Oops, you're using an old version of your browser so some of the features on this page may not be displaying properly.

MINIMAL Requirements: Google Chrome 24+Mozilla Firefox 20+Internet Explorer 11Opera 15–18Apple Safari 7SeaMonkey 2.15-2.23

Poster display session

4921 - Interpreting progression-free survival and overall survival data in biosimilar clinical studies: considerations based on a recent rituximab biosimilar study


09 Sep 2017


Poster display session


Cancers in Adolescents and Young Adults (AYA);  Haematological Malignancies


Jutta Amersdorffer


Annals of Oncology (2017) 28 (suppl_5): v355-v371. 10.1093/annonc/mdx373


J. Amersdorffer1, Y. Li2, S. Alexandrova2, A. Zubel3, E. Sasse3, H. Mellstedt4

Author affiliations

  • 1 Medical Affairs, Hexal AG, 83607 - Holzkirchen/DE
  • 2 Biostatistics, Sandoz Inc., Princeton/US
  • 3 Medical Affairs, Hexal AG, Holzkirchen/DE
  • 4 Department Of Oncology-pathology, Karolinska University Hospital Solna, Stockholm/SE


Abstract 4921


Oncology trials often report data on progression-free survival (PFS) or overall survival (OS), but such endpoints are prone to be less sensitive than short-term overall response (ORR) for confirming biosimilarity when long median PFS or OS are expected. We outline considerations when interpreting survival data with a sensitivity analysis from a recent rituximab biosimilar study.


A confirmatory phase III study compared the efficacy of the EMA approved biosimilar rituximab, GP2013 (n = 314) with reference rituximab (R) (n = 315) in patients with previously untreated, advanced follicular lymphoma. Patients received CVP chemotherapy during induction and responders received GP2013 or R maintenance monotherapy. Primary endpoint was equivalence of ORR at the end of induction. Secondary endpoints included PFS and OS, and hazard ratios (HRs) were estimated by a Cox proportional hazard model, with treatment allocation as the main effect and FLIPI score as a stratification factor.


As of 31 Dec 2016, the median follow-up was 23.6 and 24.2 months for GP2013 and R, respectively. Equivalence criteria for the primary endpoint were met, confirming biosimilarity. For time-dependent endpoints, there was a high level of censoring without PFS (∼70%) or OS (∼90%) events. Median PFS or OS could not be estimated. HRs for PFS and OS were 1.31 (90% CI [1.02–1.69]) and 0.77 (90% CI [0.49, 1.22]), respectively. Kaplan-Meier analysis showed that PFS survival curves diverged between 12–24 months yet ran parallel outside this period, violating the proportional hazards assumption of the Cox proportional hazard model. Complete response (CR) rates were similar between treatments at all time points, including 33 months.


Small sample size, low event rate, data immaturity and/or other aspects of study design can subject survival analyses to chance findings and decrease sensitivity for biosimilarity assessments. In this study, HRs for PFS and OS had opposite directions and CR rates between treatments were similar across time, emphasizing these challenges. The PFS and OS results should be interpreted with caution as they may not reflect a difference, or lack thereof, between treatments.

Clinical trial identification


Legal entity responsible for the study

Hexal AG, a Sandoz company, part of the Novartis group


Hexal AG, a Sandoz company, part of the Novartis group


J. Amersdorffer: Employee of Hexal AG, Holzkirchen, Germany. Y. Li, S. Alexandrova: Employee of Sandoz Inc. Princeton, NJ, USA. A. Zubel, E. Sasse: Employee of Hexal AG, Holzkirchen, Germany. H. Mellstedt: Received speaker or consultancy honoraria from: F. Hoffmann-La Roche Ltd; Amgen Inc.; Pfizer Inc.; Hospira; Boehringer Ingelheim Pharmaceuticals, Inc.; AbbVie Inc.; Sandoz.

This site uses cookies. Some of these cookies are essential, while others help us improve your experience by providing insights into how the site is being used.

For more detailed information on the cookies we use, please check our Privacy Policy.

Customise settings
  • Necessary cookies enable core functionality. The website cannot function properly without these cookies, and you can only disable them by changing your browser preferences.