The impact of gap duration on local control in anal canal carcinoma treated by split-course radiotherapy and concomitant chemotherapy

Purpose: To investigate the potential benefit of reducing the intersequence gap in patients with anal cancer treated with split-course chemoradiothera py. Methods: The study group consisted of 90 patients with anal squamous carcin oma treated between 1981 and 1998, using concomitant chemotherapy (CT) and radiation (RT), Median age was 65 years (range 41-87). RT was delivered in a split course, with a median gap of 37.5 days (range 4-97) between sequenc es. First (pelvic) sequence delivered a median dose of 40 Gy (range 36-50.4 ), using AP/PA megavoltage photon beams. Boost treatment (median dose 20 Gy , range 13-26) consisted of either Iridium-192 implantation (49 patients) o r external beam RT (41 patients). CT consisted of 1-2 cycles of a 5-day con tinuous infusion of 5-fluorouracil and bolus mitomycin C, usually administe red during the first week of each RT course. Median follow-up was 76.2 mont hs. Univariate and multivariate analyses were performed to determine the fa ctors associated with locoregional control (LRC), Results: Five-year actuarial LRC was 72.5%. Factors associated with poorer LRC (univariate) were: age I 65, male gender, and gap > 37.5 days. Number o f CT cycles (1 vs. 2 or more), boost technique (brachytherapy vs. external) , and T-stage were not significantly associated with LRC, m multivariate an alysis, only age (p = 0.01), and gap (p = 0.02) retained their significance . In patients older than 65 years, LRC was 92.3% and 75% for shorter and lo nger gaps, respectively. In younger patients, the corresponding values for LRC were 73.7% and 50%. Conclusion: In anal cancers, split-course RT with > 50 Gy dose delivery is difficult to avoid because of acute toxicity. The present analysis suggests that shortening the gap contributes to optimizing LRC, Gaps longer than 5 weeks correlated with poorer I,RC, with especially unsatisfactory results o bserved in younger patients. (C) 2001 Elsevier Science Inc.