AN AI SOLUTION FOR EARLY DETECTION OF BRAIN ANEURYSMS

A federal grant for an innovative approach to the early detection of aneurysms sees a collaboration between Macquarie Medical Imaging, Macquarie University and two software development companies turn to artificial intelligence to improve the accuracy of scans.

Approximately two percent of people have a brain aneurysm – with the condition often affecting those in their thirties and forties. While often clinically silent, a brain aneurysm can lead to death or disability if it ruptures or bleeds.

A new research project at MQ Health aims to develop a novel artificial intelligence (AI) approach to early detection of this potentially fatal condition.

Professor John Magnussen, Director of Macquarie Medical Imaging (MMI), and Professor Marcus Stoodley, neurosurgeon with Macquarie University, have teamed up with Fujitsu and GE to develop a revolutionary software system that will allow radiologists to automatically detect brain aneurysms that might otherwise be missed.

“The new tool will use machine-learning to process large amounts of complex 3D images that are challenging even for sub-specialist radiologists to review,” explained Professor Magnussen.

“The changes in blood vessels that indicate an aneurysm can vary significantly in shape, size and location in the brain, adding to the challenge of accurate detection.”

Professor Stoodley said that the new AI tool will be of significant help to those patients who develop the condition.

“Because brain anatomy is so complex, brain aneurysms can be difficult to find on brain scans produced by current diagnostic technology,” he said.

“The new technology will provide radiologists, neurosurgeons and patients with significantly more accurate information about a possible aneurysm. This, in turn, can make the decision to proceed with brain surgery much more clear-cut, with better outcomes for patients.”

The partnership sees MMI provide clinical expertise, data and images to enable Fujitsu to develop prototype software to detect aneurysms. GE will turn the software into products by integrating them into existing CT brain software – Stroke VCAR – to validate the technology and, ultimately, make it accessible to every radiologist across the globe.

Ramy Ibrahim, Head of Co-Creation and Innovation for Fujitsu Australia and New Zealand, said that the company is excited about the newly formed partnership.

“With Macquarie University and Macquarie Medical Imaging just across the road from our headquarters, we hope that this is the first of many partnerships including GE Healthcare,” he said. “We have applied our AI expertise to lung CT in the past and look forward to jointly solving this complex problem for clinicians.”

Regional Research Leader for GE Healthcare Professor Tim O’Meara said that the company is at its best when it partners in this way.

“This kind of collaboration enables us to solve complex problems that make a difference to patients and clinicians,” he said. “In this case, artificial intelligence applied to medical imaging has the potential to improve the accuracy of diagnosis, better track subtle changes over time and address the expertise gap that can exist in many small, isolated or resources-poor radiology practices across Australia and elsewhere.”

While enhanced quality of life is a major benefit of the new technology, in terms of the burden of health on the Australian healthcare system, the ability to diagnose and treat someone in their mid-thirties who has a potentially disabling condition is of great value.