MRI guidance technology development in a large animal model for hyperlocal analgesics delivery to the epidural space and dorsal root ganglion

Abstract

Background: Development of new analgesic drugs or gene therapy vectors for spinal delivery will be facilitated by "hyperlocal" targeting of small therapeutic injectate volumes if spine imaging technology can be used that is ready for future clinical translation.

New method: This study provides methods for MRI-guided drug delivery to the periganglionic epidural space and the dorsal root ganglion (DRG) in the Yucatan swine.

Results: Phantom studies showed artifact-corrected needle localization with frequency encoding parallel to the needle shaft, while maximizing bandwidth (125 KHz) minimized needle artifact. A custom constructed 8-12 element surface coil (phased array) wrapped over the spine in conjunction with lateral recumbent positioning achieved diagnostic quality signal to noise ratio at the depth of the DRG and afforded transforaminal access via anterolateral or posterolateral vectors, as well as interlaminar access. Swine epidural anatomy was homologous with human anatomy. Injectate containing 2% gadolinium allowed imaging of injectate volumes in increments as small as 10 microliters and discrimination of epidural flow from intraparenchymal injectate delivery into a DRG. All technical and technological elements of the procedure appear clinically translatable.

Comparison with existing methods: Computed tomographic or fluoroscopic guidance cannot directly visualize drug delivery into the DRG due to contrast medium toxicity, nor reliably identify epidural injection volumes of < 50 microliters.

Conclusions: MRI-guided hyperlocal delivery in swine provides a translatable and faithful model of future human spinal novel drug- or gene therapy vector delivery.

Keywords: Animal model; Epidural; Intraganglionic; MRI-guidance; Spinal injection.

Fig. 1.

Correction of MRI needle artifacts. (a) Photograph of a 22-gauge needle (arrows) fixed to a plastic grid, with its tip at the intersection of the grid lines (outlined in green). (b) The same 22-gauge needle (arrows) is imaged with the frequency encoding direction oblique to the needle shaft. The true location of the needle mounted on the grid is again outlined in green. The needle appears displaced in the right-left dimension from its true location and it appears larger in diameter; both observations are MRI artifacts. (c) Frequency encoding is parallel to the needle shaft resulting in accurate localization of the needle image; this corrects the localization artifact. Furthermore, the size artifact is much reduced by increasing the bandwidth from 31 KHz to 125 KHz. Bar length = 1 cm.

Fig. 2.

Epidural injectate distribution, intraganglionic injectate targeting, and alternate outcome detection. (a–c) Images with sequence [6], Supplementary Table 1. In a, 10 ul of injectate (arrow) is visible adjacent to the DRG. In b and c, 20 ul and 30 ul have been injected, respectively, layering around the DRG (arrows). (d-f) 50 ul of epidural injectate delivered at the L5-6 foramen (imaged using sequence [7], Supplementary Table 1) length = 1 cm. extends cephalad to the L4-5 foramen (d, arrowhead), about the L5 root sleeve at the L5 pedicle level (e, arrowhead) and the exiting L5 DRG (f, arrowhead). (g) A guide needle is contiguous with the DRG (arrow, imaging sequence [5], Supplementary Table 1). (h) Successful hyperlocal delivery. Contrast accumulation is exclusively in the DRG (arrow). (i and j) Alternate outcome. Fat saturated, post infusion image (j, sequence [6], Supplementary Table 1) shows no gadolinium in the DRG (arrow, imaging sequence [6], Supplementary Table 1) and only minimal dorsal epidural gadolinium (arrowhead) consistent with an intravascular loss of injectate.