Systemic BMD to Trajectory-Specific Bone Quality

To the Editor:

We read with great interest the recent article by Ogura et al demonstrating that femoral bone mineral density (BMD) measured by dual-energy X-ray absorptiometry (DEXA) correlates more strongly with pedicle screw insertional torque than lumbar BMD or Hounsfield units (HU) derived from standard lumbar vertebral regions in patients undergoing lumbar fusion surgery. ¹ This finding highlights a well-recognised limitation of lumbar DEXA in degenerative lumbar spine disease, where osteophytes, endplate sclerosis, facet hypertrophy, and vascular calcifications artifactually elevate areal BMD measurements, rendering it unreliable for predicting local fixation strength.^2-4

While the conclusion that femoral BMD reflects systemic bone quality more reliably than lumbar DEXA is instructive, we propose that the more fundamental limitation lies in the inherent inability of any DEXA-based measurement — femoral or lumbar — to capture localised bone quality along the actual pedicle screw trajectory. Pedicle screw fixation strength is determined by pedicle cancellous density, cortical shell integrity, and trabecular architecture specifically at the screw–bone interface^5-7 — factors that neither femoral nor lumbar DEXA can directly assess. The modest correlation coefficient reported by Ogura et al (r = 0.557 for femoral BMD vs insertional torque) supports this contention: even the superior DEXA predictor explains only approximately 30% of torque variance, leaving the majority of fixation strength unaccounted for.

Importantly, the concept of trajectory-specific CT-derived HU measurement is not entirely unexplored. Matsukawa et al published a seminal retrospective cohort study in 2018 demonstrating that regional HU values along the actual screw trajectory (measured within a cylindrical ROI matched to the screw diameter) correlated with insertional torque at r = 0.75 — substantially higher than femoral BMD (r = 0.59) or lumbar BMD (r = 0.55). ⁸ Furthermore, trajectory-based HU was found to be an independent predictor of 12-month screw loosening. More recently, three-dimensional pedicle screw trajectory HU measurement has been shown to outperform vertebral body HU and DEXA-based BMD in predicting screw loosening in degenerative lumbar disease. ⁹

The critical distinction that the work of Ogura et al — and much prior literature — overlooks is that standard HU measurements are derived from mid-vertebral body regions or global lumbar averages, not from the specific bone volume that the screw will engage. We propose that the field’s question should shift from ‘femoral vs lumbar DEXA’ to ‘systemic DEXA vs trajectory-specific CT metrics.’ The latter directly quantifies the cancellous and cortical bone encountered along the planned screw path and is now technically feasible on standard clinical CT with or without dedicated planning software.

We are currently conducting a prospective cohort study comparing lumbar and femoral DEXA T-scores with track-specific CT HU along planned pedicle screw trajectories in adults aged ≥50 years undergoing instrumented lumbar fusion for degenerative disease. For each screw, we are measuring mean and minimum cylindrical-ROI HU, cortical shell thickness at the pedicle–vertebral body junction, and whole-vertebral HU, and correlating these with intraoperative insertional torque, intraoperative screw redirection, cement augmentation rate, and 12-month screw loosening on CT. If our prospective data confirm the superiority of trajectory-specific HU — and replicate the r = 0.75 reported by Matsukawa et al in an independent cohort with standardised DEXA comparison — this would provide compelling evidence that preoperative surgical planning in degenerative lumbar spine disease should transition from DEXA-based assessment to CT-based, trajectory-specific bone quality evaluation. Such findings could justify automated software tools for preoperative track-density mapping and inform future clinical guidelines on instrumentation and augmentation strategies.

We commend Ogura et al for a rigorous in vivo study that adds to the growing evidence base on this important clinical question. We hope that future studies will incorporate trajectory-specific CT metrics alongside DEXA to more definitively characterise their relative predictive value.