Excited to announce that our computer vision surface imaging breast DIBH project has been funded by the CU Anschutz Cancer Center, with very positive reviews all round.
"In an ongoing cross-institutional collaboration with the Department of Computer Science at CU Boulder and the Department of Bioengineering at CU Anschutz, we are using artificial intelligence (AI) and computer vision to improve real time patient alignment and tracking for breast cancer patients. This project, “Computer vision enhanced breast DIBH-RT “, will allow improved protection of the heart and lungs, lowering the risk of cardiac and lung toxicity and reducing the risk of heart disease and lung cancer."
In this project we will develop, optimize, and quantify the clinical improvement from a pipeline for a patient-specific anatomical and posable skin surface model based on computer vision to enable ‘Avatar-guided DIBH-RT’. The integrated approaches of this study require specific expertise in radiation therapy, biomechanical modelling, and computer vision (a sub-discipline of artificial intelligence). This research collaboration will build logically on our prior biomechanical motion tracking (Gaffney Lab), computer vision research (Gurari Lab) and tumor tracking research (Thomas Lab) to develop an ‘Avatar-guided DIBH-RT’.
"In an ongoing cross-institutional collaboration with the Department of Computer Science at CU Boulder and the Department of Bioengineering at CU Anschutz, we are using artificial intelligence (AI) and computer vision to improve real time patient alignment and tracking for breast cancer patients. This project, “Computer vision enhanced breast DIBH-RT “, will allow improved protection of the heart and lungs, lowering the risk of cardiac and lung toxicity and reducing the risk of heart disease and lung cancer."
In this project we will develop, optimize, and quantify the clinical improvement from a pipeline for a patient-specific anatomical and posable skin surface model based on computer vision to enable ‘Avatar-guided DIBH-RT’. The integrated approaches of this study require specific expertise in radiation therapy, biomechanical modelling, and computer vision (a sub-discipline of artificial intelligence). This research collaboration will build logically on our prior biomechanical motion tracking (Gaffney Lab), computer vision research (Gurari Lab) and tumor tracking research (Thomas Lab) to develop an ‘Avatar-guided DIBH-RT’.
Fig. 1 shows sagittal views of the heart are shown, with inhalation CT scan overlaid with a free breathing scan. Body contours for both are shown (green = inhalation, orange = free breathing), as well as the heart contour (red) at inhalation. (a) shows large expansion of the both the thoracic (24mm) and abdominal (20mm) surface, corresponding to similar motion of the heart (24mm). (b) shows <3mm increase of the thoracic surface, but the large abdominal motion (21mm) is a better surrogate for heart motion (31mm).
Fig. 2 (b) one of five virtual camera views monitoring patient pose. (c) the overlay of the 3D avatar on the 2D image in (b). (d) real-time tracking of avatar, couch, and gantry motion during treatment.
