Assessment of spinal cord injury using ultrasound elastography in a rabbit model in vivo

Abstract

The effect of the mechanical micro-environment on spinal cord injury (SCI) and treatment effectiveness remains unclear. Currently, there are limited imaging methods that can directly assess the localized mechanical behavior of spinal cords in vivo. In this study, we apply new ultrasound elastography (USE) techniques to assess SCI in vivo at the site of the injury and at the time of one week post injury, in a rabbit animal model. Eleven rabbits underwent laminectomy procedures. Among them, spinal cords of five rabbits were injured during the procedure. The other six rabbits were used as control. Two neurological statuses were achieved: non-paralysis and paralysis. Ultrasound data were collected one week post-surgery and processed to compute strain ratios. Histologic analysis, mechanical testing, magnetic resonance imaging (MRI), computerized tomography and MRI diffusion tensor imaging (DTI) were performed to validate USE results. Strain ratios computed via USE were found to be significantly different in paralyzed versus non-paralyzed rabbits. The myelomalacia histologic score and spinal cord Young's modulus evaluated in selected animals were in good qualitative agreement with USE assessment. It is feasible to use USE to assess changes in the spinal cord of the presented animal model. In the future, with more experimental data available, USE may provide new quantitative tools for improving SCI diagnosis and prognosis.

Figure 1

Comparison of multimodal imaging results in representative transverse planes and summary of quantitative results obtained from USE. (A) Schematic of the rabbit animal model and USE imaging of the spinal cord in vivo: relative position of the scan in the animal’s dorsal view and in 3-D. The expected compression field is uniform when sufficient contact is ensured (transducer not shown and tissue approximated as a cuboid for clarity of illustration). (B) B-mode image (column 1), CT image (column 2) and T1-weighted MR image (column 3) alone (rows 1–2) and compounded with axial normal strain elastogram (spinal canal segmented, rows 3–4) from a non-paralyzed rabbit (rows 1 and 3) and a paralyzed rabbit (rows 2 and 4). (C) Mean strain ratios computed from non-paralyzed rabbits (n = 6) and paralyzed rabbits (n = 5); *p < 0.05.

Figure 2

Association between USE and histologic analysis. (A) Axial planes to be analyzed referenced to the vertebral bone model reconstructed from CT. B: T1-weighted MR images compounded with axial normal strain elastograms (spinal canal segmented, row 1) acquired from the designated planes and H&E histological sections (row 2) acquired from the designated planes. In (A, B), the left column corresponds to a non-paralyzed rabbit and the right column corresponds to a paralyzed rabbit. (C) Scatter plots of the strain ratio measured by USE and semi-quantitative histologic score of malacia. The malacia score was evaluated in the dorsal and ventral areas of the spinal cords.

Figure 3

Association between USE and mechanical testing. (A) T1-weighted MR images alone (top) and compounded with axial normal strain elastogram (bottom, with spinal canal segmented) from a non-paralyzed rabbit (left) and a paralyzed rabbit (right). (B) Transverse plane engineering stress–strain curves of spinal cords extracted from the same rabbits as in (A). (C) Scatter plot of mean strain ratio measured by USE and mean Young’s modulus measured by mechanical testing (strain range 10%–20%). Dashed curve indicates a hyperbola fit.

Figure 4

Association between USE and DTI tractography. (A) White matter fiber model reconstructed from DTI tractography alone and (B) compounded with vertebral bone model reconstructed from CT. (C) T2-weighted MR images alone and (D) compounded with axial normal strain elastogram (with spinal canal segmented). The left column corresponds to a non-paralyzed rabbit and the right column corresponds to a paralyzed rabbit. WM White matter.