Effect of primary tumor resection on survival in patients with metastatic colorectal cancer in the post-2000 era: A California Cancer Registry analysis.

Abstract

504 Background: Resection of primary tumors in the setting of metastatic colorectal cancer (MCRC) is controversial. Fewer primary tumor resections are being performed due to the improved tumor response and disease control rates associated with modern systemic therapy. Recent studies suggest a survival benefit for patients with MCRC who had primary tumors resection prior to systemic therapy. This analysis evaluates the independent prognostic impact of primary tumor resection on overall survival (OS) for patients with MCRC using the California Cancer Registry (CCR). Methods: We queried the CCR for all patients with MCRC diagnosed between 2003 and 2010. Patients were categorized by whether or not they had primary tumor resection at time of diagnosis. Covariates included gender, age, race/ethnicity, socioeconomic status (SES), and rural-urban commuting area (RUCA) score of patients. Univariate comparisons were made using the Kaplan Meier method. Multivariate comparisons were performed using the Cox proportional hazards regression method. Results: 19,836 patients met the criteria for analysis of whom 11,566 (58%) had primary tumor resection. Primary tumor resection rates declined over this time period (63% in 2003 v. 52.8% in 2010, p<0.0001), varied by SES (55% v. 62%: lowest versus highest, p<0.001) and residence (63% in rural versus 58% for urban, P=0.0160). On multivariate analysis, overall survival was significantly better for patients that had primary tumor resection (HR: 0.467 [95% CI: 0.467-0.482]; p<0.0001). Survival was statistically longer in younger patients (HR: 1.385 for age 65-75, HR: 2.217 if greater than age 75), highest SES (HR: 0.869, p<0.0001), Hispanics (HR: 0.884, p<0001), and Asian/Pacific Islanders (HR:0.892, p<0.0001). Overall survival was worse for African Americans (HR:1.105, P=0.0001). Conclusions: Our study demonstrates the independent prognostic value on survival of primary tumor resection in patients with MCRC. There is significant variability of resection rates by SES and rural-urban residence. As analysis of CCR data cannot eliminate the influence of patient and provider biases, a prospective randomized trial is warranted.