Novel MRI-Based Pedicle Bone Quality Score Independently Predicts Pedicle Screw Loosening after Degenerative Lumbar Fusion Surgery

Pedicle screw loosening after posterior lumbar fusion is associated with poor bone quality, which often determines screw pull-out strength, insertion torque, and vertebral body loading characteristics. Magnetic resonance imaging (MRI)-based vertebral bone quality (VBQ) score were associated with poor bone quality. Current evidence suggests that pedicle bone quality (PBQ) has a greater impact on screw stability. However, the correlation between MRI-based PBQ score and screw loosening has not been reported.

Purpose: To introduce and evaluate an MRI-based PBQ score to determine its effectiveness in predicting pedicle screw loosening following lumbar fusion surgery.

Methods: The retrospective study analyzed 244 patients who underwent posterior lumbar interbody fusion (PLIF) with pedicle screws between December 2017 and December 2021, with CT and MRI imaging before surgery. Data collected included patient demographics and preoperative radiological data. Radiographic screw loosening was measured at 12 months postoperatively. Clinical assessments included pain visual analog scale (VAS) and Oswestry Disability Index (ODI) scores. The PBQ score was measured using MRI scans. We use univariate analysis for preliminary screening of the risk factors of screw loosening. Subsequent analysis involved multivariate logistic regression to identify independent predictive factors for screw loosening. We constructed the receiver operating characteristic (ROC) curve to ascertain the discriminative capacity of the PBQ score. The area under the curve (AUC) quantified its predictive accuracy. Additionally, we evaluated the association between PBQ score and screw loosening using Spearman's correlation analysis.

Results: Overall, 244 patients who underwent PLIF with pedicle screw fixation participated in this study, including 35 in the loosening group and 209 in the non-loosening group. PBQ score in the loosening group was significantly higher than that in the non-loosening group. On multivariate logistic regression, the higher PBQ score (OR = 8.481, 95% CI: 3.158-22.774; p < 0.001) and the lower mean Hounsfield unit (HU) value of L1-4 (OR = 0.967, 95% CI 0.951-0.984; p < 0.001) were the variables that significantly predicted screw loosening. The AUC for the PBQ score and HU value were 0.751 (95% CI: 0.673-0.828) and 0.702 (95% CI: 0.612-0.791). The PBQ score optimal cutoff to differentiate patients with loosening and with non-loosening was calculated as 3.045 with a sensitivity of 85.7% and specificity of 76.9%, while the optimal cutoff of the HU value was 151.5 with a sensitivity of 64.6% and specificity of 89.5%.

Conclusions: The association between the PBQ score and the propensity for lumbar pedicle screw loosening was found to be substantial. As a predictive measure, the PBQ score outperformed the HU value in forecasting the likelihood of screw loosening post-posterior lumbar fusion.