Evaluating the exploratory classification performance of multimodal magnetic resonance imaging in neuropsychiatric systemic lupus erythematosus: a prospective cohort study

TL;DR

This study explores how MRI scans can help identify brain changes in patients with neuropsychiatric lupus, using a combination of imaging and clinical data.

Contribution

The study introduces an exploratory multimodal MRI classification model that combines clinical and imaging data for NPSLE diagnosis.

Findings

NPSLE patients had lower DTI-ALPS values and lumen area, and higher NWI compared to other groups.

The exploratory classification model achieved excellent performance with an AUC of 0.933.

Abstract

To investigate the associations between magnetic resonance imaging (MRI) quantitative parameters with neuropsychiatric systemic lupus erythematosus (NPSLE), and to evaluate the exploratory classification performance for NPSLE.

This study recruited 60 systemic lupus erythematosus (SLE) patients and 20 healthy controls (HC). Participants underwent multimodal MRI and clinical assessment. The diffusion tensor imaging analysis along the perivascular spaces value (DTI-ALPS) and Normalized Wall Index (NWI) were derived using Diffusion Tensor Imaging (DTI) and high-resolution vessel wall imaging (HR-VWI). MRI quantitative parameters were analyzed using Kruskal-Wallis test and Dunn’s test pairwise comparisons. Partial correlation analysis was conducted on MRI parameters and clinical data. An exploratory multimodal classification model was developed.

The NPSLE group showed lower DTI-ALPS values and lumen area (LA), and higher NWI compared to the other groups (P < 0.01). After adjusting for age, gender, and body mass index (BMI), DTI-ALPS correlated negatively with total cholesterol (TC) and immunoglobulin A (IgA) (r = –0.308, P = 0.025; r = –0.301, P = 0.028). LA correlated negatively with low-density lipoprotein (LDL) and disease duration (r = –0.293, P = 0.032; r = –0.474, P < 0.001). The exploratory multimodal classification model combining clinical data, white matter hyperintensities (WMH) status, and MRI parameters showed excellent classification performance (area under the curve (AUC) = 0.933).

NPSLE patients exhibited low DTI-ALPS values and high NWI compared to other groups. The exploratory model based on multimodal MRI quantitative parameters may provide imaging insights into cerebrovascular alterations and glymphatic dysfunction associated with NPSLE.

Figures