AI Insight
Researchers developed an AI-enhanced magnetic resonance fingerprinting (MRF) technique to quantify molecular biomarkers associated with myelin, proteins, and lipids in a mouse model of multiple sclerosis. The method successfully detected significant decreases in semisolid magnetization transfer and relayed nuclear Overhauser effect signals in the corpus callosum as early as week 4 of disease induction, before changes were visible with conventional MRI techniques. These imaging biomarkers correlated strongly with lipid concentrations in phantom validation studies and agreed with histological findings in the animal model.
Why it matters
This technique could provide earlier and more specific detection of multiple sclerosis-related tissue damage compared to current clinical MRI methods, potentially enabling better disease monitoring and treatment decisions. The quantitative molecular information may also help distinguish MS lesions from other conditions that appear similar on conventional MRI scans.
Understand the Science
⚠️ Preprint – Noch nicht peer-reviewed
Dieser Artikel wurde noch nicht von unabhängigen Experten begutachtet. Die Ergebnisse sind vorläufig und sollten mit Vorsicht interpretiert werden.
Magnetic resonance imaging (MRI) is the imaging modality of choice for the diagnosis, characterization, and monitoring of multiple sclerosis (MS). Nevertheless, the contrasts manifested by MS lesions often overlap with those of other pathological conditions, highlighting the need for additional disease biomarkers. In addition, while saturation transfer (ST) MRI provides molecular information associated with myelin, protein, and lipids, quantifying the underlying proton exchange parameters remains challenging. Here, we describe a strategy that extends and modifies AI-boosted ST magnetic resonance fingerprinting (MRF) imaging at 7T. This approach was used to quantify the dynamics of the semisolid magnetization transfer (MT) and the aliphatic relayed nuclear Overhauser effect (rNOE at -3.5 ppm and -1.6 ppm relative to water) in a longitudinal cuprizone MS mouse model (n=12). In lipid phantoms, the reconstructed proton volume fractions were strongly correlated with known lipid concentrations across all three proton pools (r>0.96, p<0.001). In vivo, semisolid MT and rNOE proton volume fractions in the corpus callosum demonstrated a significant decrease (p<0.01) as early as week 4 of cuprizone feeding, preceding changes detected by conventional water relaxometry. ST-MRF based biomarkers were in agreement with histological findings. Overall, our results demonstrate the feasibility of rapid, multi-pool ST-MRF quantification for MS characterization.