# Patient‐Reported Outcome Measures as Predictive Tools for Disease Control in Chronic Rhinosinusitis With Nasal Polyps: A Prospective Study

## TL;DR

This study shows that combining patient-reported outcomes with clinical data can predict long-term disease control in nasal polyps after surgery.

## Contribution

The study introduces a predictive model using patient-reported outcomes and clinical features to forecast postoperative disease control in CRSwNP.

## Key findings

- Eosinophil and neutrophil counts are key predictors of suboptimal disease control after surgery.
- A logistic regression model using CRS-PRO scores and clinical features achieved an AUC of 0.866 and 83.3% accuracy.
- The model supports personalized risk assessment and postoperative management in CRSwNP patients.

## Abstract

Chronic rhinosinusitis with nasal polyps (CRSwNP) significantly impairs the quality of life, and disease control is now considered the primary treatment goal. Although patient‐reported outcome measures (PROMs) such as the 22‐item Sinonasal Outcome Test (SNOT‐22) and CRS‐PRO are widely used, their utility in predicting long‐term postoperative disease control remains limited.

This prospective follow‐up study aimed to evaluate postoperative recovery and identify the predictors of suboptimal disease control in patients with CRSwNP by integrating preoperative PROMs with objective clinical features. A total of 102 patients with CRSwNP undergoing functional endoscopic sinus surgery (FESS) were enrolled, of whom 89 completed at least 12 months of follow‐up. Preoperative and postoperative PROMs were compared across disease control groups classified based on the European Position Paper on Rhinosinusitis and Nasal Polyps 2020 criteria. Least absolute shrinkage and selection operator regression was applied to select objective clinical predictors, which were then combined with either CRS‐PRO or SNOT‐22 item scores to develop and compare the nine machine learning models. Model performance was assessed using area under the curve (AUC), decision curve analysis, sensitivity, specificity, and other metrics.

Eosinophil and neutrophil counts were identified as key objective predictors of suboptimal disease control after FESS. Among all models, logistic regression incorporating CRS‐PRO scores and selected clinical features achieved the best performance, yielding an AUC of 0.866, accuracy of 83.3%, sensitivity of 72.7%, specificity of 89.5%, and F1‐score of 76.2%. This model demonstrated a strong discriminatory ability and potential utility in individualized clinical decision‐making.

Integrating preoperative CRS‐PRO item scores with selected objective clinical parameters enables the accurate prediction of suboptimal disease control in patients with CRSwNP following FESS. This approach supports personalized risk stratification and postoperative management strategies.