A nomogram to predict lymph node metastasis before resection in intrahepatic cholangiocarcinoma.

BACKGROUND:In this study, we developed and validated a nomogram to predict lymph node metastasis before surgery in patients with intrahepatic cholangiocarcinoma (ICC). METHODS:Using the data from January 2006 to January 2015, we enrolled a total of 218 eligible patients with clinicopathologically confirmed ICC as a primary cohort to develop the nomogram. After various variables before surgery were analyzed by multivariable logistic regression, we combined the preoperative carbohydrate antigen 19-9, primary site of tumor, lymphonodus size on computed tomography imaging, tumor growth pattern, and (if applicable) histologic grade to make two different predictive nomograms. Then, the results were validated in 62 consecutive ICC patients from February 2015 to December 2016. We also compared the performance of the different nomograms via calibration, discrimination, and clinical use. RESULTS:The nomogram displayed fine discrimination (the concordance index, 0.761) and fine calibration in the primary cohort. When applied to the validation cohort, the nomogram also showed fine discrimination (concordance index, 0.794) and fine calibration. After adding the histologic grade to the nomogram, the integrated discrimination for predictive performance improved significantly. Finally, the clinical usefulness of predictive nomogram was proven via the decision curve analysis. CONCLUSIONS:The proposed nomograms can be selectively used to achieve more accurate lymph node metastasis predictions before surgery in patients with ICC, and this information can help with clinical management.