skip to main content
Research Highlight | December 08, 2020

Team turns to machine learning to predict patient response to therapy, accelerate precision medicine

MRIs, CT scans and ultrasounds all help doctors and researchers see cancers with astonishing accuracy, revealing important information that helps guide cancer detection, diagnosis and treatment. But what if the images produced by these tools hold more information than the human eye can see?

That is a question that Dr. Gregory Czarnota, a radiation oncologist at Sunnybrook Health Sciences Centre in Toronto, and his team of Terry Fox-funded researchers are trying to answer. To do so, they’ve enlisted the help of an unlikely ally: machine learning.

Sometimes referred to as artificial intelligence (AI), machine learning is a term used to describe “computer systems that are able to learn and adapt without following explicit instructions, by using algorithms and statistical models to analyze and draw inferences from patterns in data,” according to the Oxford English Dictionary. In other words, it is a tool that helps researchers analyze huge amounts of complex information and find patterns and connections that a human eye might never see. 

Dr. Czarnota and his team are amongst a number of pioneering cancer researchers using this technology to analyze images of a patient’s tumour. In doing so, they’ve been able to make impressive findings that can help determine how cancer patients will react to treatment.

In a recent paper published in Scientific Reports, for instance, the team was able to deploy this new approach to determine how patients with locally advanced breast cancer (LABC) would respond to neoadjuvant chemotherapy, based on an analysis of quantitative CT (qCT) scans. In the end, they demonstrated that qCT biomarkers can be used to predict LABC tumour response to chemotherapy pre-treatment, “with high sensitivity and specificity.”

In other recent published studies the team applied machine learning to predict how patients with brain metastasis will respond to radiotherapy based on MRI images and how patients with locally advanced breast cancer will respond to neoadjuvant chemotherapy by using quantitative ultrasound.

Being able to predict how a cancer patient will respond to chemotherapy or radiation is an important step towards personalizing cancer treatments, says Dr. Czarnota.

“Such predictions of therapy response could facilitate the development of personalized medicine, which is expected to improve survival and quality of life for cancer patients.”

Recent papers

Recent papers in this area by the team include:


This research was partially funded by The Terry Fox New Frontiers Program Project Grant in Ultrasound and MRI for Cancer Therapy with funds from the Lotte & John Hecht Memorial Foundation.