Evidence for CEA utilization following curative resection of colorectal cancer

Objective: Evaluate the value of routine carcino-embryonic antigen (CEA) in follow-up of patients with resected colorectal cancer. Data Sources: Cancerlit 1983-1995, MEDLINE 1966-1995; colon neoplasm, rectal neoplasms, carcino-embryonic antigen, CEA, follow-up; any language, human data. Study selection: 18 articles: curative resection, CEA primary mode of follow-up, follow-up greater than or equal to 2 years. Data extraction: Guidelines for data quality and validity were agreed upon by four of the authors in conference. Extraction was completed by a single observer. Data synthesis: The data pooled from the 18 articles meeting the selection criteria indicated that: 33% of the total colorectal carcinoma patient population will develop recurrent disease, 18% will have the recurrence first detected by raised CEA levels, 8% will have second look laparotomy, potentially including 3.0% pelvic recurrence, 6.5% hepatic recurrence and 0.6% lung recurrence; 3.7% would have radical reresection with curative intent, including 1.5% for pelvic recurrence, 1.9% for hepatic recurrence and 0.3% for lung recurrence; a fraction of whom might have enhanced survival, while 1% would have second look laparotomy unnecessarily. Conclusions: Since routine use of CEA measurement in follow-up of patients with curatively resected colorectal carcinoma may benefit only a fraction of 3.7% of that population, acceptance of this test as standard practice is not supported by available evidence, and any further conclusion awaits the results of a large randomized study.