The medical records of 87 patients with 89 malignant colorectal polyps removed endoscopically between 1971 and 1983 were reviewed retrospectively. Fifty-five polyps contained carcinoma-in-situ. Four polyps had “pseudo-invasion” by displaced mucosal glands. Thirty polyps contained invasive carcinoma. No patients with carcinoma-in-situ or “pseudo-invasion” had either local residual disease or metastatic disease at the time of colectomy or which was detected during subsequent follow-up. Four patients (14%) with invasive cancer would have been inadequately treated by polypectomy alone, since one had residual disease at the polypectomy site, one had nodal metastases, one had liver metastases at the time of colectomy, and one subsequently developed liver metastases. Three histologic criteria correctly predicted all four cases where residual or recurrent disease was present: involvement of the polypectomy resection margin, lymphatic invasion within the polyp, and poorly differentiated histology. Polyp size, histology (villous adenoma, adenomatous polyp, or villo-adenomatous polyp), or anatomic location did not identify those patients who warranted further therapy. We conclude that polypectomy alone is adequate treatment for polyps containing carcinoma-in-situ. Polypectomy alone is also adequate treatment for most polyps containing invasive carcinoma. However, patients with lymphatic involvement within the polyp, poorly differentiated cancer, or resection margin involvement should probably undergo colectomy.