3-Dimensional quantification of ischemic brain injury in post-cardiac arrest patients with hypoxic-ischemic brain injury.

Abstract

Brain tissue hypoxia and ischemia are central to the pathophysiology of hypoxic-ischemic brain injury (HIBI). Magnetic resonance imaging (MRI) techniques including diffusion-weighted imaging (DWI) and associated apparent diffusion coefficient (ADC) maps offer a non-invasive method for identifying ischemic tissue, but their use is often limited to the provision of summary data. We developed a reproducible, semi-automated analysis pipeline to three-dimensionally quantify and anatomically localize ischemic brain injury using DWI-ADC data in HIBI patients. A retrospective cohort of post-cardiac arrest patients with HIBI (n = 10) was included. MRI preprocessing included skull stripping, spatial normalization to MNI152 space, and anatomical parcellation using FreeSurfer atlases. Ischemic regions were defined by an ADC threshold (<650e^-6 mm²/s), and lesion burden was computed voxel-wise by segment, hemisphere, and tissue type. Susceptibility-weighted imaging (SWI) hypointensities were also quantified. Ischemic injury was spatially heterogeneous but consistently affected the putamen, thalamus, and posterior cortical regions (e.g., pericalcarine and perirolandic cortices). The overall mean relative lesion volume was 28.4 ± 14.1% of total brain volume, with greater involvement in white matter (33.3 ± 17.8%) than gray matter (24.1 ± 11.4%; P = 0.005). Overlap of SWI and ADC hypointensities was minimal (<1% of all voxels), indicating negligible confounding by intraparenchymal blood. We present a robust, observer-independent workflow for anatomically resolved quantification of ischemic injury in HIBI. Atlas-based segmentation integrated with voxel-intensity ADC analysis, enabled generation of granular data on the pattern and extent of ischemic brain injury following cardiac arrest.