Influence of local recurrence on survival in patients with rectal cancer

Abstract

Background: Recent trials on rectal cancer have demonstrated significant improvements in local recurrence without improvements in overall survival. The aim of this paper was to define the influence of local recurrence on survival in a prospective series of patients who underwent R0 or R1 resections for rectal cancer. Methods: Patients presenting with rectal cancer from 1996 to 2012 were prospectively audited. The study included patients who underwent an R0 or R1 resection. Local recurrence was defined as cancer regrowth detected in the pelvis regardless of whether or not new metastases were found elsewhere. Kaplan–Meier curves, smoothed hazard functions and Cox models using both time since diagnosis and age as the time scale were used to define the influence of local recurrence on overall survival. Results: The study involved 483 patients, of mean age 66 years (standard deviation = 13) and a median follow-up of 5.2 years. The results at 5 years were overall survival 71% (95% confidence interval (CI) 66–75), local recurrence 7% (95% CI 5–10) and distant recurrence 18% (95% CI 14–22). Patients diagnosed with local recurrence died faster than patients diagnosed with either distant recurrence or no recurrence, and this was particularly obvious for younger patients (local hazard ratio (HR) 54, 95% CI 12–253 and distant HR19, 95% CI 4–80). Local recurrence that developed early following surgery also had worse survival outcomes. Conclusions: Within this cohort of rectal cancer patients, the early development of local recurrence was the single most important indicator of a reduced survival, and carried a worse prognosis than the development of distant metastases alone.