BACKGROUND: The appropriateness of resection in patients from whom polyps with invasive adenocarcinoma were excised has been questioned.METHODS: To determine the results of this policy, the authors reviewed the outcome of 42 patients from whom 44 such polyps were removed. Each polyp was categorized for the level of invasion according to the classification of Haggitt.RESULTS: Level 1 invasion was found in 27%; level 2, in 9%; level 3, in 11%; level 4, in 39%; and uncertain, in 14%. The histologic grade was well differentiated in 48% of patients and moderately differentiated in 52%. No polyps contained poorly differentiated adenocarcinoma; lymphatic and vascular invasion were not encountered. Excision was judged complete in 23 patients; 11 underwent resection, and in none was residual adenocarcinoma identified. In 14 patients, margins could not be evaluated; of 12 patients who underwent resection, residual adenocarcinoma was found in 1. Of the seven patients with positive margins who underwent resection, residual adenocarcinoma was found in only two. In the resected specimens in which residual carcinoma was encountered, all original lesions were designated level 4. None of the patients treated by polypectomy alone has experienced a recurrence at a mean follow-up time of 66 months (range, 12-152 months).CONCLUSIONS: The authors conclude that only patients with level 4 invasion require resection.