Development and validation of a clinical risk prediction model for chronic sinusitis with nasal polyps: a retrospective analysis.

OBJECT: This study aims to explore risk factors associated with chronic sinusitis with nasal polyps (CRSwNP) by analyzing the clinical pathophysiological factors in chronic sinusitis (CRS) patients.

METHODS

General information of 206 CRS patients was collected along with Visual Analog Scale (VAS), snot score, Lund Kennedy and Lund Mackay assessments. Univariate analysis was performed to select variables for multivariate regression. Optimal model was selected by Akaike Information Criterion (AIC) to identify key risk factors. LASSO algorithm was used to determine the final variables, then a nomogram was constructed. The nomogram scores were assigned via an integral line and a total score was calculated, and risks were predicted based on the probability scale. Model quality was assessed using the ROC curve. The calibration curve was employed to compare predicted versus actual risks, and the Decision Curve Analysis (DCA) was used to evaluate the model's clinical utility.

RESULTS

Significant differences exist between groups regarding height, alcohol consumption, bronchial asthma, nasal septum deviation, eosinophil ratio, LM score, and LK score. LASSO analysis revealed alcohol consumption, allergic rhinitis, deviated nasal septum, bacterial culture, LM score, and LK score as risks. AUC (0.96) and 95% CI (0.915-1.0) indicated good model accuracy. Calibration curve showed that predicted probability closely aligned with perfect agreement, and DCA indicated a broad effective range for clinical applicability.

CONCLUSION

Our study revealed significant risks for nasal polyps, including alcohol consumption, allergic rhinitis, deviated nasal septum, positive bacterial cultures, LM score, and LK score. The constructed model showed satisfactory prediction of CRSwNPs.