Machine Learning-Based Predictive Modeling for the Development of Chronic Rhinosinusitis Using Longitudinal Health
Justina Varghese1, Sicong Chang2, Akshay R Prabhakar1
1Department of Otolaryngology - Head and Neck Surgery, Houston Methodist Hospital, Houston, Texas, USA.
International Forum of Allergy & Rhinology
|June 18, 2026
Abstract
Key Points:
Machine learning identifies reproducible pre-diagnostic CRS trajectories in large EHR cohorts. Models show stable external performance (AUC ∼ 0.81), supporting real-world generalizability. Predictors highlight evolving CRS workups and models may serve as future decision support tools.