HCR.20.046 – AIM a DART: Autonomous Interventional MRI guided aDAptive RadioTherapy

Route: Health care research, sickness prevention and treatment

Cluster question: 104 How do we develop minimally invasive techniques and interventions for diagnosis, prognosis, and treatment?

The hybrid 1.5T MRI radiotherapy system (MR-Linac) enables to adapt the treatment to anatomical changes during radiation delivery for higher precision targeting while better sparing the surroundings.

UMC Utrecht has pioneered this field and MR-Linac treatments are now taking place in clinics around the globe, while various installations in non-academic settings are taking place impacting the radiotherapy landscape in the Netherlands. To increase adoption, productivity and therapy precision, the next technological step in MR guided radiotherapy will be autonomous MR guided radiotherapy. A key enabling technology for this is Artificial Intelligence (AI) similar as taking place in autonomous automotive. AI tools will automate and execute real-time processes needed to generate and interpret MR images and to propose, generate and qualify an adapted treatment. This way, the radiation delivery is continuously optimised for the current anatomy while taking the full radiation delivery history into account.

Such interventional radiotherapy approach requires new skills for multi-disciplinary operators as they become supervisors of an autonomous process. This requires new training and quality assurance procedures as well as the development of robust tools to ensure operator confidence. Intra- and inter-institutional consensus will be built for training and validation of these tools to enhance adoption beyond the academic developments.

Increasingly sophisticated and complex adaptive interventions will be addressed, automated gating and tracking the radiation beam will be followed by repeated intra-fraction re-planning. Similarly, different tumour sites, from rather stationary tumours such as lymph nodes, prostate and breast to more mobile and deforming tumours in the rectum, esophagus and pancreas and severely moving lung tumours will be addressed.

We welcome parties with expertise in medical imaging, radiotherapy clinic and technology, AI and real-time control of processes. We seek stakeholders such as Applied Universities, to shape the educational programs according to the needs of MR guided radiotherapy.

Keywords

MRI Radiotherapy adaptive real-time AI cancer

Submitter

Organisation UMC Utrecht (UMCU)
Name Prof. Dr. Bas W. Raaymakers
E-mail b.w.raaymakers@umcutrecht.nl
Website ww.umcutrecht.nl