ESTRO 2024 - Abstract Book

S4563

Physics - Machine learning models and clinical applications

ESTRO 2024

Results:

Accuracy, AUC and κ for generated RF/LR models were 0.89±0.12/0.90±0.09, 0.98±0.03/0.97±0.03 and 0.83±0.19/0.86±0.13, respectively. According to the interpretation by Landis and Koch, κ values of 0.63/0.69 point at ‘substantial agreement’ between clinician plan labeling and ML prediction. Both classifiers found gender, age, smoke and total dose of anthracyclines the most important features for ML prediction.