Dynamic prediction of Radiotherapy toxicities in Head and neck cancer using clinical and imaging data.
Authors
Affiliations (16)
Affiliations (16)
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada. Electronic address: [email protected].
- Centre de Recherche du CHUM, Montréal, QC, Canada; Polytechnique Montréal, Montréal, QC, Canada. Electronic address: [email protected].
- Centre de Recherche du CHUM, Montréal, QC, Canada; Polytechnique Montréal, Montréal, QC, Canada. Electronic address: [email protected].
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada. Electronic address: [email protected].
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada. Electronic address: [email protected].
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada. Electronic address: [email protected].
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada. Electronic address: [email protected].
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada. Electronic address: [email protected].
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada. Electronic address: [email protected].
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada. Electronic address: [email protected].
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada. Electronic address: [email protected].
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada. Electronic address: [email protected].
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada. Electronic address: [email protected].
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada. Electronic address: [email protected].
- Centre de Recherche du CHUM, Montréal, QC, Canada; Polytechnique Montréal, Montréal, QC, Canada. Electronic address: [email protected].
- Centre Hospitalier de l'Université de Montréal (CHUM), Montréal, QC, Canada; Centre de Recherche du CHUM, Montréal, QC, Canada. Electronic address: [email protected].
Abstract
Head and neck cancer radiotherapy (HNC RT) is effective but causes significant toxicities. We aimed to develop a dynamic deep learning model to predict three major HNC RT toxicities-nasogastric (NG) tube placement, hospitalization, and radionecrosis-by integrating clinical data and daily cone-beam CTs (CBCTs), assessing whether serial imaging or dosimetry features improve early prediction. We retrospectively analyzed 1,012 HNC RT patients treated between 2017 and 2022. A multibranch 3D ResNet50 and multilayer perceptron model was trained using 5-fold cross-validation. Inputs included anatomical deformations from daily CBCTs (converted to Jacobian determinant matrices, J<sub>f</sub>), radiomics, and clinical variables (demographics, tumor and treatment details, early weight loss). Each toxicity was modeled using weighted binary cross-entropy loss to address class imbalance. Prediction at the 10th RT fraction was compared with and without J<sub>f</sub> integration. The cohort was 78% male, median age 63 (range 35-84). Primary sites were mainly oropharynx (47%), larynx (19%), and oral cavity (16%). Concurrent chemoradiation was given to 57%, induction chemotherapy to 7%, and postoperative RT to 18%. Incidences of NG tube, hospitalization, and radionecrosis were 16.6%, 4.2%, and 4.6%. Clinical features alone yielded highest predictive accuracy: 70% for NG tube, 67.3% for hospitalization, and 74.2% for radionecrosis. Early J<sub>f</sub> or radiomics did not improve performance. Early weight loss was the strongest predictor. NG tube prediction accuracy improved with later RT fractions (up to 75% at fraction 25). Clinical data combined with weight loss remains the most reliable early predictor without added benefit from imaging data.