Abstract / Summary
Abstract Background Several modifications of the MRI-derived vertebral bone quality (VBQ) score have been proposed to improve assessment of vertebral bone quality, including anterior-modified, posterior-modified, calibrated, and T2-weighted adaptations. However, these methodologies have not been systematically evaluated in the cervical spine. This study compared the performance of VBQ-derived metrics for predicting cage subsidence following single-level anterior cervical discectomy and fusion (ACDF). Methods Patients undergoing single-level ACDF with preoperative MRI and postoperative radiographic follow-up were retrospectively analysed. Traditional VBQ, anterior-modified VBQ (MVBQ), posterior-modified VBQ (PVBQ), calibrated VBQ (cVBQ), and T2-weighted scores were calculated from preoperative MRI. Associations with cage subsidence and predictive performance were evaluated using multivariable logistic regression and receiver operating characteristic analysis. Results A total of 170 patients were included, of whom 67 (39.4%) developed cage subsidence. All T1-weighted bone quality metrics were significantly higher in patients with subsidence and remained independently associated after adjustment for clinical and surgical factors. Among the evaluated methodologies, cVBQ demonstrated the strongest association with subsidence and the highest discriminatory performance (AUC: 0.83). T2-weighted metrics showed weaker predictive ability, with one methodology demonstrating no significant association with subsidence. Conclusion MRI-derived bone quality metrics are associated with cage subsidence following single-level ACDF. Among the evaluated approaches, cVBQ demonstrated the highest predictive performance, supporting its potential role as a practical tool for preoperative risk stratification and identification of patients at increased risk of mechanical complications.