 (99) 256 (98) 515 (99)
*Protocol violations: wrong tumor diagnosis, lack of pathological confirmation of malignancy or lack of confirmation of adenocarcinoma of the pancreas
15
PA.3 TrialPatient Disposition
E/G(N=261)
P/G(N=260)
n (%) n (%)
Patients Never Treated 2 (<1) 4 (2)
Patients Off drug therapy 251 (96) 252 (97)
Reasons Off drug therapy
Progressive Disease 162 (62) 185 (71)
Adverse Events* 5757 (22)(22) 3737 (14)(14)
Intercurrent Illness 10 (4) 10 (4)
Patient Refusal 2222 (8)(8) 1515 (6)(6)
Death 25 (10) 21 (8)
Other 6 (2) 8 (3)
On drug therapy 8 (3) 4 (2)
Note: * adverse events that lead to drug discontinuation
16
PA3. Pancreatic Carcinoma TrialsOverall Survival analysis population
Survival time was defined as the time from the date of randomization to the date of death *Cox proportional adjusted by ECOG PS and Extent of disease. EG: Erlotonib/gemcitabine. PG: EG: Erlotonib/gemcitabine
0.03
0.06
0.02
Strat . Log Rank
p value
0.810.68 to 0.96
0.800.65 to 0.98
0.790.64 to 0.97
HR *adjusted95% CI
5.956.37443(data cutoff
5.946.47381(original
sample size)
5.946.37504(updated database)
P/GMedian OS
E/GMedian OS
Number of Deaths
0.810.68 to 0.96
0.800.65 to 0.98
0.790.64 to 0.97
HR *adjusted95% CI
5.956.37443(data )
5.946.47381(original
sample size)
5.946.37504(updated database)
P/GMedian OS
E/GMedian OS
95% CI(months)
Number of Deaths
95% CI(months)
17
Kaplan Meier Survival CurveErlotinib/Gem vs Placebo/Gem (504 deaths)
Time (months)
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 5 10 15 20 25 30 35
Pro
port
ion
surv
ivin
g
* Stratified log-rank test
Erlotinib/Gem: median OS: 6.37 mos (95% CI: 5.84 to 7.33)
Placebo/Gem: median OS: 5.94 mos (95% CI: 5.09 to 6.70)
HR: 0.811 (95% CI: 0.68 to 0.97)
p= 0.017
18
Sensitivity AnalysesOverall survival in analysis population minus
patients with major protocol violations
• Exclusion of patients without pathological confirmation
19
Erlotinib/Gem in Advanced Pancreas Sensitivity Analyses of Overall Survival using
Stratified* Log-Rank test
All unadjusted analyses showed NS results. *adjusted for ECOG PS and extent of disease. ** adjusted cox proportional hazard ratio
18 excluded 0.0350.82 0.69-0.98
0.0600.82 0.67-1.000418 excluded
HR** 95 % CI p*
381 deaths 0.0550.81 0.66-0.99
504 deaths 0.0280.81 0.68-0.97
0.0310.81 0.67-0.9818 excluded
443 deaths 0.0170.79 0.65-0.95
20
Exploratory analysesOverall survival
• Role of baseline characteristics
• Role of rash
21
Survival by Pretreatment Characteristics
0.71 0.4-1.2 66
0.76 0.5-1.3 70EGFR pos
EGFR neg
HR 95 % CI N
0.0 0.5 1.0 1.5 2.0
HR
0.76 0.6-1.0 274Age < 65 ys
0.75 0.6-1.0 273male
distant-rand.
0.5 0.3-0.8 88Ps 2-at rand.
E/G: P/G 0.83 0.7-1.0 521
Canada/US 0.74 0.6-0.9 280
Pain ≤ 20
0.77 0.6-1.0 369
0.70 0.5-0.9 238
0.93 0.7-1.3 152local- at rand.
0.89 0.7-1.1 433Ps 0-1 at rand.
Rest of World 0.96 0.7-1.3 241
0.91 0.7-1.2 247Age ≥ 65 ys
0.95 0.7-1.2 248female
Pain > 20 0.99 0.8-1.3 268
22
PA.3 trialPresence of Rash and Survival
Pro
port
ion
surv
ivin
g
Time (months)
Note: * log-Rank test. rash ≥ grade 2 (95% CI: 8.0 to 12.6 and : 4.2 to 8.8) and ≤ grade 1 (95 % CI: 4.8 to 6.0 and 4.9 to 6.6)
≥ grade 2 rash
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 5 10 15 20 25 30
E/G: OS: 10.7 mos
P/G: OS: 7.1 mosp= 0.02*
Pro
port
ion
Sur
vivi
ngTime (months)
≤ grade 1 rash
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
0 5 10 15 20 25 30 35
E/G: OS: 5.7 mos
P/G: OS: 5.9 mosp= 0.35*
23
PA.3 trialSecondary Endpoints
E/G P/G P value
Response rate (PR/CR, %) 1 8.6 7.9 ND
Duration of response (wks) 2 23.9 23.3 ND
Median PFS (months) 3 3.8 3.6 0.006
Survival by EGFR (months)
Positive4
Negative
6.6
6.0
5.3
6.0
NS
NS
Note: 1 number of patients with measurable disease: 244 vs 241, respectively. number of patients with responses: 21 vs 19, respectively. 2 Range (3.7 to 56.0+ vs 6.7+ to 65.3 +). 3 Adjusted cox proportional: 0.76. 4 total number of patients with EGFR positive and negative: 41 vs 34 and 32 vs 29, respectively. NS: not significant. ND: not done
24
PA.3 TrialSecondary endpoints (cont’d)Quality of Life assessment by EORTC QLQ-C30
• Mixed results do not confirm QoL benefit
• E/G worse diarrhea (p<0.001)
• E/G worse in some variables (e.g. cognitive
functioning, fatigue, dyspnea, appetite, global
QOL) while improved in other categories (e.g.
pain, sleep, social functioning, constipation).
• EG worse in global health status (32% vs 25%).
25
Category E/G (N=259) P/G (N=256)
n (%) n (%)
Patients with at least one AE 256 (99) 248 (97)
AEs (worst severity) regardless of causality
Grade ≤2 76 (29) 85 (36)
Grade 3/4 Grade 3/4 * 180 (69) 163 (64)
Patients with at least one SAE* Patients with at least one SAE* 131 (51) 99 (39)
Patients who discontinued study due to AEs*Patients who discontinued study due to AEs* 31 (12) 19 (7)
Patients who died on treatment or within 30 Patients who died on treatment or within 30 days of last treatment*days of last treatment*
81 (31) 68 (27)
PA.3 TrialSafety Profile
*Similar results for treatment-related AEs and regardless of causality
26
PA. 3 TrialDeath on therapy or within 30 days of therapy
Erlotinib/Gem
(N=81)
Placebo/Gem
(N=68)
N % N %
Toxicity from protocol Rx 2 2.5 0 0
Combination of pancreatic cancer and Rx
3 3.7 0 0
Other conditions 10 12.3 9 13.2
Other primary malignancy 1 1.2 0 0
Pancreatic cancer 65 80.2 59 86.8
27
PA. 3 TrialSevere AEs (grade ≥ 3)
Note: Severe: ≥ Grade 3 AEs. Stroke (ischemic and hemorrhagic). CHF: congestive heart failure
E/G (N=259) P/G (N=256)N (%) N (%)
Cardiovascular
Edema 12 (4.6) 5 (2)
Syncope 6 (2.3) 4 (1.6)
Arrhythmias 6 (2.3) 2 (0.8)
CHF 2 (0.8) 3 (1.2)
Hypertension 3 (1.2) 5 (2)
Hypotension 3 (1.2) 4 (1.6)
Ischemic events
Myocardial ischemia/infarction 7 (2.7) 3 (1.2)
Peripheral ischemia 1 (0.4) 0 (0)
Troponin elevation 1 (0.4) 0 (0)
Stroke 6 (2.3) 0 (0)
28
PA.3 Safety AnalysisIncidence of Stroke in erlotinib group
• There were 6 patients with stroke (incidence 2.3 %)– 5 ischemic– 1 hemorrhagic
• Median time to stroke was 24 days.
• The earliest case of stroke occurred by 2 days from drug initiation and the latest was 35 days after drug initiation.
29
PA.3 TrialSevere AEs (cont’d)
E/G (N=259) P/G (N=256)N (%) N (%)
Thrombotic
DVTs 9 (3.5) 3 (1.2)
PE 6 (2.3) 5 (2)
Other thrombosis 17 (6.6) 18 (7)
TTP 2 (0.8) 0 (0)
Pulmonary
Interstitial Lung disease 6 (2.3) 1 (0.4)
Pneumonitis 4 (1.5) 1 (0.4)
Hypoxia 1 (0.4) 1 (0.4)
Dyspnea 15 (5.8) 13 (5.1)
Pneumonia/lung infiltration 13 (5.0) 6 (2.4)
Note: Severe: ≥ Grade 3 AEs. DVTs: deep venous thrombosis. PE: pulmonary embolism. TTP: thrombotic thrombocytopenic Purpura. ARDS: acute respiratory disorder syndrome
30
PA.3 TrialSevere AEs (cont’d)
EG (N=259) P/G (N=256)N (%) N (%)
Gastrointestinal
Diarrhea 15 (5.8) 5 (2)
LFT elevation 19 (7.3) 15 (5.9)
GI bleeding 15 (5.8) 8 (3.1)
Ileus 4 (1.5) 1 (0.4)
Pancreatitis 5 (1.9) 2 (0.8)
Odynophagia 3 (1.2) 0 (0)
Hematology
Hemolytic anemia 2 (0.8) 0 (0)
Thrombocytopenia 4 (1.5) 1 (0.4)
Bleeding disorders 3(1.2) 1 (0.4)
Rash 12(4.6) 3(1.2)
Note: GI: gastrointestinal, LFT: liver function test. Severe: ≥ Grade 3 AEs.
31
PA.3 TrialSevere AEs (cont’d)
E/G (N=259) P/G (N=256)
N (%) N (%)
Central Nervous System
Neuropathy 5(1.9) 1(0.4)
Depression 5(1.9) 3(1.2)
Infections
Sepsis/bacteremias 22(8.5) 22(8.6)
Other infections 13(5.0) 2(0.8)
Bone
Bone pain 13(5.0) 9(3.5)
Aseptic bone necrosis 2(0.8) 1(0.4)
Others
Renal failure 3(1.2) 0(0)
Severe: ≥ Grade 3 AEs.
32
Causes E/G (N=259) P/G (N= 256)Total number of patients 22 15
Severe (≥ Grade 3) toxicity 12 3
Other causes 10 12
AEs E/G (N=12)** P/G (N= 3)**Liver enzymatic elevations 6 0
Deep venous thrombosis 3 0
Sepsis 3 0
Pneumonia 2 0
GI bleed 2 0
Others* 3 3
Causes for dropouts in patients that refused further therapy
Note: Others: EG: Congestive heart failure (N=1), Ileus/dehydration (N=1), Dysphagia (N=1), PG: Aseptic necrosis of bone (N=1), Arrhythmia (N=1), Arthralgias/muscle weakness (N=1). ** Total does not add up as patients could refuse further therapy due to more than 1 event
33
PA.3 trialSummary of Toxicity profile
• Erlotinib/gemcitabine arm has higher incidence of: – Grade 3/4 (regardless of causality and treatment related)
– SAEs (regardless of causality and treatment related)
– Discontinuation due to AEs (regardless of causality and treatment related)
– Toxic deaths
– Refusal of therapy
– Death on treatment or within 30 days of last treatment
• Most frequent AEs in the E/G group: Rash and diarrhea
• E/G has higher incidence of ILD-like disease
• Other severe AEs were higher in the EG arm: Stroke, TTP, myocardial
infarction, arrhythmias, edema, renal failure, bleeding disorders (GI and
non-GI related), ileus, pancreatitis, odynophagia and neuropathy.
34
FDA Guidance of Evidence of effectiveness from a single study without
independent substantiation• Large multicenter study: No single investigator
or site disproportionally responsible for favorable effects
• Consistency across study subjects (e.g. age, gender, disease state)
• Multiple endpoints (primary and secondary) involving different events are positive
• Statistically very persuasive finding (e.g. very low p-value): unethical to repeat trial
*Taken from guidance: Clinical Evidence of Effectiveness for human Drug and biological products, 5/98, pp 12-16
35
PA.3 trial Pancreatic Carcinoma Conclusions
• Single add-on trial: erlotinib adds marginal efficacy (clinical
and statistical) and adds severe toxicity.
– E/G increased overall survival (median difference ~ 12 days)
– E/G increased PFS (median difference ~10 days)
– No difference in response rate or duration of response
– No improvement in quality of life. However, diarrhea significantly worse in the E/G arm.
– No relationship between EGFR expression and survival was shown
– E/G had worse safety profile: greater number of AEs Grade 3/4, SAEs, discontinuation due
to AEs, refusal of therapy, toxic death and death on treatment or within 30 days of last
treatment.
– Stroke, TTP and other toxicities in the E/G are a safety concern.