ABSTRACT
BACKGROUND: We compared the prognostic performance of the International Union Against Cancer/American Joint Committee on Cancer (UICC/AJCC) 7th edition pN stage, number of metastatic LNs (MLNs), LN ratio (LNR), and log odds of MLNs (LODDS) in patients with perihilar cholangiocarcinoma (PCC) undergoing curative surgery in order to identify the best LN staging method. METHODS: Ninety-nine patients who underwent surgery with curative intent for PCC in a single tertiary hepatobiliary referral center were included in the study. Two approaches were used to evaluate and compare the predictive power of the different LN staging methods: one based on the estimation of variable importance with prediction error rate and the other based on the calculation of the receiver operating characteristic (ROC) curve. RESULTS: LN dissection was performed in 92 (92.9%) patients; 49 were UICC/AJCC pN0 (49.5%), 33 pN1 (33.3%), and 10 pN2 (10.1%). The median number of LNs retrieved was 8. The prediction error rate ranged from 42.7% for LODDS to 47.1% for UICC/AJCC pN stage. Moreover, LODDS was the variable with the highest area under the ROC curve (AUC) for prediction of 3-year survival (AUC = 0.71), followed by LNR (AUC = 0.60), number of MLNs (AUC = 0.59), and UICC/AJCC pN stage (AUC = 0.54). CONCLUSIONS: The number of MLNs, LNR, and LODDS appear to better predict survival than the UICC/AJCC pN stage in patients undergoing curative surgery for PCC. Moreover, LODDS seems to be the most accurate and predictive LN staging method.