C-VBQ and CT hounsfield units for DXA-defined low bone mass screening and their associations with bone turnover...

This study aimed to evaluate the relationships among the Cervical Vertebral Bone Quality score (C-VBQ), mean Hounsfield units on cervical CT (C-HU), the lowest T-score from dual-energy X-ray absorptiometry (DXA), and bone turnover markers (BTMs). It also compared the performance of C-VBQ and C-HU in screening for DXA-defined low bone mass.

This retrospective study enrolled 91 patients undergoing surgery for degenerative cervical spondylosis who had preoperative cervical MRI, CT, DXA, and BTMs. Based on the lowest DXA T-score, patients were categorized as having normal bone mass (T ≥ -1.0), osteopenia (-2.5 < T < -1.0), or osteoporosis (T ≤ -2.5). Differences in C-VBQ and BTMs across groups were analyzed. Spearman correlation and multiple linear regression assessed associations between C-VBQ, imaging parameters, and BTMs. Additional partial correlation analyses were performed after controlling for age, sex, and BMI, and Benjamini-Hochberg false discovery rate correction was applied to the correlation heatmaps. Receiver operating characteristic (ROC) curves evaluated the ability of C-VBQ and C-HU to identify low bone mass (T < -1.0).

C-VBQ showed significant differences across DXA groups (P < 0.001), while C-HU and the lowest T-score decreased progressively with worsening bone status (P < 0.001). C-VBQ correlated negatively with C-HU (ρ = -0.523, P < 0.001) and the lowest T-score (ρ = -0.512, P < 0.001), and C-HU correlated positively with the lowest T-score (ρ = 0.613, P < 0.001); these core imaging-based associations remained significant after FDR correction. After adjustment for age, sex, and BMI, C-VBQ remained associated with C-HU (partial r = -0.421, P < 0.001) and the lowest T-score (partial r = -0.419, P < 0.001). A weak association was observed between C-VBQ and β-CTX after adjustment (partial r = 0.221, P = 0.039), but the corresponding unadjusted heatmap association did not survive FDR correction (FDR q = 0.090). Multiple linear regression identified age (P = 0.016) and β-CTX (P = 0.037) as independent correlates of C-VBQ. For screening low bone mass, the area under the curve (AUC) was 0.702 for C-VBQ (optimal cutoff: 3.48; sensitivity 0.492, specificity 0.906) and 0.773 for C-HU (optimal cutoff: 324.57 HU; sensitivity 0.729, specificity 0.781).

In patients undergoing cervical spine surgery, C-VBQ demonstrates consistent associations with C-HU and DXA T-scores, supporting its potential for opportunistic preoperative bone quality assessment. The association between C-VBQ and β-CTX was weak and exploratory. Integrating radiological indicators such as C-VBQ and C-HU with DXA and BTMs may enable a more comprehensive, but still cautiously interpreted, perioperative bone health evaluation.