Construction and Validation of a Risk Prediction Model for Postoperative Nausea and Vomiting in Patients with Liver Cancer.

Abstract

Background: Postoperative nausea and vomiting (PONV) is a common and distressing complication following liver cancer surgery. This study aimed to develop and validate a risk prediction model for PONV in patients undergoing hepatectomy for hepatocellular carcinoma.

Methods: This prospective study enrolled patients who underwent liver resection between February 2024 and December 2024. Based on risk factors identified through univariate and binary logistic regression analyses, a nomogram prediction model was constructed. The model's discrimination was evaluated using the area under the receiver operating characteristic curve (AUC-ROC) and the consistency index (C-index). Calibration was assessed with calibration curves, and internal validation was performed via the bootstrap method.

Results: A total of 512 patients were included in the modeling cohort. The incidence of PONV was 47.5%. Significant predictors incorporated into the nomogram included age, gender, duration of surgery (min), time of hepatic portal vein occlusion during operation (min), and history of PONV. The model demonstrated an AUC of 0.717 (95% CI: 0.673-0.761), with a sensitivity of 69.9% and a specificity of 62.0% at the optimal cut-off value of 0.413. Bootstrap internal validation yielded a C-index of 0.717, and the calibration curve indicated good agreement between predicted and observed outcomes.

Conclusion: The developed risk warning model shows an acceptable predictive effect in identifying the risk of PONV in patients with liver cancer. It can assist clinical medical staff in early risk assessment and individualized intervention, and has certain predictive value.