The Needs and Benefits of Continuous Model Updates on the Accuracy of RT-induced Toxicity Prediction Models within a Learning Health System

Development of outcome prediction models from clinical data can form the foundation for a learning health system offering precision radiotherapy. We hypothesize that regular prediction model updates along with prospective data collection is important to maintain the prediction accuracy. Prediction models for grade ≥2 xerostomia were developed by bivariate logistic regression. Four methods of model updating (sliding training period/increasing training period/conditionally increasing training period/no updates) were compared. Updating prediction models was effective for maintaining the prediction performance.
Source: International Journal of Radiation Oncology * Biology * Physics - Category: Radiology Authors: Tags: Clinical Investigation Source Type: research