Construction and application of a risk prediction model for sputum viscosity in patients with laryngeal malignant tumors after operation

This study aimed to evaluate the impact of sputum viscosity on postoperative recovery among patients with laryngeal malignancies. Additionally, it sought to identify risk factors associated with postoperative sputum viscosity and to develop a risk prediction model for its occurrence. Convenience sampling was employed to enroll 130 laryngeal malignancy patients hospitalized at a Grade III-A hospital in Shanxi Province between January 2022 and January 2024. Patients were divided into two groups based on postoperative sputum viscosity: the viscous group (n = 52) and non-viscous group (n = 78). General patient data were collected through an electronic medical record system. A logistic regression model was constructed using a 50% cross-validation method, with risk prediction visualization achieved via line graphs. The model’s fit was evaluated using the Hosmer-Lemeshow test, and its predictive performance was assessed through receiver-operating characteristic (ROC) curves. Additionally, 49 surgical patients with laryngeal malignancies treated at another Grade III-A hospital in Shanxi Province between June 2024 and June 2025 were selected for external validation. Data collected between 2022 and 2024 demonstrated a 40% incidence of postoperative sputum viscosity. Significant risk factors identified included age, smoking, placement of a drainage tube, and Ca²⁺ levels. The Hosmer–Lemeshow test indicated a good model fit (x² = 14.617, P = 0.067). The area under the ROC curve was 0.914 (95% CI: 0.749–0.991) with a sensitivity of 0.90, specificity of 0.938, and overall accuracy of 0.891. The clinical application ROC curve results showed that the area under the curve (AUC) was 0.955 (95% confidence interval: 0.906-1.000), accuracy: 0.735, sensitivity: 0.350, and specificity: 1.000. The developed predictive model demonstrates robust performance in identifying individuals at high risk of postoperative sputum viscosity following laryngeal tumor surgery. This model holds potential as a valuable tool for clinical teams to facilitate timely evaluation and management of postoperative sputum viscosity, thereby improving patient outcomes. 1) Postoperative sputum viscosity incidence is notably high among individuals undergoing surgery for malignant laryngeal tumors, with identified risk factors including age, smoking history, drainage tube placement, and Ca²⁺ levels. 2) The proposed risk prediction model exhibits high sensitivity, specificity, and accuracy in forecasting postoperative sputum viscosity. 3) This model can assist healthcare providers in proactive assessment and management strategies, enhancing postoperative care for individuals with laryngeal malignancies.