Rui Zhu (1), Dan Lu (1), Yu-Waye Chu (1), Akiko Chai (1), Michelle Green (2), Nancy Zhang (2), Jin Yan Jin (1)
(1) Genentech, Inc., California, USA; (2) Quantitative Solutions, Inc., California, USA.
Objectives: Demonstration of clinically meaningful improvements in progression-free survival (PFS) and/or overall survival (OS) in clinical trials of novel therapeutic agents for the treatment of non-Hodgkin lymphoma (NHL) are necessary for their regulatory approval. However, NHL trials with OS or PFS as primary endpoint require large sample sizes and long follow-up durations. In this study, correlations between early and late efficacy endpoints were assessed with published clinical trial-level data to identify potential surrogate endpoints for OS or PFS in each of the three major NHL subtypes: diffuse large B cell lymphoma (DLBCL), follicular lymphoma (FL), and mantle cell lymphoma (MCL).
Methods: Phase II and phase III studies in patients with DLBCL, FL, or MCL, published from 1993 to 2013, encompassing 108 unique trials with 129 trial arms, were identified and used in the analysis. Approximately 71%, 52%, and 48% of trials in DLBCL, FL, and MCL included rituximab-containing therapies, respectively. Correlations between efficacy endpoints were analyzed using weighted linear regression and Pearson’s correlation.
Results: In trials of newly-diagnosed DLBCL, six-month PFS was strongly correlated with 2-year OS (R2=0.81, 95% confidence interval [CI] 0.51-0.96). Six-month PFS was strongly correlated with 3-year PFS (R2=0.89, 95% CI 0.62-0.96) in FL and was moderately correlated with 2-year OS (R2=0.69, 95% CI 0.40-0.91) in MCL trials. Linear regression determined that a 10% increase in 6-month PFS would predict for a 22.6% +/- 1.1% increase in 3-year PFS in FL. In addition, both 6-month PFS and complete response (CR) rate showed moderate correlations with median PFS in FL trials with R2=0.66 (95% CI 0.52-0.98) and R2=0.69 (95% CI 0.22-0.89), respectively. However, correlations between CR rate and median OS were not evaluable due to limited data in all three NHL subtypes.
Conclusions: Six-month PFS may be a potential surrogate endpoint for 2-year OS in newly-diagnosed DLBCL and MCL and for 3-year PFS in FL, and both 6-month PFS and CR rate may be potential surrogate endpoints for median PFS in FL patients. Confirmation and validation of these correlations may facilitate early interpretation of NHL trials.
Reference: PAGE 24 () Abstr 3436 [www.page-meeting.org/?abstract=3436]
Poster: Methodology - Other topics