Molecular subtypes and outcomes in regorafenib-treated patients with metastatic colorectal cancer (mCRC) enrolled in the CORRECT trial

Abstract

3558 Background: In the CORRECT Ph3 trial (NCT01103323), regorafenib improved overall survival (OS) and progression-free survival (PFS) vs placebo in patients with mCRC who progressed on standard therapies (Grothey 2013). Initial biomarker analyses suggested that regorafenib provides a clinical benefit in various mutational subgroups (Jeffers 2013). Here we present additional exploratory analyses to evaluate clinical benefit in CRC subgroups defined by gene expression. Methods: Gene expression analysis (Affymetrix ST1.0 array) was conducted on archival tumor tissue from 281 of the 760 patients (37%) enrolled in CORRECT. Next-generation sequencing (NGS) was done on 239 specimens (FoundationONE). Gene expression data were subjected to hierarchical molecular tumor classification (Marisa 2013). Cox proportional hazards models were used to identify potential prognostic or predictive biomarkers. Results: The distribution of the 6 different CRC subtypes characterized by gene expression clusters originally defined by Marisa and the classification of the 281 patients from CORRECT are shown in the table. The six Marisa subtypes derived a similar OS benefit from regorafenib. Although the numbers of patients in the subgroups are small, a greater PFS benefit for regorafenib was seen in patients in the ‘high-risk’ subgroup (C4 +C6, n = 26; HR = 0.10; 95%CI 0.02 - 0.35; p = 0.0009) than the ‘low-risk’ subgroup (C1+C2+C3+C5, n = 255; HR = 0.58; 95%CI 0.44 - 0.77; p = 0.002). NGS analyses suggested that the chromosomal instability group (CIN) was the predominant subgroup in CORRECT. Conclusions: Molecular characterization by gene expression analysis may allow the identification of CRC subgroups that correlate with regorafenib clinical benefit. Data suggest this does not depend on the presence of mutations. Additional exploratory analyses, including the use of the recently defined Consensus molecular subtypes (Dienstmann ASCO 2014), to validate these results are ongoing. Clinical trial information: NCT01103323. Subtype CINImmuneDown (C1) dMMR (C2) KRASmutant (C3) CSC (C4) CINWntUp (C5) CINnormL (C6) Marisa (2013) 21% 19% 13% 10% 27% 10% CORRECT (n) 36% (100) 12% (33) 11% (32) 7% (19) 32% (90) 2% (7)