Distribution of carcinoembryonic antigen and biologic behavior in colorectal carcinoma

PURPOSE: Carcinoembryonic antigen is assumed from the results of several experiments to be associated with invasion of colorectal carcinoma by adhesion or contact inhibition. The patterns and the intensity of carcinoembryonic antigen distribution in colorectal carcinoma were assessed to verify whether they were correlated with malignant potential from those biologic characteristics. METHODS: Carcinoembryonic antigen distribution was tested in the archival samples of 149 colorectal carcinomas by immunohistochemistry, using three characterized anti-carcinoembryonic antigen monoclonal antibodies: T84.66, PR1A3, and PR3B10. The distribution patterns in neoplastic tissue were categorized into unstained, apicoluminal, and diffuse-cytoplasmic patterns. Tumor, invasive tumor margin, and tissue surrounding the tumor were examined. RESULTS: Although all three antibodies revealed a positive correlation, T84.66 showed better discrimination than the others. Although none of the negative staining of the tumor or invasive tumor margin showed recurrence, the apicoluminal pattern showed recurrence, and the diffuse pattern showed the most frequent recurrence (P < 0.01). Recurrence was also associated with staining intensity in the apicoluminal pattern in both the tumor and invasive tumor margin (P < 0.05). Infiltrative tumor growth and lymph node metastasis were more frequent in cases of positive staining in tissue surrounding the tumor. Patients with the apicoluminal pattern achieved longer survival than patients with the diffuse-cytoplasmic pattern in the invasive tumor margin (P = 0.024) by a multivariate analysis including tumor stage and histologic differentiation. CONCLUSION: The distribution of carcinoembryonic:antigen in tumors and surrounding tissue seems to be closely correlated with invasiveness and metastatic behavior in colorectal carcinoma. Carcinoembryonic antigen immune staining can be considered as an efficient tool to determine groups with risk of recurrence.