Qualitative and quantitative comparison of tissue microstructural and mechanical properties among post-mortem and in-vivo human subjects

Introduction: : Traumatic brain injury (TBI) is a serious health condition affecting tens of millions across both military and civilian populations. Understanding the biomechanics of TBI, however, is challenging due to the ethical concerns associated with performing experimental biomechanical studies in living humans [1]. To circumvent this issue, post-mortem human specimens (PMHS) and animal models are frequently used as alternatives. Animal models, however, cannot replicate the intrinsic biomechanical properties of human brains due to the variations in shape, size and tissue structure. While PMHS models overcome some of these limitations, changes in brain tissue properties are inevitable [2], and thus need to be investigated to understand how the biomechanical response may be different from in vivo. Advances in magnetic resonance imaging (MRI), particularly magnetic resonance elastography (MRE), have enabled the quantification the mechanical properties of tissue, such as stiffness and damping ratio [3], both in vivo and post-mortem. These advances complement diffusion weighted imaging (DWI) methods quantifying microstructural properties and free water in brain tissue, which may elucidate post-mortem tissue degradation processes and the mechanisms behind the changes in mechanical properties. In this novel study, we performed diffusion MRI and MRE on in-vivo subjects and two PMHS to investigate differences in tissue properties and microstructure between in-vivo and PMHS brains. This data provides insights into the limitations of using PMHS for experimental studies of brain biomechanics.

Materials and

Methods: : Two PMHS heads, namely PMHS1 [Age 91y, Female] and PMHS2 [Age 79y, Male], were obtained within 48 hours and 72 hours of death, respectively, with necks sealed to preserve fluids. Based on age range, previously collected 11 in-vivo human subjects were divided into 3 groups: age group 10-30 (N=3), 31-60 (N=4), 61-80 (N=4) [3]. During MRE, an excitation at 50Hz was applied along the anterior-posterior direction of the head while placed on a soft “pillow” driver [Figure a]. MRE phase images [Figure a], acquired using a single-shot echoplanar imaging sequence, were unwrapped and Fourier transformed to obtain displacement fields. Using a nonlinear inversion algorithm [4], stiffness and the damping ratio (DR) were estimated from the displacement data [Figure-a]. The DWI data were acquired using a Philips 3T scanner with TR=155.908s, TE=82ms, at a spatial resolution of 2x2x2mm for b=0 and 1000〖s/mm^2〗 for PMHS1 and for b=0 and 2000〖s/mm〗^2 for PMHS2. For in vivo subjects, data were acquired for b=0 and 1000〖s/mm^2〗. After performing correction for eddy currents, images were corrected for susceptibility induced distortion using DRBUDDI [5]. After denoising the data, free water was eliminated from each voxel and tissue water volume fraction (TF) was estimated using FERNET [6]. Microstructural properties, namely, fractional anisotropy (FA) and mean diffusivity (MD) were calculated from the diffusion tensor estimated from FERNET. The diffusion properties were registered to MRE space for comparison. With the implementation of SLANT [7], the brain regions were extracted from T1-w image.

Results, Conclusions, and Discussions::

Results:
Summary statistics were computed voxel-wise in MRE space. The results showed an increase in stiffness, but a decrease in FA, MD, TF and DR for PMHS models compared to in-vivo subjects [Figure-b]. More specifically, the highest stiffness was found in cerebellum for PMHS1 (7.22±1.84kPa) and in cerebrum for PMHS2 (9.12±2.04kPa) model, both higher than in-vivo subjects [Figure c]. For DR, opposite trends were found in respective regions for PMHS models. While all diffusion properties were lower in both PMHS models compared to in-vivo subjects, MD was lowest in PMHS1 model (MD: 0.13±0.073x10^(-3) 〖mm〗^2/s) whereas FA and TF were found lowest in cerebral region for PMHS2 (FA: 0.150±0.1, TF: 0.078±0.044) model. Comparing white and gray matter properties [Figure d], white matter was found to be stiffer than gray matter in PMHS models compared to in-vivo subjects. A significantly lower (47% lower) TF was observed in white matter of PMHS2 model compared to PMHS1, suggesting tissue degradation and alteration in microstructure in both PMHS models compared to in-vivo subjects, as observed in the spatial images [Figure e].

Discussion:
This is the first study showing both qualitative and quantitative comparison of tissue microstructure, material and diffusion properties for PMHS and in-vivo subjects using DWI and MRE techniques. While all the diffusion properties were found lower in PMHS models, the material properties, namely stiffness, was found higher in both PMHS model compared to in-vivo subjects. Spatial variation in the tissue properties and microstructure could be due the variation in age, subject specificity and time interval after death of the PMHS specimens. One limitation of this study is the low number of the PMHS specimens. In the future, incorporating more PMHS specimens could provide more insights on the change in tissue properties, as well as alteration in brain biomechanics.

Conclusion:
In this study we compared tissue microstructure, diffusion and mechanical properties among in-vivo and 2 PMHS specimens using DWI and MRE. The results showed lower diffusion and viscous properties along with tissue volume fraction, while higher stiffness, in PMHS models compared to in-vivo subjects.

Acknowledgements and/or References (Optional)::

References:
1. Legrand, L. et al Jou of Neuroradiology,Volume 47, Issue 1,2020.
2. Chung BS, J Korean Med Sci. 2019 Aug;34(34):e218.
3. Bayly, P. V. et al. Ann Biomed Eng 49, 2677-2692, (2021).
4. McGarry, M. et al, Med Image Anal 78, 102432, (2022).
5. Irfanoglu MO, Neuroimage 106, 284-299.
6. Parker D, PLoS One. 2020 May 29;15(5).
7. Huo, Y. et al, NeuroImage 194, 105-119 (2019).

Abstract disclaimer:
This work was partially supported by NINDS grant U01-NS112120, the Department of Defense in the Military Traumatic Brain Injury Initiative (MTBI2), the intramural research program of the NIH and an NIH Bench-to-Bedside grant Award Number: 736331. The authors have no conflicts of interest to disclose. The opinions and assertions expressed herein are those of the author(s) and do not reflect the official policy or position of the Henry M. Jackson Foundation for the Advancement of Military Medicine Inc., the Uniformed Services University of the Health Sciences, or the Department of Defense.