BACKGROUND: Increasing evidence suggests that a substantial proportion of colorectal carcinomas develop without a preexisting polypoid adenomatous lesion, but it is difficult to detect the possible origin of advanced carcinomas. The purpose of this study was to test the validity and significance of a new histopathologic classification system based on the histologic analysis of the tumor edge. METHODS: One hundred eighty-six unselected cases of colorectal carcinoma were included. A new classification method to distinguish polypoid and nonpolypoid growth type was based on the presence or absence of elevation of tumor as compared with adjacent mucosa. Inter- and intraobserver agreement of classification was tested. Association with other clinicopathologic features including histopathologic characteristics of the tumors, presence or absence of lesional and concurrent adenoma, K-ras mutations, and prognosis was evaluated. RESULTS: Classification could be made in 75% of the tumors, and 25% were unclassifiable, mostly due to absence of tumor margin in sections. Of the classifiable carcinomas, 45% were classified as polypoid, of which 52% had lesional adenoma. Nonpolypoid tumors formed 48% of classifiable cases, and only 2% had lesional adenoma. Features of both polypoid and nonpolypoid carcinomas were present in 7% of cases. Concurrent extralesional adenomas were found more frequently in association with polypoid carcinomas. K-ras mutations were more common in polypoid (43%) than in nonpolypoid tumors (8%; P = 0.018). Nonpolypoid carcinomas were significantly (P = 0.03) more aggressive than polypoid carcinoma, with 38% and 20% recurrence rates, respectively. CONCLUSIONS: The authors' results indicate that advanced colorectal carcinomas can be classified according to growth pattern by observing the tumor edge. This classification has prognostic significance because nonpolypoid carcinomas appeared to have a worse prognosis than polypoid ones. Copyright 2000 American Cancer Society.
BACKGROUND: Increasing evidence suggests that a substantial proportion of colorectal carcinomas develop without a preexisting polypoid adenomatous lesion, but it is difficult to detect the possible origin of advanced carcinomas. The purpose of this study was to test the validity and significance of a new histopathologic classification system based on the histologic analysis of the tumor edge. METHODS: One hundred eighty-six unselected cases of colorectal carcinoma were included. A new classification method to distinguish polypoid and nonpolypoid growth type was based on the presence or absence of elevation of tumor as compared with adjacent mucosa. Inter- and intraobserver agreement of classification was tested. Association with other clinicopathologic features including histopathologic characteristics of the tumors, presence or absence of lesional and concurrent adenoma, K-ras mutations, and prognosis was evaluated. RESULTS: Classification could be made in 75% of the tumors, and 25% were unclassifiable, mostly due to absence of tumor margin in sections. Of the classifiable carcinomas, 45% were classified as polypoid, of which 52% had lesional adenoma. Nonpolypoid tumors formed 48% of classifiable cases, and only 2% had lesional adenoma. Features of both polypoid and nonpolypoid carcinomas were present in 7% of cases. Concurrent extralesional adenomas were found more frequently in association with polypoid carcinomas. K-ras mutations were more common in polypoid (43%) than in nonpolypoid tumors (8%; P = 0.018). Nonpolypoid carcinomas were significantly (P = 0.03) more aggressive than polypoid carcinoma, with 38% and 20% recurrence rates, respectively. CONCLUSIONS: The authors' results indicate that advanced colorectal carcinomas can be classified according to growth pattern by observing the tumor edge. This classification has prognostic significance because nonpolypoid carcinomas appeared to have a worse prognosis than polypoid ones. Copyright 2000 American Cancer Society.