Background context:

Osteoporosis is a risk factor for instrumentation failure in spine surgery. Bone strength is commonly assessed by bone mineral density (BMD) as a surrogate marker. However, BMD represents only a portion of bone strength and does not capture the qualitative dimensions of bone. Recently, the magnetic resonance imaging (MRI)-based vertebral bone quality (VBQ) score was introduced as a novel marker of bone quality. However, it is still unclear if the VBQ score correlates with in-vivo bone microstructure.

Purpose:

The aims of the study were 1) to demonstrate differences in MRI-based (VBQ) and in-vivo (microcomputed tomography; μCT) bone quality between osteopenic/osteoporotic and normal bone, 2) to show the correlation between VBQ, bone microstructure and volumetric BMD (vBMD), and 3) to determine the predictive value of the VBQ score for the prevalence of osteopenia/osteoporosis.

Study design/setting:

Retrospective cross-sectional study.

Patient sample:

267 patients who underwent posterior lumbar fusion surgery from 2014 to 2021 at a single academic institution. Bone biopsies were harvested intraoperatively in 118 patients.

Outcome measures:

VBMD, VBQ score, and bone microstructure parameters derived from μCT.

Methods:

Quantitative computed tomography (QCT) measurements were performed at the lumbar spine and the L1/L2 average was used to categorize patients with a vBMD ≤ 120mg/cm³ as osteopenic/osteoporotic. The VBQ score was determined by dividing the median signal intensity of the L1-L4 vertebrae by the signal intensity of the cerebrospinal fluid using sagittal T1-weighted MRI scans. Intraoperative bone biopsies from the posterior superior iliac spine were obtained and evaluated with μCT. VBQ scores and μCT parameters were compared between the normal and the osteopenic/osteoporotic group. Correlations between VBQ score, μCT parameters and vBMD were assessed with Spearman’s correlation (ρ). Receiver operating characteristic (ROC) analysis was performed to determine the VBQ score as a predictor for osteopenia/osteoporosis. Multiple linear regression analysis with vBMD L1/L2 as outcome was used to identify independent predictors from VBQ, μCT parameters and demographics.

Results:

267 patients (55.8% female, age 63.3 years, BMI 29.7 kg/m²; n=118 with bone biopsy) with a prevalence of osteopenia/osteoporosis of 65.2% were analyzed. In the osteopenic/osteoporotic group the VBQ score, structured model index (SMI), and trabecular separation (Tb.Sp) were significantly higher, whereas bone volume fraction (BV/TV), connectivity density (Conn.D) and trabecular number (Tb.N) were significantly lower. There were significant correlations between VBQ and μCT parameters ranging from ρ=-0.387 to ρ=0.314 as well as between vBMD and μCT parameters ranging from ρ=-0.425 to ρ=0.421, and vBMD and VBQ (ρ=-0.300, p<0.001). ROC analysis discriminated osteopenia/osteoporosis with a sensitivity of 84.7% and a specificity of 40.6% at a VBQ score threshold value of 2.18. Age, BV/TV and trabecular thickness (Tb.Th), but not VBQ, were significant independent predictors for vBMD (corrected R²=0.434).

Conclusions:

This study demonstrated for the first time that the VBQ score is associated with trabecular microstructure determined by μCT. The bone microstructure and VBQ score were significantly different in patients with impaired vBMD. However, the ability to predict osteopenia/osteoporosis with the VBQ score was moderate. The VBQ score appears to reflect additional bone quality characteristics and might have a complementary role to vBMD. This enhances our understanding of the biological background of the radiographic VBQ score and might be a take-off point to evaluate the clinical utility of it as non-invasive screening tool for bone quality.