Should carcinoembryonic antigen be used in the management of patients with colorectal cancer?

The contribution of carcinoembryonic antigen carcinoembryionic antigen for the effective management of colorectal cancer patients remains a controversial issue. The aim of this study is to attempt to get some valid answers to its function in the diagnosis, prognosis, and overall management of colorectal cancer patients. METHODS: A retrospective review of colorectal cancer patients managed and prospectively registered by the authors between 1985 and 1998 was performed. Serum carcinoembryionic antigen levels were determined preoperatively in 209 patients with primary colorectal cancer and postoperatively in 196 patients who had undergone curative resection of their tumors, according to a fixed schedule. A maximum value of 5 ng/ml was accepted as being normal. With the exception of endoscopy, all other diagnostic techniques were only used after an abnormal carcinoembryionic antigen result (a raised value found twice consecutively). RESULTS: carcinoembryionic antigen preoperative values were raised only in 40 percent of patients and were related to disease stage, with the highest values found in patients with Stage IV disease. However, an elevated preoperative carcinoembryionic antigen value had a very marked prognostic importance, with a statistically significant difference in survival curves (Kaplan-Meier); the same was valid for curatively resected patients (Stages I, II, and III) and for Stages II and III patients considered separately. Multivariate analysis using the Cox proportional hazards technique confirmed these results, showing preoperative carcinoembryionic antigen to have an independent prognostic value, with a relative risk of recurrence of 3.74 for patients with raised preoperative carcinoembryonic antigen levels. In postoperative follow-up, carcinoembryionic antigen elevation was found to be a very accurate marker of recurrence (sensitivity, 77 percent; specificity, 98 percent), mainly in liver metastasis (sensitivity, 100 percent), and the best marker of asymptomatic recurrence (63 percent of cases). However, carcinoembryionic antigen's impact on overall survival was negligible because of the poor results of surgical treatment of recurrences. CONCLUSIONS: Preoperative carcinoembryionic antigen is a very important prognostic indicator and should be considered in future trials. Postoperative carcinoembryionic antigen elevation is a very sensitive marker of recurrence and even of asymptomatic recurrence, but its impact on overall survival does not seem to be relevant. Nevertheless, carcinoembryionic antigen should continue to be used in colorectal cancer patients until better methods of diagnosis and treatment of recurrence are developed.