Australian National University researchers develop AI tool to classify brain tumours quicker and more accurately
Australian National University (ANU) researchers have developed an artificial intelligence (AI) tool able to more quickly and accurately classify brain tumours.
The current gold standard for brain tumour identification is profiling based on DNA methylation, which can take weeks.
Lead researcher Dr Danh-Tai Hoang said the new AI tool, called DEPLOY, classifies brain tumours into 10 subtypes in an hour.
Identifying brain tumours is essential in making a treatment plan, so the longer it takes the longer a patient must wait to begin treatment.
“The time it takes to do [DNA methylation] testing can be a major drawback, often requiring several weeks or more when patients might be relying on quick decisions on therapies,” Dr Hoang said.
“There’s also a lack of availability of these tests in nearly all hospitals worldwide.”
The tool, created in collaboration with experts from the National Cancer Institute in the United States, draws on microscopic pictures of a patient’s tissue.
DEPLOY was trained and validated on a dataset of about 4,000 patients across the US and Europe, and had an “unprecedented accuracy of 95 per cent”.
“When given a subset of 309 particularly difficult to classify samples, DEPLOY was able to provide a diagnosis that was more clinically relevant than what was initially provided by pathologists,” Dr Hoang said.
“This shows the potential future role of DEPLOY as a complementary tool, adding to a pathologist’s initial diagnosis, or even prompting re-evaluation in the case of disparities.”
Dr Tim Wang of the University of Sydney’s Faculty of Medicine and Health has worked in the AI and medical imaging space for about seven years.
Dr Wang said in his research into the use of AI-based monitoring for multiple sclerosis MRIs, and his work as co-founder of Sydney Neuroimaging Analysis Centre, he’s seen the value AI can bring to repetitive diagnostic tasks.
He said using AI for technical processes, like analysing microscopic images of patients’ tissue, makes it faster and helps reduce human error.
“A lot of those diagnostic processes are really long, and also technically complex, and this is exactly where the opportunity with the current AI technology is,” Dr Wang said.
He said while the tool is very helpful to identify different tumours, diagnoses and treatment plans should still be coming from a trained medical professional.
“For the diagnosis, and even for the monitoring, it has to be working together with a human expert,” Dr Wang said.
“The final call, I believe, still needs to be a human for now, but the AI can very rapidly accelerate that process.”
Dr Hoang said he hopes to one day see the technology used more widely to identify different types of cancers.
“We are expanding this work to other cancer types as well, like we want to classify breast cancer into other some types,” he said.
SOURCE: ABCNEWS