Development and validation of a model for predicting short-term complications after hepatectomy in patients with...

Currently, surgical resection remains the preferred treatment for primary liver cancer. Nevertheless, the incidence of postoperative complications remains considerable. This study sought to develop and validate a model for predicting short-term postoperative complications in patients undergoing hepatectomy for primary liver cancer in order to provide a reference for perioperative management strategies.

A retrospective analysis was conducted on patients with primary liver cancer who underwent hepatectomy in Nantong First People's Hospital from January 2015 to October 2024. Participants were randomly assigned to a training set and a validation set at a 7:3 ratio, and baseline variables were compared. Potential predictors were screened via least absolute shrinkage and selection operator (LASSO) regression analysis within the training set. Subsequently, multivariate logistic regression was employed to identify the independent predictive factors of postoperative complications, which were then incorporated into a predictive nomogram. The performance of the model was evaluated according to receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA) in both the training and validation sets.

A total of 397 patients were included in the study, of whom 136 (34.26%) experienced complications within 30 days after hepatectomy. Four variables emerged as significant predictive factors: operative time [odds ratio (OR) =1.01; 95% confidence interval (CI): 1.00-1.01; P=0.003], lowest intraoperative heart rate (HR) (OR =1.17; 95% CI: 1.12-1.23; P<0.001), lowest intraoperative mean arterial pressure (MAP) (OR =0.93; 95% CI: 0.89-0.97; P=0.001), and prognostic nutritional index (PNI) (OR =0.92; 95% CI: 0.86-0.98; P=0.02). The nomogram constructed with the above four parameters demonstrated robust predictive accuracy, with area under the ROC curve (AUC) values of 0.870 (95% CI: 0.824-0.916) in the training set and 0.874 (95% CI: 0.809-0.939) in the validation set. Calibration curves indicated strong agreement between the predicted and observed outcomes in both sets. DCA confirmed the clinical utility of the nomogram across datasets.

The model we developed could effectively predict short-term complications in patients with primary liver cancer after hepatectomy and was based on four parameters: operation time, lowest intraoperative MAP, lowest intraoperative HR, and PNI. This tool offers valuable support for risk stratification and clinical decision-making in perioperative management.