Researchers at Stanford University in the US have developed an artificial intelligence (AI) tool to facilitate better detection of brain aneurysms.
Based on an algorithm called HeadXNet, the new tool is designed to highlight the areas of a brain scan that could contain aneurysms, which have the potential to leak or burst open and could lead to a stroke, brain damage or death.
Researchers noted that the AI solution improved clinicians’ ability to correctly detect aneurysms and also consensus among the interpreting clinicians.
Stanford University radiology associate professor Kristen Yeom said: “Search for an aneurysm is one of the most labour-intensive and critical tasks radiologists undertake.
“Given inherent challenges of complex neurovascular anatomy and potential fatal outcome of a missed aneurysm, it prompted me to apply advances in computer science and vision to neuroimaging.”
The team trained the AI algorithm using clinically significant aneurysms detectable from 611 computerised tomography (CT) angiogram head scans.
The solution analyses each voxel of a scan for the presence of an aneurysm and generates its conclusions in the form of a semi-transparent highlight.
Stanford University computer science graduate student Pranav Rajpurkar said: “Rather than just having the algorithm say that a scan contained an aneurysm, we were able to bring the exact locations of the aneurysms to the clinician’s attention.”
When eight clinicians assessed HeadXNet by examining 115 brain scans for aneurysm, the tool allowed correct identification of more aneurysms, minimising the miss rate.
However, further studies are required to determine the algorithm’s generalisability before clinical application across various hospital centres. To address this, the team is planning for a multi-centre alliance.
It is expected that HeadXNet’s machine learning technique could also be trained to detect other diseases as well.