Title: Aging multiple sclerosis lesions on structural magnetic resonance images
Authors: Elizabeth Sweeney - Weill Cornell (United States) [presenting]
Abstract: Brain structural magnetic resonance imaging (sMRI) is a tool that uses a magnetic field to produce detailed images of the brain. Patients with multiple sclerosis (MS) have lesions in their brains which are visible on sMRI. Part of the 2017 McDonald diagnostic criteria for MS requires that a patient have one clinical MS attack as well as lesions that display dissemination in time and space. Patients would like to have a diagnosis as soon as possible, but showing dissemination in time is difficult from a single MRI scan and requires the invasive technique of administering intravenous gadolinium for an MRI scan. We have developed an algorithm for estimating the age of multiple sclerosis lesions using non-invasive sMRI and Quantitative Susceptibility Mapping (QSM) MRI from a single MRI study. The algorithm creates numerous radiomic features from the sMRI and QSM, performs principal component analysis (PCA) for dimension reduction, and then feeds into a random forest which predicts whether a lesion is older or younger than a year. In a validation cohort, the algorithm achieved an area under the receiver operating characteristic curve (AUC) of .90. The results show promise for revising the diagnostic criteria of MS and allowing for a faster and less invasive diagnosis.