The aim of this retrospective study is to evaluate the results of local excision (LE) for rectal cancer for curative purposes. From 1969 to December 1997, a total of 456 operations were performed for surgical treatment of rectal carcinoma (262 males and 194 females, mean age 66 years). 20 patients (4.1%) underwent LE, 7 males and 13 females, median age 65 years. Patients were selected for LE if they met the following criteria during preoperative staging: tumors staged as T1-T2, N0, M0, grading G1 or G2, achievable location. As far as the type of LE is concerned, 13 transanal excisions (Francillon technique), 2 Mason, 2 endoscopic excisions and 3 TEM were performed. Among patients who underwent LE there was no operative mortality. 13 tumors were T1 and 7 were T2; pathologic findings included 20 adenocarcinoma, 14 G1 and 6 G2. There was no postoperative specific morbidity, while aspecific morbidity was minimal (5%). There were no local recurrences but 2 patients (10%) had secondary lesions. Five year overall survival following LE was 87.4%. Comparing T1 and T2 tumors after APR and SSR (17 T1 and 42 T2, all adenocarcinoma), operative mortality and specific morbidity were respectively 1.7% (p = 0.55) and 28% (p = 0.007). There were 5 (8.5%) local recurrences (p = 0.17) and 6 (10.2%) metastatic lesions. Five year overall survival was similar to LE (88.3%; p = 0.76). In conclusion the authors stress the importance that IE for rectal carcinoma must be performed only in selected patients provided there is correct preoperative staging. In these cases five year overall survival, local recurrence and operative mortality were similar to APR and SSR, while there was a statistically significative difference following LE in terms of specific morbidity.