Prostate cancer is the most frequently diagnosed type of cancer in men in Australia and is the leading cause of cancer death. Recently, researchers developed a diagnostic tool powered by artificial intelligence that can detect asymptomatic prostate cancer in seconds.
Men frequently avoid the doctor, delaying diagnosis tests until it is too late. Their research was published in the journal Scientific Reports. Now, an artificial intelligence (AI) programme developed at RMIT University in Melbourne, Australia, may allow for early detection of the disease through routine computed tomography (CT) scans.
The technology, which was developed in collaboration with clinicians at St Vincent’s Hospital Melbourne, works by analysing CT scans for tell-tale signs of prostate cancer, something that even the most trained human eye struggles to do.
Due to the high radiation doses involved in CT imaging, it is not suitable for routine cancer screening, but the AI solution could be used to conduct a cancer check whenever men have their abdomen or pelvis scanned for other reasons.
Dr Ruwan Tennakoon of RMIT said that while CT scans were excellent for detecting bone and joint problems, even radiologists had difficulty detecting prostate cancers on the images.
“We’ve trained our software to see what the human eye cannot, with the goal of detecting prostate cancer by chance,” he explained.
“It’s similar to training a sniffer dog in that we can teach the AI to see things that our own eyes cannot, just as a dog can smell things that human noses cannot.”
Prostate cancer is a slow-growing disease that is typically discovered incidentally, which means it can go undiagnosed for years. By 2020, it was estimated to be responsible for approximately 12% of male cancer deaths in Australia.
How it operates
Researchers from RMIT and St Vincent’s Hospital Melbourne examined CT scans of asymptomatic patients with and without prostate cancer for the study, which was published in Nature’s Scientific Reports.
The researchers trained the AI software to look for disease-related features in a variety of scans and to pinpoint their location precisely, avoiding the need to manually crop the images.
The AI outperformed radiologists who viewed the same images, quickly detecting cancerous growths.
Additionally, the AI improved with each scan, adapting and learning to read images from a variety of machines in order to detect even the tiniest irregularities.
Professor John Thangarajah, Head of Artificial Intelligence at RMIT, said the study demonstrated how AI can and should be used to advance the public good.
“Our health care system requires smarter solutions, and AI can help, but we’re just getting started,” he said.
“There is a great deal of good that artificial intelligence can do for the world, which is our focus at RMIT, and this study is a significant part of that.”
Dr Mark Page, Head of CT in Diagnostic Imaging at St Vincent’s Hospital Melbourne, stated that early prostate cancer intervention is critical for a positive health outcome.
“Australia does not have a prostate cancer screening programme, but with this technology, we hope to detect cases early in patients who are undergoing other types of scanning,” he explained.
“For instance, emergency patients undergoing CT scans could be screened for prostate cancer concurrently.
“If we can detect it earlier and refer them to specialist care more quickly, we can improve their prognosis significantly.”
The technology has the potential to be scaled up and integrated with a variety of diagnostic imaging equipment such as MRI and DEXA machines – subject to additional research.
“It was fantastic to leverage RMIT’s AI expertise, and we look forward to future opportunities to analyse additional radiology scans,” Page said.
The multidisciplinary team, which includes researchers from RMIT’s School of Engineering and School of Computing Technologies, is seeking commercial partners to develop software that will further integrate the AI technology with hospital equipment in preparation for possible clinical trials.
(Only the headline, some content and picture of this report may have been reworked by the Mixpoint Team; the rest of the content is auto-generated from a syndicated feed. The meaning of the content has not been altered in any way.)