Probabilistic Assessment of Nerve Regeneration with Diffusion MRI: Validation in Rat Models of Peripheral Nerve Trauma

Abstract

Nerve regeneration after injury must occur in a timely fashion to restore function. Unfortunately, current methods (e.g., electrophysiology) provide limited information following trauma, resulting in delayed management and suboptimal outcomes. Herein, we evaluated the ability of diffusion MRI to monitor nerve regeneration after injury/repair. Sprague-Dawley rats were divided into three treatment groups (sham=21, crush=23, cut/repair=19) and ex vivo diffusion tensor imaging (DTI) and diffusion kurtosis imaging (DKI) was performed 1-12 weeks post-surgery. Behavioral data showed a distinction between crush and cut/repair nerves at 4 weeks. This was consistent with DTI, which found that thresholds based on the ratio of radial and axial diffusivities (RD/AD=0.40±0.02) and fractional anisotropy (FA=0.53±0.01) differentiated crush from cut/repair injuries. By the 12 th week, cut/repair nerves whose behavioral data indicated a partial recovery were below the RD/AD threshold (and above the FA threshold), while nerves that did not recover were on the opposite side of each threshold. Additional morphometric analysis indicated that DTI-derived normalized scalar indices report on axon density (RD/AD: r=-0.54, p<1e-3; FA: r=0.56, p<1e-3). Interestingly, higher-order DKI analyses did not improve our ability classify recovery. These findings suggest that DTI can distinguish successful/unsuccessful nerve repairs and potentially identify cases that require reoperation.