Exploratory Radiomic Analysis of Conventional vs. Quantitative Brain MRI: Toward Automatic Diagnosis of Early Multiple Sclerosis

Conventional magnetic resonance imaging (cMRI) is poorly sensitive to pathological changes related multiple sclerosis (MS) in normal-appearing white matter (NAWM) and gray (GM), with the added difficulty of not being very reproducible. Quantitative MRI (qMRI), on other hand, attempts represent physical properties tissues, making it an ideal candidate for quantitative medical image analysis or radiomics. We therefore hypothesized that qMRI-based radiomic features have diagnostic value MS compared cMRI. This study investigated ability cMRI (T1w) qMRI extracted from (WM), NAWM, GM distinguish between patients (MSP) healthy control subjects (HCS). developed exploratory classification models a dataset comprising 36 MSP HCS recruited CHU Liege, Belgium, acquired qMRI. For each type region interest, diagnosis were training subset validated testing subset. Radiomic based entire externally open-source datasets 167 10 MSP. Ranked by best performance was achieved whole WM. Here model magnetization transfer (a qMRI) yielded median area under receiver operating characteristic curve (AUC) 1.00 sub-cohort. type, models, AUCs 0.79 (0.69–0.90, 90% CI) NAWM 0.81 (0.71–0.90) GM. The external validation T1w AUC 0.78 (0.47–1.00) WM, demonstrating large 95% CI low sensitivity 0.30 (0.10–0.70). indicates radiomics could provide efficient information using be useful fast automated check conventional WM abnormalities once acquisition reconstruction heterogeneities been overcome. Further prospective needed, involving more data better interpretation generalization results.