The preoperative carcinoembryonic antigen test in the diagnosis, staging, and prognosis of colorectal cancer.

A study of preoperative carcinoembryonic antigen (CEA) levels was conducted in 319 patients with surgically treated colorectal cancer, 272 of whom had disease resectable with curative intent. Only three patients could not be completely followed. All of the remaining 316 patients have been followed for a minimum of 5 years or until death. From the standpoint of diagnosis, the CEA test was more frequently positive (greater than 5 ng/ml) in patients with advanced stage disease, with larger primary tumors, and with more differentiated histopathologic characteristics. It was grossly insensitive in diagnosis of resectable cancer (26%) and was only reasonably reliable (72%) in patients with unresectable and metastatic disease. In relationship to surgical pathology of colorectal cancer, CEA levels were significantly correlated with stage of disease and with size of the primary tumor in Dukes' B lesions, but not with extent of nodal metastasis in Dukes' C lesions. In advanced stage lesions, CEA was inversely correlated with degree of anaplasia. In the overall patient group, and also among resectable patients, the preoperative CEA level was strongly associated with survival after adjustment for the effects of a number of other prognostic factors. Within stages of resectable disease, however, CEA was not significantly associated with survival among patients with Dukes' A and B lesions or Dukes' C lesions with one to three nodes involved. CEA was found to be a significant and independent prognostic determinant only in patients with Dukes' C lesions who had four or more metastatically involved lymph nodes. Under these circumstances, a preoperative CEA level could perhaps be of some value for stratification of Dukes' C patients in randomized colorectal cancer surgical adjuvant trials. The value of this test as a prognostic guide in clinical practice, however, would seem to be limited because of a lack of sensitivity in identifying individual poor prognosis patients.