Risk Prediction And Multimodality Therapy Of Liver Metastases From Gastric Cancer

Purpose The aim of this study is to develop a scoring system for patients with gastric cancer based on the most relevant clinicopatho logical variables to predict the risk of liver metastasis and assess different treatments for gastric cancer patients with liver metastasis (GCLM), finding the optimal therapy combination and personalizing the post-operation follow-up plan.Methods Between January 2004 and December 2013,1,935 patients were enrolled at four centers. Among these patients,176 had gastric cancer with liver metastasis, while the other 1759 patients had gastric cancer without any distant metastases. We use these 1935 patients to build and evaluate the predictive ly scoring system. And then we analyzed the outcome of different treatment combination in 122 patients with complete information and follow-up records who had received gastrectomy, also conducted meta-analyses combining our results with previously published studies.Results According these results of multivariate stepwise logistic regression analysis, the equation was: Z=-1.862-2.966 (well histologic differentiation degree)-3.377 (moderately histotogic differentiation degree)-4.438 (poorly histologic differentiation degree)-5.230 (undifferentiation or signet-ring cell)+2.608 (T2) + 3.086 (T3) + 3.480 (T4a) +4.934 (T4b) + 0.791 (N1) + 1.033 (N2) + 0.886 (N3a) + 0.882(N3b). We chose 80 as the cut-off point between low-risk group and high-risk group. The AUC was 79.7% by logistic regression model and 77.5% by artificial neural network model. Multivariate analysis showed that degree of histologic differentiation (RR=1.397, P=0.020)and type of liver metastases (RR=0.643, P=0.009)were independent prognostic factors of GCLM. The median overall survival of the 122 patients was 15 months:18 months in the 45 patients of group I and 12 months in the 77 patients of group II. The cumulative 1-,3-, and 5-year survival rates were 80.0%,17.8%, and 11.1%, respectively, in group â… and 55.8%,5.2%, and 0, respectively, in group â…¡ (x2= 12.690, P=0.000). The therapeutic composition of gastrectomy plus hepatic resection plus chemotherapy plus TACE resulted in a 5-year survival rates of 33.3%, which is higher than the worldwide average (22.2% according to meta-analyses). Five patients survived for more than 5 years.Conclusions Patients identified as high risk by the predictively scoring system are more susceptible to liver metastasis after gastrectomy and are inclined to have a shorter survival time. They are good candidates for revised follow-up plan and post-operation adjuvant treatment. Metachronous metastases and moderately-differentiated degree of primary tumors were independent determinants for favorable prognosis of patients with GCLM. Hepatic resection is an effective treatment for GCLM patients who have operative indications. Chemotherapy and TACE should be recommended after resection of both gastric cancer and hepatic metastases.