Vertebral bone density variation and site-specific computed tomography measurements for risk of subsequent fracture
Dana Niry1, Khalil Salame2, Zvi Lidar2
1Department of Radiology, Tel Aviv Sourasky Medical Center, Affiliated to the Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel.
Background:
Percutaneous vertebral augmentation (PVA) is an established treatment for selected patients with osteoporotic vertebral compression fractures (VCFs) but carries a substantial risk of subsequent vertebral fractures (SVFs). Opportunistic bone mineral density (BMD) assessment using CT-derived Hounsfield units (HU) enables site-specific vertebral measurements, which may provide additional insight into fracture risk beyond whole-vertebra or dual-energy X-ray absorptiometry (DXA)-based averages. This study aims to determine the distribution of vertebral body density along the spinal column, and to evaluate the predictive value of low site-specific preoperative HU measurements for SVF risk following cement augmentation.
Methods:
We conducted a retrospective case-control study of patients who underwent PVA for osteoporotic VCFs between 2007 and 2017. HU values were measured in three locations on preoperative computed tomography (CT) vertebrae within thoracic, thoracolumbar, and lower lumbar spine. HU attenuation values were compared between the SVF and non-SVF groups, by univariant Cox regression and General Estimating Equation (GEE) model, for SVF risk factor prediction.
Results:
Sixty-six patients (50 women, mean 77.9 years) underwent single-level thoraco-lumbar percutaneous cement augmentation (kyphoplasty/vertebroplasty) for low-energy osteoporotic compression fracture. Twenty-two patients (33.3%) had SVF, within 31.2 months. Across the cohort, mean HU values were significantly different between locations (P<0.001): below superior endplate (57 HU), middle vertebral body (85.9 HU) and superior to inferior endplate (81.9 HU). Significant differences were also noted between mean HU values at thoracic spine (T5-T10, 100.6 HU), thoracolumbar junction (T11-L2, 80 HU) and lower lumbar levels (L3-L5, 75.7 HU) (P<0.001). Lower mean HU values were observed in the SVF group (47.22 vs. 53.86 HU, P=0.63 and 74.74 vs. 85.11 HU, P=0.36, respectively). Fractures of the L2 vertebra correlated significantly with predicted risk of SVF (P=0.012). In GEE analyses, no clinical variables correlated with SVF risk.
Conclusions:
Significant novel site-specific findings include: lowest bone density at superior aspect of cancellous vertebral body in thoraco-lumbar junction; L2 vertebral fractures, a specific risk factor for SVF. No significant correlation between low site-specific bone-density and the risk of SVF was identified.
