Skip to content

Evaluation of Predictive Performance of CT Hounsfield Units and Modified VBQ Scores in Lumbar Pedicle Screw Loosening.

Jul 2026 · Academic Radiology · 0 citations · 41 references
Medicine

Abstract

Rationale

AND

Objectives

Evaluating and Comparing the Predictive Ability of computed tomography (CT)-based Hounsfield units (HU), Magnetic resonance imaging (MRI)-based vertebral bone quality score (VBQ) and modified VBQ scores for pedicle screw loosening (PSL) in patients with lumbar degenerative diseases.

Materials And Methods

Retrospective study of patients undergoing transforaminal lumbar interbody fusion (TLIF) for degenerative diseases (Dec 2019 - Jan 2021). Screw loosening was assessed using 12-month postoperative lumbar radiographs and defined by a ≥1 mm radiolucent halo surrounding the screw. Various VBQ scores and HU value were measured using preoperative MRI and CT, respectively. Predictive performance was evaluated using receiver operating characteristic (ROC) curve analysis.

Results

Among all patients (n = 185), several modified VBQ scores showed higher areas under the curve (AUCs) than the VBQ score (0.776, 95% CI: 0.701-0.851) and CT HU value (0.770, 95% CI: 0.684-0.857), without statistical significance (PAUC>0.05). No significant differences were observed in lumbar disc herniation (LDH) or spondylolisthesis (LS) subgroups. In lumbar spinal stenosis (LSS) patients, the AUC of T1-VBQFAT (0.876, 95% CI: 0.788-0.965) was significantly higher than that of CT HU value (0.718, 95% CI: 0.583-0.852) (PAUC=0.040).

Conclusion

In the overall cohort and other subgroups, HU values and several modified VBQ scores showed numerically higher AUCs without statistical significance (PAUC>0.05). In the LSS subgroup, T1-VBQFAT demonstrated superior predictive performance compared with HU values (PAUC=0.040); notably, this finding did not remain statistically significant after correction for multiple comparisons (PAUC>0.0026). These exploratory results suggest that T1-VBQFAT may serve as a potential adjunct for predicting PSL in patients with LSS.

View source