Anil Mistry

My Topol fellowship problem / project:

The rapid development of radiotherapy treatment planning in the last decade has allowed us to target tumours more accurately with radiation beams and spare more of the surrounding healthy tissue. We use sophisticated software to create complex treatment plans that are personalised to each patient. The combination of increasingly complex delivery and growing patient numbers means the workload for physicists and dosimetrists is significant and can involve 10’s – 100’s of checks per treatment plan to ensure that treatment is safe, optimised, clinically acceptable and deliverable. However, errors can occasionally propagate through from planning to treatment when they are missed by manual checks. Not only does this impact upon patient safety and cause delays in treatment starts, but it also increases workload for dosimetry staff, doctors and radiographers alike.

I aim to build a service development framework based around the API of the treatment planning software to automate many of the routine tasks. This will improve treatment plan quality, reporting and the speed of patient throughput, resulting in safer patient care and a reduction in workload to free up staff for further service developments and research.

By developing in-house skills for automating parts of treatment planning and checking, we can then automate audits and implement extensive data mining to allow us to conduct better service development and research. I will also investigate the introduction of AI models into the clinical pathway via the treatment planning system API.

An important outcome of this project is growing the digital skills among all staff working within radiotherapy, so that everyone can become knowledgeable stakeholders and contribute to the improvements that automation offers.