Accuracy of MRI-Based Vertebral and Endplate Bone Quality in Predicting Cage Subsidence in Anterior Cervical Spine Surgery: A Systematic Review and Meta-Analysis

Introduction: Conventional radiographs and bone mineral density are commonly used to predict cage subsidence, though their accuracy can be limited by factors several factors. MRI-based Vertebral Bone Quality (VBQ) and Endplate Bone Quality (EBQ) offer a more detailed assessment of bone quality and subsidence risk. This study aims to evaluate the accuracy of MRI-based VBQ and EBQ in predicting cage subsidence in anterior cervical spine surgery patients. Methods: A comprehensive search was conducted across databases including PubMed, Europe PMC, ScienceDirect, and Google Scholar using keywords such as “Magnetic Resonance Imaging”, “MRI”, “Vertebral Bone Quality”, “Endplate Bone Quality”, “Cage Subsidence”, “Anterior Cervical Spine Surgery”, “ACDF”, “ACCF”, combined with Boolean operators "AND" and "OR" up to November 2024. The risk of bias was assessed using Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2), with the primary outcome focusing on VBQ and EBQ values, and secondary outcomes examining the AUC, sensitivity, and specificity of VBQ and EBQ. Results: Nine retrospective cohort studies involving 911 patients were included. Pooled analysis showed that high VBQ and EBQ scores were significantly associated with increased subsidence risk (OR 2.63, 95% CI: 1.51–4.60, p = 0.0007). Pooled sensitivity and specificity for VBQ were 78.9% (95% CI: 72.5%–85.3%) and 85.7% (95% CI: 82.2%–93.1%), respectively, with an AUC of 0.865. EBQ demonstrated pooled sensitivity of 81.8% and specificity of 88.7%, with an AUC of 0.890. The cutoff values for VBQ and EBQ ranged from 2.68–3.445 and 1.973–4.7, respectively. Conclusions: MRI-based VBQ and EBQ scores demonstrate strong diagnostic accuracy for predicting cage subsidence in anterior cervical spine surgery. These scores could aid in preoperative risk assessment, though further research should aim to standardize MRI protocols and validate cutoff values across diverse populations.

Publikationsverlauf

Eingereicht: 13. April 2025

Angenommen nach Revision: 20. Januar 2026

Accepted Manuscript online:
21. Januar 2026