The effect of T2 signal heterogeneity on gadolinium enhancement in pediatric myelin oligodendrocyte glycoprotein antibody-associated disease: a quantitative magnetic resonance imaging study
摘要
T2-hyperintensity and gadolinium enhancement mark inflammation in myelin oligodendrocyte glycoprotein antibody-associated disease (MOGAD), yet the significance of T2 signal heterogeneity remains underexplored.
ObjectiveTo investigate the relationship between quantitative T2 signal heterogeneity metrics and lesion enhancement.
Materials and methodsWe retrospectively reviewed 70 MOGAD pediatric patients with 190 brain lesions (99 enhanced, 91 non-enhanced). Twenty-six patients underwent follow-up magnetic resonance imaging (MRI) (median 39 days) evaluating 57 baseline-enhanced lesions. Quantitative analysis included lesion area, T2 signal intensity (mean, standard deviation [SD], maximum [Max], minimum [Min]), and normalized ratios to the corpus callosum (CC).
ResultsAcute enhanced lesions exhibited larger axial areas and elevated T2 heterogeneity ratios (Max/CC, SD/CC, Max–Min/CC and Max/Min, all P<0.001) and Mean/CC (P=0.001) versus non-enhanced lesions. Follow-up lesions showed smaller areas (P<0.001) and normalized T2 ratios (Max/CC, SD/CC, Max–Min/CC and Max/Min, all P<0.001) versus their acute-phase counterparts. Analysis of covariance (ANCOVA) identified significant group×lesion area interaction effects for Max/CC, Max–Min/CC, and Max/Min (all P<0.05). Acute non-enhancing lesions and follow-up lesions showed no significant differences across all quantitative parameters (all P>0.05). The Max/Min ratio demonstrated superior diagnostic performance for lesion-level enhancement prediction (area under the receiver operating characteristic curve [AUC] 0.802, 95% confidence interval [CI] 0.739–0.864) and patient-level classification (AUC 0.837, 95% CI 0.739–0.934).
ConclusionT2 signal heterogeneity can suggest inflammatory activity in MOGAD. The Max/Min ratio provides a clinically applicable predictor for lesion enhancement and disease monitoring.
Graphical abstract