Clustering CRSwNP Patients for Predicting Uncontrolled Outcomes Based on Clinical Features.

PURPOSE: Chronic rhinosinusitis with nasal polyps (CRSwNP) is highly complex and heterogeneous. Many patients still respond poorly to current medications that combine with surgical treatment strategies, resulting in uncontrolled outcomes. However, identifying uncontrolled CRSwNP remains challenging. We aimed to develop an effective predictive procedure to assess uncontrolled CRSwNP based on clinical features. METHODS: The clinical features of 952 adult CRSwNP patients were subjected to a decision tree analysis, with the uncontrolled outcome at follow-up considered the positive predictive event. RESULTS: A predictive procedure was developed to categorized CRSwNP patients into 6 clusters with different uncontrolled rates. classification indicators were determined as the total computed tomography (CT) scores and age, as well as tissue and blood eosinophil counts . The uncontrolled rates in Clusters 1-6 were 2.75%, 12.31%, 21.28%, 33.16%, 13.54%, and 38.27%, respectively. Additionally, Cluster 1 patients had the lowest tissue and blood eosinophil count and ratio, and the lowest total CT score. Cluster 3 patients had the highest tissue eosinophil count and ratio. Cluster 5 patients >2-fold tissue eosinophil count and ratio than Cluster 2 patients. Cluster 6 patients had the highest value for blood eosinophil count and ratio, total CT score, and endoscopic score. After surgery, the primary disturbing symptoms were nasal congestion (11.01% in Cluster 1 patients and 22.31% in Cluster 2 patients), rhinorrhea/postnasal drip (27.66% in Cluster 3 patients), and olfactory dysfunction (43.68%, 26.56%, and 50.62% in Clusters 4-6 patients, respectively). CONCLUSIONS: The decision tree constructed from the total CT scores, tissue and blood eosinophil counts, and age can generate an effective predictive procedure to guide the identification of uncontrolled CRSwNP.