Evaluating Patient Selection for Surgery in Older Patients with Non-Metastatic Colorectal Cancer in Ontario: A Population-Based Cohort Study

Abstract

BACKGROUND: Surgeons are faced with an increasing number of patients over the age of 80 presenting with non-metastatic colorectal cancer (CRC). Justifying major surgery in a comorbid and frail population requires individualized decision making and remains a challenge for surgeons. Few studies compared survival in these patients with and without surgery, but all had important biases. In addition, no study has evaluated survival from a surgeon's case selection perspective. This study evaluated the quality of patient selection in Ontario octogenarians and nonagenarians with non-metastatic colorectal cancer by comparing survival by surgery status in important patient subgroups. METHODS: A retrospective population-based cohort study was conducted using population-based health administrative data in Ontario, Canada from 2010 to 2020. We included all patients aged 80 and over with stage I-III colorectal cancer. We first conducted an unadjusted analysis using methods described by Simon and Makuch, then we used a proportional hazards model to measure the association of surgical cancer resection with all-cause survival adjusted for: patient age, sex, and frailty; cancer stage and location; and year. Surgical status was a binary variable expressed as a time-dependent covariate. Interactions between surgical status and all other covariates were included. These results were visualized using heat maps. RESULTS: We identified 5782 patients; 4779 underwent elective colorectal surgery and 1003 did not. The surgery group was younger (84.4± 4.5 vs 86.9± 3.6), had more colon/recto-sigmoid cancer (83.7% vs 62.6%) and the difference in frailty between the two groups was small (standardized difference 0.21). Patients selected for surgery had significantly improved survival in almost all subgroups including stage I to II colon and rectal cancer and almost all patients with stage I-III colon cancer. Survival benefit of surgery was greatest in stage 1 disease and progressively decreased with stage 2 and 3 disease. Both male and female patients with stage III rectal cancer having increased levels of frailty undergoing surgery did not have better survival. The most important covariates associated with survival (from most to least important) were surgery, frailty, age, stage, and sex. Tumor type had the smallest independent influence on survival. CONCLUSION: Surgeons in Ontario are very good at identifying older patients with colorectal cancer in whom surgery improves survival. Age and frailty alone should not preclude patients from being considered for surgery.