What is the Radiation Planning Assistant (RPA)?

What is the RPA?

Whilst 50% of cancer patients require radiotherapy during their disease course, only 10-40% of patients in low and middle-income countries (LMICs), have access to it. This shortage in specially trained staff is one of the biggest barriers to improving radiotherapy capacity.

Currently, the process of planning radiotherapy is complicated and takes a long time. Although it takes hours in the UK, it can take weeks in LMICs, due to staff and equipment limitations. As the number of cancer cases is predicted to continue rising, even more staff are required. It is predicted that using artificial intelligence (AI) to automate the planning of radiotherapy will help meet demands.

The ARCHERY study plans to evaluate the Radiotherapy Planning Assistant (RPA) that has been developed at the MD Anderson Cancer Cancer Center by Dr Laurence Court and his team. The RPA uses AI to automate the planning of radiotherapy treatment and the mapping of the treatment onto patients' scans. The study will assess the quality of these plans and whether it saves time and money.

How does the RPA work?

The RPA is an web-based AI software that carries out two main process in radiotherapy planning:

1. 1. When radiotherapy is planned, a team of specially trained hospital staff outline the areas of the tumour that need to be exposed to radiotherapy beams. They also make sure non-cancerous tissue is protected and areas where the tumour could potentially spread are treated, if necessary. This is the contouring.
   2. The team will then work out the size, shape and number of radiotherapy beams to most effectively treat the cancer. This is the treatment planning.

Both steps require a lot of expertise and time. Once a patient CT scan is uploaded, the RPA can plan both the contouring and treatment planning on its own. The RPA is ‘trained’ to do this using many previous manual plans. This means it will have ‘learned’ from the expertise of the staff that have created those plans.

The RPA can also be changed to fit the hospital it is being used at. The software can be added to the hospitals existing treatment planning system, so it can be programmed to work with the radiotherapy devices they use. The staff team can then review, edit and approve the plan before it is used for treatment.

Within the ARCHERY study, the AI-generated plans will be reviewed and compared to the manual plans that will be used for patients’ treatment, as normal. The plans and contours will be checked against a criteria of what is currently acceptable for treatment. The automated and manual plans will also be compared for speed of planning. The amount of staff labour and cost will also be compared.

Find out more about the RPA on the MD Anderson website.