This clinical study was conducted to evaluate the predictive value of tumor markers for recurrence and the clinical significance of false positive findings after curative gastrectomy in patients with gastric cancer.

Two hundred ninety patients with gastric cancer who underwent gastrectomy with curative intent were evaluated retrospectively. We analyzed the correlations between changes in tumor markers (CEA, CA 19-9, AFP, and CA-125) and clinicopathologic data, and basis for changes in tumor markers without recurrence during the follow-up period.

The sensitivity, specificity, positive predictive value, negative predictive value, and accuracy of tumor markers for recurrence were 75.0%, 64.6%, 23.1%, 94.8%, and 65.9% respectively. Among 36 patients with recurrences, 10 patients (27.8%) had elevated tumor markers prior to positive findings on imaging studies, while 13 patients (36.1%) had concomitant elevation in tumor markers. At least 1 of the 4 tumor markers increased in 90 of 290 patients during the follow-up period; however, there was no evidence of tumor recurrence. Twenty patients had persistently elevated tumor markers, while the tumor marker levels in 70 patients returned to normal level within 9.08+/-7.2 months. The patients with pulmonary disease, hepatobiliary disease, diabetes, hypertension, or herbal medication users had elevated tumor markers more frequently than patients without disease (P<0.001).

Although detecting recurrence of gastric cancer with tumor markers may be useful, false positive findings of tumor markers are common, so surgeons should consider other chronic benign diseases and medical conditions when tumor markers increase without evidence of recurrence.