A new artificial intelligence (AI) tool developed by a team at the University of Toronto might be in a position to significantly reduce the time required to create radiation therapy treatment programs for individuals with cancer.
To check the AI-produced relevant treatment programs, the investigators looked at 217 patients with head and neck cancer who had their radiation therapy schedules developed through traditional methods. The plans were similar.
"There are other AI optimization engines which were developed, but the concept behind ours is that it more closely imitates the current medical best practice," states Aaron Babier, the lead author of the research from the University of Toronto Engineering Department.
At the moment, developing radiation therapy plans for each individual patient's tumor can take days, precious time for patients since the cancer often continues to grow and evolve, but also for doctors spending some time designing these intricate treatment strategies.
Head and neck cancers are notoriously hard to design treatment programs for as tumors can be remarkably different from patient to patient. The researchers expect that as the instrument worked so well on this tricky, complex cancer kind, it should have the ability to handle more prevalent tumor types that don't exhibit as much variation, like prostate cancer.
Babier is keen to worry in this case that AI is not supposed to become a replacement for health care professionals, but may save them time by performing some significant groundwork. Once the program has made a treatment program, it would nevertheless be reviewed by means of a radiation physicist and additional altered, taking a couple more hours.
AI is well known to play a major part in the future of cancer diagnosis, monitoring and therapy, but concerns have been raised by some healthcare professionals about the ethics of using machine learning tools to make clinical decisions. 1 such concern published in an article earlier this year at the New England Journal of Medicine by researchers and MDs in Stanford said:
'Physicians must satisfactorily understand how calculations are created, seriously assess the origin of the information used to create the statistical models developed to predict outcomes, understand how the models function and guard against becoming overly dependent on these.'
Herein lies a fairly common issue with new technological developments in medicine nowadays--the demand for MDs to achieve specialist understanding of the brand new diagnostic methods they use so they can fully understand just how much to rely on them to affect their decisions about patients. A similar discussion is ongoing for perpetually-controversial liquid biopsies for cancer.
Despite all these issues, investment in AI from people in the health care industry is commonplace, with huge companies such as Microsoft and IBM using it for various programs currently. Many firms seem to see AI as a possible solution to try and streamline the lengthy and obscenely costly drug-discovery procedure. Toronto-based biotech firm BenchSci has as of today counted 28 pharmaceutical companies and 97 startups currently using AI for their drug-discovery processes.
In the event of using AI to help radiation therapy treatment design, Babier says that his specific instrument is more an expansion of what's currently available to health care employees, rather than a revolution.
"It is basically a fairly simple plugin to aid with what's presently there in a clinical setting but using more intelligent parameters than currently available," said Babier.
The University of Toronto team are not the only ones working on optimizing radiation treatment with AI.