Microcomputed Tomography (microCT)‐based Three‐Dimensional (3D) Image Registration to Assess Bone Loss in Spinal Cord Injured Rats

Abstract

BACKGROUND3D image registration is a technique where in‐vivo microCT scans are collected at different timepoints and regions of interest (ROI) are constructed and aligned to improve the precision of determining bone microstructure. In the rodent spinal cord injury (SCI) model, the rapid bone loss occurring at the distal femur precludes the use of standard 3D registration strategies.PURPOSETo (1) adapt a microCT‐based 3D registration protocol to our rodent SCI model, (2) determine the degree of cancellous bone loss at the distal femoral epiphysis after SCI, and (3) assess the effects of bodyweight‐supported treadmill training (TM) or passive bicycle training (PBT) on bone loss after SCI.METHODS16‐wk old male Sprague‐Dawley rats were stratified to receive: 1) T9 laminectomy (SHAM) (n=9), 2) severe T9 contusion (SCI) (n=10), 3) SCI+TM (n=10), or 4) SCI+PBT (n=14). TM and PBT began 1‐wk post‐surgery (post‐sx, two 20‐min bouts/day, 5‐d/wk for 3‐wks). In‐vivo microCT scans were performed pre‐sx and 2‐ and 4‐wks post‐sx. Images were aligned with a 3D registration protocol. ROIs were developed to assess cancellous bone microstructure at the distal femoral epiphysis using two separate protocols that either included or excluded new bone formed by periosteal bone expansion over the 4‐wk experiment.RESULTSDifferences were noted between the ROI protocols, with the ROI that included periosteal bone growth underestimating SCI‐induced bone loss. As such, the results reported hereafter were derived from the ROI that excluded new bone resulting from periosteal bone expansion. No differences in bone outcomes were present in SHAMs at any timepoint except for a slightly higher trabecular separation (Tb.Sp) at 4‐wks (p<.05). At 2‐wks, SCI displayed 14% lower cancellous bone volume (BV/TV) than pre‐sx (p<.01), characterized by 13% lower trabecular number (Tb.N) (p<.05) and 7% higher Tb.Sp (p<.05). Bone loss was more pronounced at 4‐wks after SCI, evidenced by lower trabecular thickness (Tb.Th) and higher Tb.Sp vs 2‐wks (both p<.01). SCI+TM and SCI+PBT displayed a similar magnitude of bone loss to SCI at 2‐wks (1‐wk after starting exercise). Thereafter, SCI+TM displayed no further bone loss, resulting in 9% less BV/TV loss than SCI (p<.01). In comparison, PBT increased BV/TV 15% from 2‐ to 4‐wks (p<.01), due to 5% higher Tb.Th (p<.01) and 12% higher Tb.N (p<.05), ultimately restoring BV/TV to pre‐sx levels. Structural model index (SMI) and trabecular pattern factor (Tb.Pf) increased in SCI (p<0.05) and SCI+TM at 4‐wks (p<.05 to <.01), signifying transition from rod‐like to weaker plate‐like trabeculae and a less connected trabecular network, respectively. In comparison, SMI (p<.05) and Tb.Pf (p<.01) increased in SCI+PBT at 2‐wks before returning to pre‐sx levels by 4‐wks.CONCLUSIONUsing our 3D registration protocol, we determined that SCI causes severe cancellous bone loss and changes indicative of an overall weakening of bone architecture at the distal femoral epiphysis. TM attenuated bone loss at this skeletal site, while PBT promoted new bone formation and restored BV/TV to pre‐sx levels.Support or Funding InformationThis work was supported by the APS UGSRF.