Abstract
Lower field strength images produced on the onboard MR system have better soft tissue contrast than photon-based onboard image modalities, like cone beam computed tomography images, but have lower quality when compared to diagnostic MR images (1.5 tesla and above). The pulse sequence and field strength used to capture the images are new to radiation oncology. The purpose of this study was to investigate the changes of radiomics texture features extracted from low field strength (0.35 tesla) magnetic resonance (MR) images for early prediction of disease response. Radiomics texture features are predefined statistical descriptors that quantify the frequencies of patterns within radiographic images by describing image heterogeneity. Features that are predictive of response during the early stages of MR-guided radiotherapy (MRgRT) could provide supplemental information to physicians to allow for modification of the remaining treatment regime. By providing patient specific response prediction, delta-radiomics texture analysis could support individualized cancer treatment for each patient using information produced during a clinical workflow routine.