Quantitative susceptibility mapping (QSM) and R2* of silent cerebral infarcts in sickle cell anemia

Abstract

Silent cerebral infarction (SCI) is the most commonly reported radiological abnormality in patients with sickle cell anemia (SCA) and is associated with future clinical stroke risk. To date, there have been few histological and quantitative MRI studies of SCI and multiple radiological definitions exist. As a result, the tissue characteristics and composition of SCI remain elusive. The objective of this work was therefore to investigate the composition of segmented SCI lesions using quantitative MRI for R2* and quantitative magnetic susceptibility mapping (QSM). 211 SCI lesions were segmented from 32 participants with SCA and 6 controls. SCI were segmented according to two definitions (FLAIR+/–T1w-based threshold) using a semi-automated pipeline. Magnetic susceptibility (χ) and R2* maps were calculated from a multi-echo gradient echo sequence and mean SCI values were compared to an equivalent region of interest in normal appearing white matter (NAWM). SCI χ and R2* were investigated as a function of SCI definition, patient demographics, anatomical location, and cognition. Compared to NAWM, SCI were significantly less diamagnetic (χ = –0.0067 ppm vs. –0.0153 ppm, p < 0.001) and had significantly lower R2* (16.7 s−1 vs. 19.2 s−1, p < 0.001). SCI definition had a significant effect on the mean SCI χ and R2*, with lesions becoming significantly less diamagnetic and having significantly lower R2* after the application of a more stringent T1w-based threshold. SCI-NAWM R2* decrease was significantly greater in patients with SCA compared with controls (–2.84 s−1 vs. –0.64 s−1, p < 0.0001). No significant association was observed between mean SCI–NAWM χ or R2* differences and subject age, lesion anatomical location, or cognition. The increased χ and decreased R2* in SCI relative to NAWM observed in both patients and controls is indicative of lower myelin or increased water content within the segmented lesions. The significant SCI–NAWM R2* differences observed between SCI in patients with SCA and controls suggests there may be differences in tissue composition relative to NAWM in SCI in the two populations. Quantitative MRI techniques such as QSM and R2* mapping can be used to enhance our understanding of the pathophysiology and composition of SCI in patients with SCA as well as controls.

Longitudinal Changes in Iron and Myelination Within Ischemic Lesions Associate With Neurological Outcomes: A Pilot Study

“…To date, studies that have used QSM and R2* relaxometry analyses for cerebrovascular diseases have adopted a cross-sectional design. 12–14 In these studies, ischemic lesions demonstrated increased magnetic susceptibility and decreased R2* values compared with normal-appearing white matter areas, suggesting a lower myelin content within the ischemic lesion. However, the tissue composition of iron and myelin could not be cross-sectionally inferred because the distribution of iron and myelin differs in each anatomic region of the brain.…”

Quantitative susceptibility mapping identifies hippocampal and other subcortical grey matter tissue composition changes in temporal lobe epilepsy

“…There are a range of dipole inversion methods to choose from and, after comparison of several state‐of‐the‐art direct and iterative methods, iterative Tikhonov regularisation (Karsa et al, 2020), non‐linear total variation (TV; Milovic et al, 2018) and weak harmonic QSM (WH‐QSM) (Milovic et al, 2019) were selected. Iterative Tikhonov was chosen for its applicability to head (and neck) imaging (Karsa et al, 2020) and its use in clinical QSM research (Murdoch, Stotesbury, Kawadler et al, 2022; Murdoch, Stotesbury, Hales et al, 2022). Total variation‐based approaches were shown to be the most accurate in the QSM Challenge 2.0 (Bilgic et al, 2021) and non‐linear TV (FANSI), in particular, was chosen because it scored the highest in Stage 2 of the Challenge.…”