Fundamental problems in magnetic resonance imaging (MRI)—a critical tool for meeting the challenges of multiple sclerosis (MS)—have limited our ability to diagnose and care for the more two million people worldwide who suffer from the disease. Two recent methods studies by Russell (“Taki”) Shinohara, PhD, offer examples of potentially game-changing techniques he looks forward to advancing in collaboration with clinical epidemiologists.

One paper addresses a biomarker that is very effective for diagnosing MS as opposed to diseases that mimic it: the central-vein sign, which appears in a patient’s white-matter brain lesions. Differences among raters and the time burden of confirming the sign in all of a patient’s lesions make it impractical to use this technique in the clinical setting. Dr. Shinohara and PhD student Jordan Dworkin presented an automated technique for detecting the sign—an efficient way of distinguishing the patients who truly have MS. 

The second paper looks at quantifying a patient’s lesions via MRI—a common approach for measuring a patient’s outcomes. The team, again with Dworkin in a lead role, introduced a statistical technique for counting pathologically distinct lesions that blend together cross-sectionally, a method that overcomes problems associated with existing techniques.

“In order to get these techniques into the clinical setting, we’d want to refine the methods—enhancing accuracy—and do broad multi-center validation studies,” says Dr. Shinohara. “This would give us an idea how well these methods work in the ‘real world,’ with different kinds of MRI scanners and broader populations of interest. It would be fantastic to get clinical epidemiology faculty involved, to help us better design and conduct these validation studies and to learn more about MS by using these tools in larger-cohort studies.”