International trial proves AI can design gold-standard cancer treatment plans

An AI technology is effective at planning the delivery of life-saving radiotherapy for cervical cancer and prostate cancer, according to results from a large international trial led by researchers at University College London (UCL) and the London School of Hygiene & Tropical Medicine (LSHTM).

For cervical cancer, 94 % of deaths occur in low- and middle-income countries (350,000 women died in 2022). Radiotherapy is the main curative treatment, but currently only 10 % of people who need radiotherapy in low-income countries receive it (the figure is 40 % in middle-income nations), with a shortage of skilled professionals one of the major barriers.

The researchers say that by producing high-quality radiotherapy plans quickly, the AI technology saves time and cost, enabling more people to be treated and cured, by bridging the workforce gap.

The trial, called ARCHERY, was conducted at hospitals in India, South Africa, Jordan, and Malaysia, and involved more than 1,000 cancer patients with three main cancers, cervix, prostate, and head and neck. It assessed whether the AI technology can achieve international best-practice radiotherapy planning, usually undertaken by oncologists and physicists, to enable its use across all resource settings, including countries such as the UK.

Planning radiotherapy involves several complex steps requiring often many hours of work from an oncologist (to outline tumours on a CT scan as well as areas at risk of tumour spread, and areas of healthy tissue at risk of radiation damage) and from a physicist (to identify the best position, size and shape of the radiation beam). The AI-based software automatically does this by identifying target structures and determining optimal radiation beam configurations - tasks that traditionally take days or weeks.

The new results, presented at the European Society for Radiotherapy and Oncology (ESTRO) congress in Stockholm, found that the AI technology could plan radiotherapy to a high standard in more than 95% of cervical cancer cases.

For prostate cancer, the technology planned radiotherapy to a high standard in 85% of cases, which would still be considered suitable for routine use in the clinical setting. The results for head and neck cancer will be available later this year.

Ajay Aggarwal, Chief Investigator and Professor, London School of Hygiene & Tropical MedicineThese results show that for cervical cancer, this AI technology achieves a very high standard, supporting its routine use in hospitals globally. In doing so it can help meet the World Health Organisation's cervical cancer elimination initiative for treatment. It can also be used to support the delivery of prostate cancer treatments in any country setting."

Aggarwal added, "In a usual workflow, planning radiotherapy can take many hours over several weeks, as it depends on the availability of specialised staff. This AI technology can reduce that time to just over an hour. This is important as it has the potential to reduce waiting times and widen access to this life-saving treatment."

Co-investigator Professor Mahesh Parmar, Director of UCL's Institute of Clinical Trials, said: "Radiotherapy is a core cancer treatment, helping to cure 40% of cancer cases. Yet millions of people around the world do not have access to it. If they did, we could save more than a million lives a year.

"Our trial also fills an important gap in rigorously testing AI technology for cancer treatment. Unlike our trial, trials testing AI tools are often small-scale, at only one centre, and can have a high risk of bias. Hardly any have been performed outside of high-income country settings, where the need is often greatest.

"This trial highlights the importance of providing the best available evidence to support implementation of the most promising AI technology to deliver improvements in health at scale."

ESTRO President, Professor Matthias Guckenberger, from University Hospital Zurich, Switzerland, who was not involved in the research, said: "Radiotherapy is a vital tool for treating and curing cancer. It is a highly precise treatment, meaning we can carefully target tumours without damaging the rest of the body. The cost a radiotherapy machine can be high, but when used effectively, the benefits to cancer patients are enormous.

"Using AI in this way could not only help treat and cure more cancer patients, but it could also save hospital time and resources. There are obvious benefits to improving the efficiency of cancer services in countries with lower resources, where there may not be enough radiotherapy machines to treat every patient who needs it. However, AI could also be used to help increase the efficiency of cancer services anywhere in the world. We know AI offers enormous potential in healthcare, but we need largescale trials like this one, involving patients in several different countries around the world, to prove that AI is genuinely beneficial to patients and their doctors."

The AI technology, known as the Radiotherapy Planning Assistant, was developed by the University of Texas MD Anderson Cancer Centre led by Professor Laurence Court and team.

Source:

University College London