Understanding the Prognostic Value of Primary Tumor Location and KRAS in Metastatic Colorectal Cancer: A Post Hoc Analysis of the OPTIMOX3 DREAM Phase III Study

International audienceIntroduction: We evaluated the prognostic value of KRAS and primary tumor location (PTL) for overall survival (OS), progression-free survival (PFS), and post-progression survival (PPS) in metastatic colorectal cancer (mCRC).Material and methods: Individual patient data from the DREAM phase III study were retrospectively analyzed. PTL was defined as right-sided or left-sided if tumor arising from the cecum to transverse colon or from the splenic flexure to the rectum, respectively. OS, PFS, and PPS were estimated using the Kaplan-Meier method and compared using log-rank test.Results: Among 700 patients included in the DREAM study, both PTL and KRAS were available for 536 (76.6%) patients. PTL showed stronger prognostic impact than KRAS status for OS (HRPTL, 1.62 vs. HRKRAS, 1.37), PFS (HRPTL, 1.27 vs. HRKRAS, 1.15) and PPS (HRPTL, 1.54 vs. HRKRAS, 1.33). Interaction between PTL and KRAS was significant (Pinteraction = .003). A negative impact of KRAS mutation was observed for OS and PPS, but not for PFS. Right-sided tumor was associated with poorer Eastern Cooperative Oncology Group performance status, anemia, and KRAS mutation, whereas left-sided KRAS wild-type tumor was associated with an increased lactate dehydrogenase. In patients with KRAS mutant mCRC, alkaline phosphatase was the main prognostic factor whatever the tumor site, whereas in those with KRAS wild-type tumors, prognostic factors varied according to PTL. The exposition to the anti-epidermal growth factor receptor (anti-EGFR) agents during and after study was similar in patients with left-sided and right-sided KRAS wild-type tumors.Conclusion: Our findings suggest that a better prognosis of patients with mCRC with left-sided tumors is driven more strongly by PPS than by PFS when compared with patients with right-sided tumors, whatever the KRAS mutation status. This phenomenon was independent from the exposition to poststudy anti-EGFR monoclonal antibody