Artificial intelligence could cut hours from radiation treatment

In collaboration with the Press Association

A new artificial-intelligence computer program could cut hours from the meticulous manual process of targeting cancer radiation treatment, say researchers from the Sloan-Kettering Cancer Centre and the Rensselaer Polytechnic in the US.

While early forms of radiation therapy took what scientists describe as a "fire hose" approach, radiotherapy techniques have become increasingly accurate in recent years.

The latest technique, Intensity Modulated Radiation Therapy (IMRT), fires thousands of tiny beams which can be precisely adjusted.

This ensures that the treatment is effective as possible against tumours and also that that the radiation does minimum damage to healthy tissue.

Use of IMRT is now increasingly widespread, but setting up the precise targeting can take many hours.

To speed up the process and reduce the hours of skilled work needed, researchers have now developed a program capable of "learning" to predict patterns from previous examples.

During testing on ten prostate cancer patients, the program was able to produce effective treatment plans for seven patients, spending between five and ten minutes on each.

The team now plans to develop a more robust prototype that can be installed on hospital computers and evaluated in a clinical setting.

They hope to see a clinical prototype in place at Memorial Sloan-Kettering in the next few years.

The researchers also plan to test the approach on tumours that are more difficult to treat with radiation therapy, such as head and neck cancers.

The study is published in the journal Physics in Medicine and Biology.