Preoperative Prediction Model for Early Recurrence of Intrahepatic Cholangiocarcinoma after Surgical Resection: Development and External Validation Study

Abstract

Purpose We aimed to develop a preoperative risk scoring system to predict early recurrence (ER) of intrahepatic cholangiocarcinoma (ICCA) after resection, utilizing clinical and computed tomography (CT) features. Materials and Methods This multicenter study included 365 patients who underwent curative-intent surgical resection for ICCA at six institutions between 2009 and 2016. Of these, 264 patients from one institution constituted the development cohort, while 101 patients from the other institutions constituted the external validation cohort. Logistic regression models were constructed to predict ER based on preoperative variables and were subsequently translated into a risk scoring system. The discrimination performance of the risk scoring system was validated using external data and compared to the American Joint Committee on Cancer (AJCC) TNM staging system. Results Among the 365 patients (mean age, 62 +/- 10 years), 153 had ER. A preoperative risk scoring system that incorporated both clinical and CT features demonstrated superior discriminatory performance compared to the postoperative AJCC TNM staging system in both the development (area under the curve [AUC], 0.78 vs. 0.68; p=0.002) and validation cohorts (AUC, 0.69 vs. 0.66; p=0.641). The preoperative risk scoring system effectively stratified patients based on their risk for ER: the 1-year recurrence-free survival rates for the low, intermediate, and high-risk groups were 85.5%, 56.6%, and 15.6%, respectively (p < 0.001) in the development cohort, and 87.5%, 58.5%, and 25.0%, respectively (p < 0.001) in the validation cohort. Conclusion A preoperative risk scoring system that incorporates clinical and CT imaging features was valuable in identifying high-risk patients with ICCA for ER following resection.