By applying a machine learning approach, scientists can now differentiate neuromyelitis optica from multiple sclerosis on magnetic resonance imaging scans of patientsâ€™ gray matter.
By applying a machine learning approach, scientists can now differentiate neuromyelitis optica (NMO) from multiple sclerosis (MS) on magnetic resonance imaging (MRI) scans of patients’ gray matter (GM). A study conducted by Arman Eshaghi, MD, of the Queen Square MS Center at the Institute of Neurology at the University College London, UK, and colleagues, published in the journal Neurology in November 2016, detailed the process.
The researchers had three aims: first, to learn whether or not GM measures from MRI scans could differentiate the two conditions; second, to find out if the same technique could be used at multiple centers; and third, to investigate which GM measures contributed to the differentiation. Although there were three aims, the researchers said, “Our primary question was whether imaging biomarkers extracted from routine MRI measures discriminate between MS and NMO.” They added that the evidence from this study is Class II.
There were 144 participants in total; 90 from one center in Tehran, Iran, and 54 from Padua, Italy. The researchers said that although previous studies have demonstrated differences in MRI scans of patients with MS and NMO, “an automatic distinction is still challenging.” They also noted, “Here, we automatically classified patients with MS or NMO on the basis of their brain MRI scans routinely acquired with clinical protocols, using a random-forest classifier.”
“Our findings showed that GM imaging measures, such as cortical thickness, cortical surface area, and subcortical GM volumes, led to an accuracy of 74% when classifying the 2 patient groups, which is higher than that obtained with each GM measure on its own,” reported the researchers. The results were more accurate when distinguishing healthy controls from patients with either MS or NMO.
The researchers concluded, “We showed that the random-forest classification robustly and automatically discriminates between MS and NMO on the basis of MRI scans in a 2-center setting,” adding “Furthermore, deep GM volumes and cortical thickness of specific key regions may give increased power to detect subtle GM features, which may facilitate the differential diagnosis between MS and NMO.”