An Evaluation of the Carcinoembryonic Antigen (CEA) Test for Monitoring Patients With Resected Colon Cancer

Abstract

OBJECTIVE: To determine the effectiveness of carcinoembryonic antigen (CEA) monitoring in detecting surgically curable recurrence of colon cancer. DESIGN: Clinical data were collected from a national surgical adjuvant trial in which CEA monitoring was elective. SETTING: Cancer centers, universities, and community clinics. PATIENTS: A total of 1216 patients with resected colon cancer, 1017 (84%) of whom had CEA monitoring. MAIN OUTCOME MEASURES: Sensitivity and specificity of CEA testing for cancer recurrence and CEA-motivated diagnostic and surgical interventions and their end results. RESULTS: Among 417 monitored patients with recurrence, 59% had a preceding elevation of CEA concentration. Sixteen percent of 600 patients without recurrence showed a false-positive test result. Carcinoembryonic antigen testing was most sensitive for hepatic or retroperitoneal metastasis and relatively insensitive for local, pulmonary, or peritoneal involvement. Surgical explorations were performed in 115 patients with CEA elevations, and 47 recurrences, usually hepatic, were resected with curative intent. On the other hand, 38 patients with normal CEA concentrations and 23 patients not monitored also underwent such resections--usually for pulmonary or local recurrence. Of all CEA-monitored patients, 2.3% are alive and disease free more than 1 year after salvage surgery (2.9% of those with CEA elevations and 1.9% of those with no elevations). Of patients with no CEA monitoring, 2.0% are also alive and disease free more than 1 year after salvage surgery. CONCLUSIONS: Cancer cures attributable to CEA monitoring are, at best, infrequent. It is questionable whether this small gain justifies the substantial cost in dollars and physical and emotional stress that this intervention may cause for patients.