Individualized Risk Prediction of Medical Postoperative Complications After Oncologic Hepatectomy: A Nomogram-Based Approach.

BACKGROUND: Liver resection remains the primary curative treatment for many malignant liver diseases. Advances in patient selection, perioperative care, and surgical technique have markedly reduced procedure-related (surgical) complications in experienced centres. However, despite these improvements, medical (non-surgical) complications continue to represent a substantial source of postoperative morbidity, particularly after major liver resections. Herein, we aim to assess the incidence, nature, and predictors of medical versus surgical complications after liver resection and to develop an individual risk calculator for estimating medical morbidity after liver resection. METHODS: This is an observational single-centre study including patients who underwent liver resection for cancer between 2013 and 2025. Postoperative complications were classified into medical and surgical categories based on clinical and diagnostic criteria. Demographic, clinical, and intraoperative data were analyzed to identify risk factors associated with each type of complication, and a multivariate logistic regression model was used to select significant variables, which were imputed in a prediction nomogram made available as an interactive web-based calculator. RESULTS: Of the 231 patients included, 36 patients (15.6%) developed postoperative complications. From multivariate analysis, independent predictors of medical complications included cirrhosis (OR 2.8, 95% CI 1.2-6.8, p < 0.05), operative time > 180 min (OR 2.0, 95% CI 1.1-7.4, p < 0.05), intraoperative blood loss > 500 mL (OR 2, 95% CI: 0.9-4.8, p < 0.05), and ASA score ≥ 3 (OR 3.7, 95% CI 1.1-12.5, p < 0.05). Major hepatic resection was the only independent predictor of surgical complications (OR 7.42, 95% CI: 1.14-48.52, p = 0.036). The logistic regression model demonstrated fair discriminative ability with an AUC of 0.682 (95% CI: 0.544-0.729). The risk-prediction nomogram showed a 24.7% risk of postoperative medical morbidity in patients with all four risk factors vs. a 5.4% risk in patients without any risk factor. CONCLUSION: Postoperative medical complications are significantly more frequent in patients undergoing oncological liver resection with an ASA score ≥ 3, history of cirrhosis, prolonged operative time, and increased intraoperative blood loss. Our logistic regression model and web-friendly nomogram may be used for external validation in larger cohorts and could support preoperative counselling and perioperative risk stratification.