AI helps spot previously 'invisible' gray matter MS lesions
A new development in artificial intelligence is giving researchers the ability to spot previously “invisible” multiple sclerosis lesions in gray matter.
Although it has been well-established that lesions in the brain’s gray matter play a significant role in the progression of MS and its resultant symptoms, there has been no reliable way to visualize these abnormalities on imaging. MRI scans are the gold standard for diagnosing and monitoring MS, but up to this point, they were limited to the detection of lesions in white matter only. Now, researchers have developed a technique that combines AI and emerging post-processing methods they believe could help physicians identify cortical lesions.
They published their findings this week in Communications Medicine.
“Detecting previously invisible cortical lesions on conventional legacy MRI scans has major implications for MS research and clinical care,” senior author Robert Zivadinov, MD, PhD, SUNY Distinguished Professor in the Department of Neurology and director of the Buffalo Neuroimaging Analysis Center in UB's school of medicine, said in a news release. “The ability to see for the first time these previously hidden indicators of MS disease progression, including cognitive impairment and disability, is an important advance.”
For their work, the team combined multiple post-processing techniques, including fluid-attenuated inversion recovery squared (FLAIR2), T1/T2 ratio, and artificial intelligence-derived double inversion recovery, to detect lesions. They also included a technique they developed themselves—MMCLE, or multimodal cortical lesion enhancement. These techniques were used alongside transformer-based semantic segmentation (to improve automated detection and delineation) on a collection of MRI exams from the FDA regulatory ORATORIO clinical trial, which is studying MS drug Ocrelizumab.
MRIs from around 700 participants were analyzed using the post-processing technique. Although the individual images did not reveal lesions outside of the white matter, post-processing uncovered multiple gray matter lesions in every participant. An average of 15 to 20 additional lesions—all in the gray matter—were identified. Combined, the technique revealed an additional 11,000 lesions in the patient cohort.
“If you look on the original scans, you generally can’t see the cortical lesions,” explained first and corresponding author Michael G. Dwyer, PhD, associate professor of neurology and biomedical informatics at UB. “But generative AI is very powerful because it can look between the scans and detect tiny differences between them. Because it sees those minor discrepancies, AI can reveal that there’s something going wrong there, that the tissue is not behaving like healthy tissue. The trained models can view multiple MRI images together and synthesize them, and synthesize what had been missing.”
The team suggested that their technique could have a “tremendous impact” on MS research moving forward.
Read more here.
In addition to her background in journalism, Hannah also has patient-facing experience in clinical settings, having spent more than 12 years working as a registered rad tech. She began covering the medical imaging industry for Innovate Healthcare in 2021.