Dissociated Magnetic Susceptibility and White Matter Microstructure in Type 2 Diabetes with Peripheral Neuropathy

Patients with diabetic peripheral neuropathy (DPN) have been confirmed to have related changes in brain structure. This study aims to compare white matter microstructural and iron-related alterations in type 2 diabetes mellitus (T2DM) patients with and without DPN.

Participants included T2DM (n = 26), DPN (n = 42), and healthy controls (HCs, n = 37). Neuropsychological assessments, electromyography, diffusion tensor imaging (DTI) and quantitative susceptibility mapping (QSM) were performed. DTI metrics of white matter tracts and magnetic susceptibility values (MSV) of deep brain nuclei were compared across groups. Associations between MSV abnormalities and clinical parameters in DPN were assessed, and the diagnostic performance of MSV-based features for distinguishing DPN from T2DM was evaluated using receiver operator characteristic curve (ROC) analysis.

Compared to the T2DM group, the DPN group showed significantly higher values in specific neuroimaging indices: in DTI, these included mean diffusivity in the left medial lemniscus (P = 0.017) and axial diffusivity in the left posterior limb of the internal capsule (P = 0.026); in QSM voxel-based analysis, MSV in the Cerebellum_Crus1, 7b, Crus2 and 8 regions (initial P < 0.001, cluster level P_FDR < 0.05) were elevated; and in QSM ROI-based analysis, significant increases were found in the right pulvinaris superficialis (Pu.sf) (t = -2.750, P_FDR = 0.023) and the left lamella medialis interpolaris (La.m.ip) (t = -3.009, P_FDR = 0.011). In DPN group, partial correlation analysis showed that cerebellar MSV with intergroup differences (DPN > T2DM) was positively correlated with fasting blood glucose (r = 0.397, P = 0.012); MSV in the right Pu.sf was negatively correlated with triglyceride (r = -0.325, P = 0.044) and positively correlated with high density lipoprotein cholesterol (r = 0.384, P = 0.016). ROC results showed that the fusion model (voxel cerebellar MSV + right Pu.sf + left La.m.ip) had the highest discriminative efficacy among all models (area under curve = 0.888, sensitivity = 0.857, specificity = 0.808).

DPN showed the separation between relatively subtle DTI changes and more significant QSM abnormalities in the thalamic cerebellar circuit, suggesting that the magnetic susceptibility changes in this pathway may be more sensitive to detect DPN.

1. DPN reveals a dissociation between subtle DTI changes and marked QSM abnormalities in the thalamic-cerebellar circuit. 2. Iron-related alterations in this sensory pathway may serve as a potential imaging marker for DPN.