1155O - Updated overall survival (OS) analysis from a phase III study of sunitinib vs placebo in patients (pts) with advanced, unresectable pancreatic neuro...

Date 29 September 2012
Event ESMO Congress 2012
Session NETs and endocrine tumors
Topics Anticancer Agents
Neuroendocrine Tumours
Biological Therapy
Presenter Sandrine Faivre
Authors S. Faivre1, P. Niccoli2, J.L. Raoul3, Y. Bang4, I. Borbath5, J.W. Valle6, X. Huang7, S. Patyna7, R. Chao7, E. Raymond8
  • 1Cancérologie (oncology), Hôpital Beaujon, 92110 - Clichy/FR
  • 2Oncology, CHU Timone, Marseille/FR
  • 3Oncology, Institut Paoli-Calmettes, 13009 - Marseille/FR
  • 4Internal Medicine, Seoul National University College of Medicine, Seoul/KR
  • 5Gastro-enterology, Cliniques universitaires Saint-Luc, BE-1200 - Brussels/BE
  • 6Medical Oncology, The Christie NHS Foundation Trust, GB-M20 4BX - Manchester/UK
  • 7Oncology, Pfizer, La Jolla/US
  • 8Hôpital Beaujon, 92110 - Clichy/FR




A phase III trial demonstrated that sunitinib improved progression-free survival (PFS) and OS vs placebo in pts with pancreatic NET (Raymond et al, NEJM 2011). Upon disease progression or trial closure, 69% of pts randomized to placebo crossed over to sunitinib, potentially confounding OS analysis. We now provide both updated OS data and updated estimates of the effect of sunitinib on OS after adjusting for crossover.


Pts were randomized 1:1 to sunitinib 37.5 mg/d or to placebo, each with best supportive care. The primary endpoint was PFS; OS was a secondary endpoint to be evaluated every 2 y for 5 y or until 95% of pts had died. OS data were analyzed using four different methods to adjust for crossover: 1) censoring of placebo-arm data at crossover, 2) Cox model analysis with treatment as a time-dependent covariate, 3) rank-preserving structural failure time (RPSFT) analysis (which models the absence of crossover while respecting the randomization), and 4) RPSFT analysis adjusted for time of crossover.


171 pts were randomized (sunitinib: 86; placebo: 85). Pts randomized to placebo crossed over to sunitinib at disease progression (38/85, 45%) or at trial closure (21/85, 25%). Median duration of follow-up was 34.1 months (95% CI: 31.1 − 35.5). By April 2011 (with 2 y of follow-up post trial closure), a total of 87 deaths (51%) had occurred (sunitinib: 40; placebo: 47). OS analysis results are provided in the table.


Updated OS based on the ITT analysis continued to favor sunitinib, although this result was not statistically significant for reasons that may include treatment crossover and limited statistical power. Four different methods of adjusting for crossover demonstrated improvements in OS with sunitinib treatment. Together, these analyses demonstrate a survival advantage and further support the clinical benefit of sunitinib for pts with pancreatic NET.

OS analyses Hazard ratioa 95% CI P value
ITT − no adjustment for crossover 0.713 0.468 − 1.088 0.115
Adjustment for crossover:
Censoring at crossover 0.428 0.239–0.767 0.004
Time-dependent Cox model 0.492 0.285–0.851 0.010
RPSFT model 0.431 0.170–1.200b 0.115c
RPSFT model adjusted for crossover time 0.568 0.184–1.086b 0.115c

aSunitinib vs placebo

bFrom 20,000 bootstrap samples

cThe RPSFT method does not alter the P value obtained using the ITT method


S. Faivre: Honoraria: Pfizer and Novartis. Research funding: Pfizer and Novartis.

P. Niccoli: Consultant or advisory role: Novartis. Honoraria: Novartis. Research funding: Pfizer.

Y. Bang: Consultant or advisory role: Pfizer. Honoraria: Pfizer. Research funding: Pfizer.

J.W. Valle: Consultant or advisory role: Pfizer. Honoraria: Pfizer.

X. Huang: Employment: Pfizer. Stock ownership: Pfizer.

S. Patyna: Employment: Pfizer. Stock ownership: Pfizer.

R. Chao: Employment: Pfizer. Stock ownership: Pfizer.

E. Raymond: Consultant or advisory role: Pfizer. Honoraria: Pfizer. Research funding: Pfizer.

All other authors have declared no conflicts of interest.