Various histologic findings exist for managing patients with malignant polyps. Our goal was to determine the criteria for a conservative approach to patients with locally excised early invasive carcinoma. In 292 early invasive tumors (local resection followed by laparotomy [80 tumors, group A], local resection only [41 tumors, group B], and primarily laparotomy [171 tumors, group C], potential parameters for nodal involvement were analyzed. The status of the endoscopic resection margin also was examined for the risk for intramural residual tumor. Unfavorable tumor grade, definite vascular invasion, and tumor budding were the combination of qualitative factors that most effectively discriminated the risk for nodal involvement in patients in groups A-C. The nodal involvement rate was 0.7%, 20.7%, and 36.4% in the no-risk, single-risk, and multiple-risks group, respectively. Thirty-two and 9 patients from group B were assigned to the no-risk and one-risk group, respectively; extramural recurrence occurred in 2 patients with risk factors. Considering quantitative risk parameters for submucosal invasion (i.e., width > or =4000 microm or depth > or =2000 microm), nodal involvement (including micrometastases) was not observed in the redefined no-risk group that accounted for about 25% of the patients from groups A and C. An insufficiency of endoscopic resection could be evaluated most precisely based on the coagulation-involving tumor, rather than the 1-mm rule for the resection margin. Provided that the criterion of sufficient excision is satisfied, the absence of an unfavorable tumor grade, vascular invasion, tumor budding, and extensive submucosal invasion would be the strict criteria for a wait-and-see policy.