Introduction

Numerous studies have quantified cadaveric intervertebral disc structure, composition, and mechanical behavior and interpreted results in the context of in vivo disc; however, there are substantial differences between in vivo and ex vivo discs. In many ex vivo studies, segments are pre-loaded and hydrated, to mimic the physiological in vivo condition. However, the differences between the physiological and experimental conditions are unknown, limiting interpretation of cadaveric study outcomes. The objective of this study is to assess geometric and hydration differences between physiological and experimental conditions using repeated magnetic resonance imaging (MRI) in the porcine spine.

Methods

Repeat 3T MRI were acquired on 3 lumbar discs from Yucatan minipigs in 3 conditions: physiological (intact torso <2 hours since sacrifice, n=4 pigs, 12 discs), multilevel (fresh-frozen intact excised spine, n=8 pigs, 24 discs), and segment (fresh-frozen vertebra-disc-vertebra motion segment, n=8 pigs, 24 discs). Each sample was in PBS-soaked gauze during thawing, imaging, and freezing. The physiological condition was reasonably represented by the fresh torso because in vivo animal MRI is performed under anesthesia in supine position which minimizes axial load and eliminates active muscle contractions [1].

Two MRI sequences were used, a T₁-weighted FLASH sequence for geometry and a T₂-weighted CPMG sequence for T₂ relaxation time. Disc volume was calculated from T₁ image segmentations using ITK-SNAP and mid-sagittal height, width, and area were calculated using custom MATLAB code (Fig 1A). T₂ times, positively correlated with hydration, were evaluated in the nucleus pulposus (NP) and adjacent vertebral bodies by fitting the intensity in the regions of interest to noise-corrected exponentials [2] (Fig 1B). Mixed model fits with fixed effect of condition, random effect of specimen, and post-hoc pair-wise Tukey honest significance tests were conducted with p < 0.05.

Results and Discussion

Physiological in vivo spines are constrained and loaded by surrounding tissues and body forces that are progressively removed when spines are dissected for ex vivo testing. This study showed that dissection altered disc geometry with increased disc volume, increased height, and reduced width (Fig 2A-C). While area was not significantly changed, the increased volume is likely due to the increased height, assuming soft tissue dissection reduced physiological constraints mostly in the axial direction. The reduced width could be related to release of annulus fibrosus circumferential pre-strain. The decrease in NP T₂ indicates reduced water content in the analyzed region: this has two potential causes, 1) water redistribution within the larger volume disc and/or 2) water loss to the vertebral bodies, which had increased T₂ time (Fig 2D). Importantly, these geometric and hydration changes will alter segment mechanical behaviors. While the effects of spinal dissection have been preliminarily quantified here, further work to evaluate the differences in physical constraints and boundary conditions between in vivo and ex vivo states is ongoing, with the ultimate goal of restoring cadaveric spinal segments to the in vivo physiologic state for testing and modeling.