Probability of Lymph Node Metastases in Patients Undergoing Adequate Lymphadenectomy during Surgery for Intrahepatic Cholangiocarcinoma: A Retrospective Multicenter Study.

Introduction: Nodal metastases (lymph node metastasis [LNM]) are one of the major determinants of prognosis following surgery for intrahepatic cholangiocarcinoma (ICC). Previous studies investigating the correlation between clinical-radiological features and the probability of LNM include patients undergoing inadequate nodal sampling. Aim of this study was to develop a model to predict the risk of LNM in patients undergoing adequate lymphadenectomy using preoperative clinical and radiological features. Methods: Patients undergoing radical surgery for ICC with adequate lymphadenectomy at seven Italian Centers between 2000 and 2023 were collected and divided into a derivation and a validation cohort. Logistic regression and dominance analysis were applied in the derivation cohort to identify variables associated with LNM at pathology. The final coefficients were derived from the model having the highest c-statistic in the derivation cohort with the lowest number of variables included (parsimony). The model was then tested in the external validation cohort, and the linear predictor was divided into quartiles to generate four risk categories. Results: A total of 693 patients were identified. Preoperative CA 19-9, clinically suspicious lymph nodes at radiology, patients' age, and tumor burden score were significantly associated with LNM. These factors were included in a model (https://aicep.website/calculators/) showing a c-statistic of 0.723 (95% CI: 0.680, 0.766) and 0.771 (95% CI: 0.699, 0.842) in the derivation and validation cohort, respectively. A progressive increase of pathological lymph node positivity across risk groups was observed (29.9% in low-risk, 45.1% in intermediate-low risk, 51.5% in intermediate-high risk, and 87.3% in high-risk patients; Conclusions: A novel model that combines preoperative CA 19-9, clinically suspicious lymph nodes at radiology, patients' age, and tumor burden score was developed to predict the risk of LNM before surgery. The model exhibited high accuracy and has the potential to assist clinicians in the management of patients who are candidate to surgery.