Early prediction of tumor recurrence after curative resection of gastric carcinoma by measuring soluble E-cadherin

BACKGROUND. Currently, to the authors' knowledge, there is no serum marker to predict disease recurrence after patients undergo curative resection for gastric carcinoma. Previous reports have indicated that serum levels of soluble E-cadherin had prognostic value in these patients. The objective of the current study was to determine whether soluble E-cadherin levels could predict disease recurrence in patients with gastric carcinoma who underwent curative surgery. METHODS. Sixty-nine patients who under-went curative surgery for gastric carcinoma after December 1997 were followed prospectively. Venous blood samples were collected preoperatively, 1 month after surgery, and every 3 months thereafter. The blood samples were assayed for soluble E-cadherin and for carcinoembryonic antigen (CEA) using commercial enzyme-linked immunosorbent assay kits. Receiver operating characteristic (ROC) curves were used to define a cut-off level of E-cadherin for the optimal sensitivity and specificity for predicting disease recurrence. RESULTS. The median follow-tip was 21 months for patients with recurrent disease (it 17 patients) and 36 months for patients without recurrent disease (n = 52 patients; P = 0.007). The optimal cut-off level of E-cadherin was 10,000 ng/mL. The sensitivity for predicting prediction disease recurrence using this cut-off level at 3 months and at 6 months postsurgery was 47% and 59% respectively, which was significantly better compared with the sensitivity of CEA using the conventional cut-off level (6% at 3 months postsurgery and 6% at 6 months postsurgery; P = 0.004 and P < 0.0001, respectively). The median time between the elevated E-cadherin level and documented disease recurrence was 13 months (range, 3-20 months), compared with 4 months (range, 1-20 months) for CEA. CONCLUSIONS. Serum soluble E-cadherin was a good marker for predicting disease recurrence in the first 3-6 months after surgery, with a median of 13 months before clinical recurrence. The use of this marker may allow time for vigilant surveillance and consideration of adjuvant therapy.