Construction and Analysis of Risk Prediction Model of Eosinophilic Chronic Rhinosinusitis With Nasal Polyps: ACross-Sectional Study in Northwest China.

Abstract

To provide guidance for clinical endotypes by constructing a risk-predictive model of eosinophilic chronic rhinosinusitis with nasal polyps (ECRSwNP). A cross-sectional study. Single-centre trial at tertiary medical institutions. A cross-sectional study included 343 CRSwNP patients divided into ECRSwNP (n = 237) and non-ECRSwNP (n = 106) groups using surgical pathology. Single-factor and multivariate analysis were used to identify statistically significant variables for constructing a nomogram, including the history of AR, hyposmia score, ethmoid sinus score, BEP and BEC. The model's performance was evaluated based on the receiver operating characteristic (ROC) curve, calibration curve and decision curve analysis (DCA). Allergic rhinitis, hyposmia score, ethmoid sinus score, peripheral blood eosinophil percentage (BEP) and eosinophil count (BEC) were retained for the construction nomogram of ECRSwNP. The nomogram exhibited a certain accuracy, with an AUC of 0.897 (95% CI: 0.864-0.930), good agreement in the calibration curve and a 0.891 C-index of internal validation. Moreover, the DCA with a threshold probability between 0.0167 and 1.00 indicated a higher net benefit and greater clinical utility. The construction of a predictive risk model of ECRSwNP based on easily accessible factors could assist clinicians in more conveniently defining endotypes to make optimal diagnoses and treatment choices.