Long-term cure in surgery for extrarectal pelvic recurrence of rectal cancer.

Abstract

Survival benefit of radical surgery for locally recurrent rectal cancer depends on whether disease is cured rather than whether death is delayed. Cured patients gain decades of life and are spared from sufferings with recurrence. Unfortunately, the majority of patients undergoing surgery, particularly those with extrarectal pelvic recurrence, have poor outcomes with occult disseminated disease. This study was designed to identify which of these patients are curable. Of 61 patients with pelvic recurrence treated by radical reexcision more than nine years before, 36 patients whose initial surgery was abdominoperineal resection were examined retrospectively. We used the logistic regression and Gamel-Boag regression models to estimate curability and identify predictors of cure. Ten patients survived five years and seven survived ten years. The cumulative disease-specific mortality curve leveled off 6.5 years after reexcision and remained at 74 percent (95 percent confidence interval, 60-89), indicating that the remaining 26 percent are curable. This value is comparable with the 23 percent curability estimated by the Gamel-Boag model, which also found that the disease-free interval from the initial surgery to the first recurrence is the best predictor of cure (P = 0.005). Of 11 patients with disease-free interval three years or more, 6 survived ten years, whereas 8 of 9 patients with disease-free interval less than one year died of second recurrence within three years of reexcision. Even patients with extrarectal pelvic recurrence may have isolated disease that is amenable to complete eradication. As a biologic marker, the disease-free interval serves to predict curability and may distinguish isolated disease from occult disseminated disease.