Artificial intelligence to predict chemotherapy toxicities in colon cancer patients

A team of Melbourne based researchers and doctors have successfully used artificial intelligence automated body composition measurements to predict the risk of chemotherapy-related toxicities to improve chemotherapy dosing in cancer treatment.

The Australian-first, retrospective study was performed on stage 3 colon cancer patients receiving oxaliplatin, a chemotherapy drug following surgery at a major Melbourne teaching hospital, Western Health.

Results:

* Between 2012-2021,129 patients were identified.
* Male 53per cent, female 47per cent.
* Mean age – 58 years.
* Dose-limiting toxicity was experienced in 84 (65.1per cent) patients and was higher in females.
* Females had significantly lower muscle indices and higher adiposity (a condition of having too much fatty tissue in the body).
* The optimal oxaliplatin cut point was identified as 3.41 mg/kg lean body mass. 
* Within the first four cycles of chemotherapy, 37 patients experienced dose-limiting toxicity, of which 29 (78per cent) received a dose equal to or greater than the predicted cut-point.

Professor Justin Yeung, a Royal Australasian College of Surgeons Fellow, is Head of the Department of Surgery, Western Precinct, University of Melbourne and is a consultant colorectal surgeon at Western Health.

Professor Yeung, led the multidisciplinary research and said the study has the potential to revolutionise the way chemotherapy drugs are administered to cancer patients around the world.

“For the past three decades we have effectively been using the same drugs and the same approach to determine chemotherapy dosages,” Professor Yeung said.

“Currently, chemotherapy dosages are based on body surface area, a derivative from the patient’s height and weight. Body make-up is not taken into consideration.

“Using artificial intelligence automated body composition measurements, we have been able to predict the risk of chemotherapy-related toxicities.

“It’s a game-changer given 40per cent of colorectal cancer patients require chemotherapy. Of those, up to 70per cent experience some form of chemotherapy-related toxicity, including nerve damage, gastrointestinal disturbances including nausea and diarrhoea, as well as increased risks of severe infection.

“Artificial intelligence is removing the guess work out of chemotherapy dosing, predicting likely toxicities and ultimately benefiting the patient.

“This has the potential to change medicine. Shortly we’ll be expanding the research to include other cancers including breast cancer patients.

“We’re hopeful that in a decade’s time, artificial intelligence using body composition measurements combined with patient characteristics to predict the risk of chemotherapy-related toxicities will be the norm in every Australian hospital.”

Professor’s Yeung’s research was unveiled at the the Royal Australasian College of Surgeons Annual Scientific Congress in Adelaide (1-5 May).

The Congress is the largest multi-disciplinary surgical meeting held in the southern hemisphere and brings together some of the top surgical and medical minds from across Aotearoa New Zealand, Australia, and the rest of the world.

Link to original article.