[ACCESS RESTRICTED TO THE UNIVERSITY OF MISSOURI AT AUTHOR'S REQUEST.] Online adaptive image-guided radiation therapy has been a longstanding topic of interest in the field of radiation oncology due to its unique ability to tailor a dose distribution to account for inter-fractional variations and motion of critical structuresthrough daily online re-planning. Efforts are now being made to optimize steps of the adaptive process so that treatment planning and dose delivery can be practically administered while the patient is on the treatment couch. Automated image deformation and segmentation algorithms, along with fast dose calculation and plan re-optimization, have been implemented to streamline the online adaptive treatment planning process. Due to the complexity of inter-fractional anatomical deformations, obtaining precise delineation of target and structure volumes through deformable image registration (DIR) and auto-segmentation is a challenge. Mapping accurate organ at risk (OAR) contours through DIR and auto-segmentation is especially challenging for abdomen and pelvis treatment sites known to have significant interfractionaanatomical variations. While others have studied the accuracy of auto-deformed contours and potential errors and risk factors in the adaptive radiotherapy (ART) process, this study aims to determine if accounting for these errors within specific regions of interest (ROIs) can produce a comparable treatment plan without compromising PTV coverage, OAR sparing or overall plan quality. Once the correlation between dosimetric differences and geometric errors has been identified, a system will be developed to guide the physician in focusing their contour edits to the locations that matter most to the non-deterministic optimization algorithm.
Clinical Implementation of Prostate Image Guided Radiation Therapy: A Prospective Study to Define the Optimal Field of Interest and Image Registration Technique Using Automated X-Ray Volumetric Imaging Software
