Individual cutoff levels of carcinoembryonic antigen and CA 242 indicate recurrence of colorectal cancer with high sensitivity

Abstract

Purpose: This study was designed to identify the sensitivity and specificity associated with recurrent colorectal cancer, principally hepatic metastases, with individual cutoff levels of carcinoembryonic antigen; identify the corresponding data for CA 50 and CA 242; compare these findings with standard cutoff levels, clinical examinations, and patients' personal health assessments; and identify the time between increasing serum levels and detection of actual relapses, particularly hepatic metastases.

Methods: A prospective study was performed that included 132 patients with colorectal cancer who underwent surgery with curative intent and who were followed up for a minimum of 5 years. Serum samples were obtained the first two years after surgery. Marker serum levels were analyzed with the commercial Delfia test kit.

Results: During the sampling period, carcinoembryonic antigen levels were very high or above the individual cutoff level in 21 of 24 cases with recurrent disease. All eight hepatic metastases detected during the sampling period were indicated by carcinoembryonic antigen 0 to 19 months earlier. In the 3rd postoperative year, nine patients with recurrence were identified, and in eight of them, carcinoembryonic antigen rose above the individual cutoff level 23 months before diagnosis. The overall sensitivity and specificity associated with recurrent disease with individual cutoff levels were 79 and 64 percent for carcinoembryonic antigen, 46 and 77 percent for CA 50, and 63 and 76 percent for CA 242, respectively, which was higher than for any other method compared.

Conclusions: With individual cutoff levels, both carcinoembryonic antigen and CA 242 had high sensitivity and acceptable specificity for indicating recurrent colorectal cancer. Because most (29/33) patients with recurrent disease identified within 3 years after surgery and all 8 with hepatic metastases identified during the sampling period were indicated by carcinoembryonic antigen with the individual cutoff, that method was established as the best for postoperative surveillance.