June 3 (UPI) -- In some scenarios, artificial intelligence is better than humans at identifying preferable outcomes for cancer radiation treatment, researchers found in a study published Thursday in Nature Medicine.
Researchers at the University of Toronto found radiation treatment plans created by AI were selected over human-based treatments 72 percent of the time in a blind clinical study.
In addition, the radiation treatments designed by artificial intelligence were found to be 60 percent faster than those produced by human counterparts.
"We have shown that AI can be better than human judgement for curative-intent radiation therapy treatment," study co-author Christ McIntosh said in a press release.
Related
"In fact, it is amazing that it works so well," said McIntosh, chair of medical imaging and AI at the Joint Department of Medical Imaging and the University of Toronto.
Researchers stressed the use of such technology is easier to control in a lab setting than in a traditional clinical setting, but expressed optimism that it will be utilized.
"There has been a lot of excitement generated by AI in the lab, and the assumption is that those results will translate directly to a clinical setting," study co-author Dr. Tom Purdie said in the release.
"But we sound a cautionary alert in our research that they may not," said Purdie, a radiation oncologist at the Princess Margaret Cancer Center and professor at the University of Toronto.
While the results of the clinical study published in Nature Medicine could have immediate implications on radiation therapy, researchers say the technology has broader applications.
Artificial intelligence is increasingly being investigated for use at improving medical diagnostics and treatment.
A Finnish team recently found artificial intelligence was able to better assist in determining medication outcomes for cancer fighting agents.
A German group similarly found data analysis undertaken by algorithms for X-rays and MRIs to produce more efficient diagnoses.
AI has also recently been used to identify heart disease and Type 2 diabetes.
