MRI Tract‐Based Quantitative Susceptibility Mapping of Limbic White Matter Alterations in Preeclampsia

Abstract

BACKGROUND: Preeclampsia is associated with long-term cerebrovascular and cognitive risk, but tract-level white matter (WM) involvement remains incompletely characterized. PURPOSE: To investigate whether preeclampsia is associated with tract-level WM susceptibility alterations on quantitative susceptibility mapping (QSM) and whether these alterations are related to circulating hypoxia-and brain injury-associated proteins. STUDY TYPE: Prospective cohort study with exploratory postpartum follow-up in a small subset. POPULATION: 350 women (152 with preeclampsia, 73 pregnant healthy controls, and 125 nonpregnant healthy controls). FIELD STRENGTH/SEQUENCE: 1.5-T MRI; multiecho gradient-echo sequence for QSM, and 3D T1-weighted turbo field echo, T2-weighted turbo spin echo, T2-weighted fluid-attenuated inversion recovery imaging, and diffusion-weighted echo-planar imaging. ASSESSMENT: Quantitative susceptibility maps were reconstructed using MEDI+0. Mean susceptibility was extracted from 42 XTRACT-defined WM tracts. Plasma hypoxia-inducible factor-1α, amyloid-β1-42, total tau, and phosphorylated tau181 were measured using enzyme-linked immunosorbent assays. Follow up MRI was performed after 1-3 years in 32 women with preeclampsia. STATISTICAL TESTS: Group differences were assessed using analysis of variance with Tukey post hoc testing and false discovery rate correction. Multivariable regression, partial Spearman correlation, and exploratory cross-lagged panel modeling were performed. RESULTS: Significant group differences in susceptibility were concentrated in limbic pathways, including the bilateral anterior thalamic radiations, right dorsal and perigenual cingulum, bilateral temporal cingulum, and bilateral fornix (FDR-adjusted p < 0.05). Compared with pregnant healthy controls, women with preeclampsia had higher susceptibility in the bilateral temporal cingulum (left: 15.75 ± 13.37 vs. 9.95 ± 11.16 ppb, p < 0.05; right: 14.86 ± 14.39 vs. 8.48 ± 11.18 ppb, p < 0.05) and left fornix (5.07 ± 10.12 vs. 1.14 ± 6.65 ppb, p < 0.05). These associations persisted in pregnancy-restricted models after extended adjustment (β = 0.246-0.386; p < 0.05). Higher susceptibility correlated weakly with HIF-1α and total tau (ρ≈0.22-0.27). The exploratory follow-up analyses of 32 women with preeclampsia suggested that baseline biomarker levels were associated with later limbic tract susceptibility. DATA CONCLUSION: Tract-based QSM demonstrated preeclampsia-associated limbic WM susceptibility alterations, mainly involving the temporal cingulum and left fornix compared with pregnant healthy controls. Exploratory postpartum findings require confirmation in larger longitudinal cohorts. TECHNICAL EFFICACY: Stage 2.