BACKGROUND: The aim of this study was to evaluate the risk of liver metastases after radical surgical treatment for gastric cancer, the potential risk factors involved, and the sensitivity of serum tumor markers during followup. STUDY DESIGN: A total of 208 patients who had undergone curative resection for primary gastric cancer and a prospective followup protocol were studied. The association between clinicopathologic variables and hepatic recurrence was investigated using standard univariate methods and multivariate Cox regression analysis. RESULTS: Mean followup time (+/- SD) for the entire patient population was 51 +/- 38 months (median 52 months) and was 88 +/- 24 months (median 81 months) for disease-free patients. Recurrence of gastric cancer was documented in 109 of 208 patients (52.4%). Liver metastases were found in 28 of 208 patients (13.5%); in most of these patients (82.1%) diagnosis was made within 2 years after surgical treatment. The estimated cumulative risk of liver metastases after 5 years was 16.4%. Cox regression analysis identified lymph node involvement (adjusted relative risk [RR] = 6.28, 95% confidence interval [CI] = 2.11 to 18.70, p = 0.001), preoperative positivity for CEA, CA 19-9, or CA 72-4 (RR = 5.18, 95% CI = 1.75 to 15.37, p = 0.003), and intestinal histotype (RR = 3.08, 95% CI = 1.06 to 8.96, p = 0.039) as independent predictors of hepatic recurrence. In 27 of 28 cases hepatic recurrence was associated with an increase in CEA, CA 19-9, or CA 72-4 serum levels above the cutoff, either before or at the time of the clinical diagnosis (sensitivity 96.4%). CONCLUSIONS: Preoperative positivity for serum tumor markers, lymph node involvement, and intestinal histotype are risk factors for liver metastases after radical surgical treatment for gastric cancer. Postoperative measurement of serum tumor markers may be useful for an early diagnosis of hepatic recurrence during followup.
BACKGROUND: The aim of this study was to evaluate the risk of liver metastases after radical surgical treatment for gastric cancer, the potential risk factors involved, and the sensitivity of serum tumor markers during followup. STUDY DESIGN: A total of 208 patients who had undergone curative resection for primary gastric cancer and a prospective followup protocol were studied. The association between clinicopathologic variables and hepatic recurrence was investigated using standard univariate methods and multivariate Cox regression analysis. RESULTS: Mean followup time (+/- SD) for the entire patient population was 51 +/- 38 months (median 52 months) and was 88 +/- 24 months (median 81 months) for disease-free patients. Recurrence of gastric cancer was documented in 109 of 208 patients (52.4%). Liver metastases were found in 28 of 208 patients (13.5%); in most of these patients (82.1%) diagnosis was made within 2 years after surgical treatment. The estimated cumulative risk of liver metastases after 5 years was 16.4%. Cox regression analysis identified lymph node involvement (adjusted relative risk [RR] = 6.28, 95% confidence interval [CI] = 2.11 to 18.70, p = 0.001), preoperative positivity for CEA, CA 19-9, or CA 72-4 (RR = 5.18, 95% CI = 1.75 to 15.37, p = 0.003), and intestinal histotype (RR = 3.08, 95% CI = 1.06 to 8.96, p = 0.039) as independent predictors of hepatic recurrence. In 27 of 28 cases hepatic recurrence was associated with an increase in CEA, CA 19-9, or CA 72-4 serum levels above the cutoff, either before or at the time of the clinical diagnosis (sensitivity 96.4%). CONCLUSIONS: Preoperative positivity for serum tumor markers, lymph node involvement, and intestinal histotype are risk factors for liver metastases after radical surgical treatment for gastric cancer. Postoperative measurement of serum tumor markers may be useful for an early diagnosis of hepatic recurrence during followup.