Development and Validation of a Predictive Model for Postoperative Nausea and Vomiting in Thyroid Cancer Patients: A Retrospective Case-Control Study

Background: Thyroid cancer is the most prevalent endocrine malignancy, with surgical intervention being the primary treatment modality. Postoperative nausea and vomiting (PONV) are common complications following thyroid surgery, significantly impacts patient recovery. Consequently, this study analyzed the risk factors for PONV in patients with thyroid cancer post-surgery and developed a predictive model. Methods: This retrospective study involved 393 patients who underwent thyroid cancer surgery at a tertiary-level Class A hospital between July 2024 and February 2025. The patients were randomly divided into a training set (n=275) and a validation set (n=118). A predictive model was constructed using multivariate logistic regression analysis, and a nomogram was developed. The model's discrimination was evaluated using receiver operating characteristic (ROC) curves, while calibration was assessed through calibration curves, and clinical applicability was determined via decision curve analysis (DCA). Results: The overall incidence of postoperative nausea and vomiting (PONV) was found to be 29.01%. The final predictive model developed from the training set incorporated four variables: age (≥50 years), a history of PONV or motion sickness, non-use of ondansetron, and a delay in the time to first oral intake (> 6 hours). The area under the curve (AUC) values for the training and validation sets were 0.76 and 0.68, respectively. Additionally, the calibration curves for both sets exhibited strong consistency. Furthermore, the decision curve analysis (DCA) results indicated good clinical applicability for both the training and validation sets. Conclusion: Patients with thyroid cancer exhibit a heightened risk of postoperative nausea and vomiting, which correlates with age, a history of PONV or motion sickness, the administration of ondansetron, and the time to first oral intake. The PONV prediction model constructed using these variables demonstrates favourable discriminatory ability, calibration, and clinical utility.