A simple and effective prognostic staging system based on clinicopathologic features of intrahepatic cholangiocarcinoma

Incidence and mortality of intrahepatic cholangiocarcinoma (ICC) are increasing. However, its prognostic predictive system associated with outcome after surgery remains poorly defined. In this study, we conducted retrospective survival analyses in a primary cohort of 370 patients who underwent partial hepatectomy for ICC (2005 and 2009). We found that seven variables were significantly independent predictors for overall survival (OS): serum prealbumin (hazard ratio [HR]: 1.447; p = 0.015), carbohydrate antigen 19-9 (HR: 1.438; p = 0.009), carcinoembryonic antigen (HR: 1.732; p = 0.002), tumor number (HR: 1.781; p < 0.001), vascular invasion (HR: 1.784; p < 0.001), regional lymphatic metastasis (HR: 2.003; p < 0.001) and local extrahepatic metastasis (HR: 1.506; p = 0.008). Using these independent predictors, we created a simple clinicopathologic prognostic staging system for predicting survival of ICC patients after resection. The validity of the prognostic staging system was prospectively assessed in 115 patients who underwent partial hepatectomy between January 2010 and December 2010 at the same institution. The prognostic power was quantified using likelihood ratio test and Akaike information criteria. Compared with the 6^th and 7^th AJCC staging systems, the new staging system in the primary cohort had a higher predictive accuracy for OS in terms of homogeneity and discriminatory ability. In the validation cohort, the homogeneity and discrimination of the new staging system were also superior to the two other staging systems. Conclusions: The new staging system based on clinicopathologic features may provide relatively higher accuracy in prognostic prediction for ICC patients after tumor resection.