Trabecular bone score predicts distal femur screw pull-out strength better than DXA bone mineral density, CT...

The purpose of the study was to evaluate trabecular bone score (TBS) technology for orthopedic application (TBSortho) as a predictor of a screw pull-out strength in a cadaveric model. This study finds that TBSortho correlates more strongly with the screw pull-out strength compared to cortical density, computed tomography (CT) Hounsfield units (HU), and dual-energy X-ray absorptiometry (DXA) bone mineral density (BMD).

Patient bone quality impacts the mechanical strength of surgical fixation constructs. Preoperative assessments of bone quality and the potential for a screw pull-out are important for surgical planning as well as postoperative rehabilitation protocols. Previous work has correlated the screw pull-out strength with the screw insertional torque, cortical thickness, and CT HU. TBS is a gray-level textural metric that can be extracted from a two-dimensional DXA scan, improves fracture prediction, and may evaluate the mechanical competence of both the cortical and trabecular bones. The purpose of the study was to evaluate TBS technology for TBSortho as a predictor of the screw pull-out strength in a cadaveric model.

Twenty paired, fresh-frozen cadaver femurs stripped of soft tissue were obtained (5 M, 5 F specimens, age range of 56-96 years). Standard clinical femur CT were performed to obtain HU, cortical thickness, and cortical density. DXA was also performed using a novel analysis technique as distal femur DXA is not routinely acquired clinically. DXA data were used to generate TBSortho values in two distal femur regions of interest. All femurs then underwent a screw pull-out testing with five lateral distal femoral 5-mm locking screws (n = 100 screws total). The correlation coefficient from Spearman tests and R-squared of the fixed effects from the linear mixed effects models were calculated.

TBSortho was found to correlate most strongly of CT and DXA measures with the screw pull-out strength, having marginal R² and standardized beta of 0.75 and 0.87 in the proximal screw cluster and 0.67 and 0.83 in the distal screw cluster, respectively. TBSortho accounted for 75% variance in the pull-out strength. CT HU and DXA bone mineral density (BMD) did not have a statistically significant correlation with the screw pull-out strength.

This study finds that TBSortho correlates more strongly with the screw pull out strength in a cadaveric distal femur model compared to cortical density, CT HU, and DXA BMD. These preliminary results suggest that TBSortho may be a valuable tool to model mechanical integrity of bone preoperatively.