Background: The serum carcinoembryonic antigen (CEA) test is mainly used for postoperative surveillance of colorectal cancer patients in Western and Japanese guidelines, but evidence to support the use of CA19-9 is scarce.
Methods: We analyzed the cohort data from 22 institutions of the Japanese Study Group for Postoperative Follow-up of Colorectal Cancer. Patients who had undergone curative surgery for primary colorectal cancer (pathological stage I-III) between 1997 and 2006 were eligible for analysis. Sensitivities of CEA and CA19-9 at the time of recurrence and the contribution of CA19-9 to detecting recurrences were assessed.
Results: A total of 17,833 patients were eligible, and the overall recurrence rate was 18 %. The sensitivity of CA19-9 in detecting recurrence was lower than that of CEA (29 vs. 57 %). Among patients with recurrence, recurrences were first suspected in 96 % using standard surveillance modalities (CEA elevation, CT scan, clinic visit, and colonoscopy), whereas recurrences were suspected because of CA19-9 elevation in an estimated 1.3 % of patients. With regard to prognosis after recurrences, the sensitivity of CA19-9 was lower than that of CEA in the detection of surgically treatable recurrences (22 vs. 49 %). In terms of overall survival after recurrences, CA19-9 and CEA had almost comparable hazard ratios (1.66 and 1.48, respectively).
Conclusions: Our data suggested that the sensitivity of serum CA19-9 test is low, and that adding it to the current standard surveillance strategies is not beneficial.
Keywords: CA19-9; Colorectal cancer; Recurrence; Surveillance.