Quantifying cortical lesions in large legacy multiple sclerosis clinical trial MRI datasets using multi-contrast post-processing and deep learning

Multiple sclerosis (MS) is a chronic neurological disease affecting both white and gray matter of the central nervous system. Despite the well-established history of gray matter involvement in MS, cortical lesions are almost never evaluated in clinical trials because of limitations in the feasibility of magnetic resonance imaging (MRI) to visualize them. Recently, a number of post-processing methods, including synthetic contrasts and artificial intelligence (AI)-based approaches, have shown potential for enhancing cortical lesion detection on conventional MRI data. These methods have the potential to reanalyze existing clinical trial data to answer key mechanistic questions about both MS development and about treatment effects. Therefore, we evaluated three of the most promising of them - FLAIR2, T1/T2 ratio, and AI-DIR - and introduced a new combined contrast called multi-modal cortical lesion enhanced (MMCLE). We also harnessed transformer-based semantic segmentation to improve automated detection and delineation of these lesions. Using the data from the large, multicenter, phase 3 ORATORIO trial, we confirmed that cortical lesions can be clearly visualized and quantified with these methods. At baseline, we detected 14.8+/-20.72 lesions per participant, 86.0% true positive rate, 8.4% false positive rate across subjects for blinded MMCLE, using simultaneous review of all contrasts as the reference. Using deep learning, we also confirmed that the simultaneous use of multiple contrasts improves quantification.