A retrospective cohort study of metastatic colorectal cancer patients treated with oxaliplatin-based chemotherapy, with an exploratory analysis of changing serum carcinoembryonic antigen levels.

Aim: To examine the relationship between changes in serum carcinoembryonic antigen (CEA) levels and survival during oxaliplatin-based chemotherapy for metastatic colorectal cancer (mCRC). Methods: A retrospective review of 142 patients with mCRC who were treated with oxaliplatin-based chemotherapies (mostly FOLFOX 6 or XELOX) by St Vincent's Hospital, from October 1999 until 30 November 2007. Survival analysis was used to determine median overall survival (OS) from commencement of chemotherapy. A CEA response was defined by =50% decline compared with baseline, maintained on two consecutive occasions at least 4 weeks apart. The Cox proportional hazard model and a landmark analysis at 3 months were used to evaluate survival differences between CEA responders (rCEA) and non-responders (non-rCEA). Results: The median OS was 14.7 months. Using an intention-to-treat analysis, 76 (53.5%) patients achieved a CEA response, while 66 (46.5%) did not. Using the landmark analysis at 3 months, rCEA had a longer survival than non-rCEA (median 16.0 vs 7.8 months, P < 0.0001). The hazard ratio for patients dying of mCRC in non-rCEA was 2.2 (P < 0.0001). In multivariate analysis, CEA response and better baseline Eastern Cooperative Oncology Group (ECOG) predicted for survival (P < 0.0001 for both), while age, gender and histology grade did not. Conclusion: The median OS of our patients is similar to published randomized trials. A CEA response of =50% at 3 months and good ECOG were independent predictors of OS of patients with mCRC treated with oxaliplatin-based chemotherapies.