Serial Metabolic Evaluation of Perihematomal Tissues in the Intracerebral Hemorrhage Pig Model

Abstract

Purpose: Perihematomal edema (PHE) occurs in patients with intracerebral hemorrhage (ICH) and is often used as surrogate of secondary brain injury. PHE resolves over time, but little is known about the functional integrity of the tissues that recover from edema. In a pig ICH model, we aimed to assess metabolic integrity of perihematoma tissues by using non-invasive magnetic resonance spectroscopy (MRS).

Materials and methods: Fourteen male Yorkshire pigs with an average age of 8 weeks were intracerebrally injected with autologous blood to produce ICH. Proton MRS data were obtained at 1, 7, and 14 days after ICH using a whole-body 3.0T MRI system. Point-resolved spectroscopy (PRESS)-localized 2D chemical shift imaging (CSI) was acquired. The concentration of N-Acetylaspartate (NAA), Choline (Cho), and Creatine (Cr) were measured within the area of PHE, tissues adjacent to the injury with no or negligible edema (ATNE), and contralesional brain tissue. A linear mixed model was used to analyze the evolution of metabolites in perihematomal tissues, with p-value < 0.05 indicating statistical significance.

Results: The perihematoma volume gradually decreased from 2.38 ± 1.23 ml to 0.41 ± 0.780 ml (p < 0.001) over 2 weeks. Significant (p < 0.001) reductions in NAA, Cr, and Cho concentrations were found in the PHE and ATNE regions compared to the contralesional hemisphere at day 1 and 7 after ICH. All three metabolites were significantly (p < 0.001) restored in the PHE tissue on day 14, but remained persistently low in the ATNE area, and unaltered in the contralesional voxel.

Conclusion: This study highlights the potential of MRS to probe salvageable tissues within the perihematoma in the sub-acute phase of ICH. Altered metabolites within the PHE and ATNE regions in addition to edema and hematoma volumes were explored as possible markers for tissue recovery. Perihematomal tissue with PHE demonstrated a more reversible injury compared to the tissue adjacent to the injury without edema, suggesting a potentially beneficial role of edema.

Keywords: intracerebral hemorrhage; magnetic resonance spectroscopy; perihematomal edema; pig ICH model; serial neuroimaging study.

FIGURE 1

A quantitative summary of the change in hematoma and edema volume in a post-ICH pig brain. Panels (A–C) illustrate a representative MRI image showing outlined boundaries of intracranial volume, perihematoma, and hematoma at days 1, 7, and 14, respectively. Panels (D,E) show Perihematoma edema (D) and hematoma (E) volumes. A double asterisk (^∗∗) denotes the significant temporal ipsilesional difference between the time point (p < 0.05). The error bars represent standard error in the mean.

FIGURE 2

Serial FLAIR MRI images overlaid with a multi-voxel grid of corresponding temporal spectra with highlighted voxels of interest (A–C). Panels (A–C) show the PHE, ATNE, and contralesional voxel location at day 1, 7, and 14, respectively, from which the metabolite spectra were measured. MR spectroscopy graphs at days 1, 7, and 14 (columns) are shown for voxels from the PHE (D–F), ATNE (G–I), and the contralesional regions (J–L).

FIGURE 3

Metabolite concentration in the perihematomal edema (PHE), adjacent tissues with no edema (ATNE), and contralesional voxels over two week follow-up post ICH. (A–C) Represent serial NAA, Cr, and Cho concentrations, respectively. The asterisk (^∗) denotes significant (p < 0.001) hemispheric differences, the double asterisk (^∗∗) denotes significant (p < 0.001) temporal difference and (‡) denotes significant (p < 0.001) change between PHE and ATNE. The error bars represent standard error in the mean.