The role of deep learning in CBCT-based workflows
,
Italy;
Guillaume Landry,
Germany;
Marco Mueller,
Australia
A Cone-Beam-CT (CBCT) scan is acquired for patient positioning in almost every radiotherapy workflow in our clinics today. It provides a 3D-image of the patient’s anatomy right before the treatment, and exploiting this information potentially enables better targeted and personalised treatment with minimal additional costs. However, geometric and quantitative limitations of CBCT scans make this a challenging task. This session focusses on how machine learning can be applied to CBCT to enable adaptive radiotherapy. The presenters discuss challenges, opportunities, and limitations of machine learning as a tool for correction and quality assurance of attenuation values, image segmentation and CBCT-based workflows.
3130
Symposium
Physics