WP A1 aims to develop AI-driven adaptive radiotherapy based on cone-beam CT, or CBCT. CBCT is already widely used on radiotherapy machines to check patient positioning, but IMAGINE aims to use it more actively: not only to verify where the patient is, but to help adapt the treatment to the patient’s anatomy on the day of treatment.

The work package focuses on improving CBCT image quality, automatically recognising tumours and organs at risk, optimising treatment plans online and accounting for motion during treatment. AI methods will help turn CBCT images into actionable information, while radiotherapy system expertise will help connect these methods to treatment delivery. In practice, this work could support more precise treatment, smaller safety margins and more efficient workflows.