To evaluate our data concerning the prognostic factors for locoregional control, survival, late complications, and sphincter conservation in a series of epidermoid cancers of the anal canal without clinical evidence of metastasis.

Between June 1972 and January 1997, 305 patients were treated with curative-intent radiotherapy (RT). The T stage according to the 1987 International Union Against Cancer classification was T1 in 26, T2 in 141, T3 in 104, and T4 in 34. Forty-nine patients had nodal involvement at presentation. The pretreatment anal function score, according to our in-house system, was 0 for 22 patients, 1 for 182, 2 for 74, 3 for 7, and 4 for 11 patients; for 9 patients, scores were unavailable. The treatment started with external beam radiotherapy (EBRT) in 303 patients (median dose 45 Gy). After a rest period of 4-6 weeks, a boost of 20 Gy was delivered by EBRT in 279 patients and by interstitial (192)Ir brachytherapy in 17 patients. Seven patients received only one course of EBRT (mean dose 49.5 Gy), and 2 patients were treated with interstitial (192)Ir brachytherapy alone (55 Gy and 60 Gy). Concomitant chemotherapy (5-fluorouracil and either mitomycin C or cisplatin) was delivered to 19 patients. The mean follow-up was 103 months (median 84).

At the end of RT, the local tumor clinical complete response rate was 96% for T1, 87% for T2, 79% for T3, and 44% for T4. Of the 61 locally progressive tumors, 27 (44%) were salvaged with abdominoperineal resection. The rate of local tumor relapse was 12%. Among 37 local tumor relapses, 20 (54%) were salvaged with abdominoperineal resection and one with interstitial (192)Ir brachytherapy. The overall local control rate (with or without salvage local therapy) was 84%. The local control rate with good anal function (score 0 or 1) was 56.5%. Of 181 available patients with their anus preserved, 94% had good anal function. For a subgroup of 15 patients with a tumor length of <2 cm and without nodal involvement, the clinical complete response rate after RT completion was 100%, the local control rate with or without local salvage treatment was 100%, and among 13 available patients with their anus preserved, the anal function score was good in 12 patients (92%). The 10-year disease-free survival rate was 74%. After multivariate analysis, three independent predictive factors significantly influenced disease-free survival: the interval between the two courses of RT (>38 days vs. < or =38 days, p = 0.0025), pretreatment anal function score (0 vs. 1 vs. 2 vs. 3 vs. 4, p = 4.4.10(-6)), and clinical complete response after RT completion (no complete response vs. complete response, p = 2.5.10(-14)).

We confirm the excellent results with RT in T1 and T2 lesions. However, to improve survival without colostomy with good anal sphincter function, chemoradiotherapy should be preferred for tumors > or =2 cm in length and for locally advanced tumors.