Modeling relationships between CT radiomic features improved distant-metastasis prediction over conventional radiomics across independent head and neck cancer cohorts.
Combined radiomics, foundation-model and clinical features predicted grade ≥2 radiation proctitis with 0.865 AUC on an independent imaging device.
Adding CTV radiomics increased internal AUROC from 0.507 to 0.754 for predicting poor response to rectal chemoradiotherapy.
Planning CT features added prognostic information for local and distant control but did not significantly improve overall or progression-free survival prediction.
A multi-regional clinical-radiomics model reached an external-test area under the curve of 0.908 for 5-year recurrence prediction.
A multimodal imaging, radiomics, and clinical deep learning model showed moderate external validation for survival risk after Gamma Knife radiosurgery.
N4 bias correction with z-score normalization produced the most stable radiomic features on a 0.35 tesla magnetic resonance linear accelerator.
In 24 surgically confirmed cases, three radiomic features differed nominally between pure osteoradionecrosis and osteoradionecrosis with recurrence.