SNAC To Use Artificial Intelligence for Brain Scan Imaging
Sydney Neuroimaging Analysis Centre (SNAC) has designed artificial intelligence tools to assist radiologists by improving the consistency and speed neuroimagery. The Australian company is looking to use its algorithms for use in clinical settings.
The algorithms produced can perform various tasks which automate a previously manual process. For instance, the director of operations, Chenyu Wang described how a system was designed using Nvidia's DGX-1 system which can perform analysis by isolating brain tissue in MRI scans while performing the task at a much quicker speed.
With a training dataset of over 15,000 three-dimensional CT and MRI images from clinical sites from across the globe, SNAC was able to create the automated process and reducing the time taken from 20-30 minutes to three minutes or even less, while also being more accurate than any human. The reason the task is quite arduous generally is due to the fact that it requires the isolation of the brain from other parts of the head, such as the venous sinuses and fluid-filled compartments around the brain.
Michael Barnett, a neurologist professor at the University of Sydney's Brain and Mind Centre, says that, within one and a half years since SNAC integrated AI into its systems, it enabled some processes to be up to 90-95% fully automated. There is always some human input, of course, for quality assurance, however in terms of productivity and cost the AI has dramatically improved such factors.
But the revolutionary AI-powered systems doesn't stop there. In addition to this, algorithms which can help radiologists examine progressive changes between brain scans have also been developed to see if a patient is improving over time.
Piling 300 slices of brain images together and comparing these between two scans is painstakingly long. Now, however, the process of scrolling up and down through these can be handled by the algorithms in a matter of seconds by immediately highlighting and quantitating changes between subsequent scans.
What's more, Barnett suggested that the algorithms are being used to seek out new biomarkers in MRI and CT scans, which could potentially find and alert the detection of critical abnormalities in the scans to bring it to a radiologist's attention.
The significant results produced by the cutting edge technology is enhancing brain scan accuracy to another degree while reducing the time consumption of these previously laborious tasks. However, Barnett added that we are only beginning to realise the extent to which AI is capable of making an impact on imaging data.