MRI-Based Score for Assessment of Bone Mineral Density in Operative Spine Patients

Spine (Phila Pa 1976). 2023 Jan 15;48(2):107-112. doi: 10.1097/BRS.0000000000004509. Epub 2022 Oct 17.

Abstract

Study design: Retrospective comparison.

Objective: The aim was to determine whether a previously developed magnetic resonance imaging (MRI)-derived bone mineral density (BMD) scoring system can differentiate between healthy and osteoporotic vertebrae and to validate this scoring system against quantitative computed tomography measurements.

Summary of background data: BMD is an important preoperative consideration in spine surgery. Techniques to measure BMD are subject to falsely elevated values in the setting of spondylosis (dual-energy X-ray absorptiometry) or require significant exposure to radiation [quantitative computed tomography (QCT)]. Previous studies have shown that MRI may be utilized to measure bone quality using changes in the bone marrow signal observed on T1-weighted MRIs.

Materials and methods: Retrospective study of patients who underwent operative lumbar procedures at a single tertiary institution between 2016 and 2021 (n=61). Vertebral bone quality (VBQ) scores were measured by dividing the median signal intensities of L1-L4 by the signal intensity of cerebrospinal fluid on noncontrast T1W MRI. Demographic data, comorbidities, VBQ scores, and QCT-derived T scores and BMD of the lumbar spine were compared between healthy ( T score ≥-1; n=21), osteopenic (-2.5 < T score < -1; n=21), and osteoporotic ( T score ≤-2.5; n=19) cohorts using analysis of variance with post hoc Tukey test. Linear regression and receiver operating characteristic curve analyses were performed to assess the predictive value of VBQ scores. Pearson correlation test was used to evaluate the association between VBQ scores and QCT-derived measurements.

Results: VBQ differentiated between healthy and osteoporotic groups ( P =0.009). Receiver operating characteristic curve analysis revealed that a greater VBQ score was associated with presence of osteoporosis (area under the curve=0.754, P =0.006). Cutoff VBQ for osteoporosis was 2.6 (Youden index 0.484; sensitivity: 58%; specificity: 90%). VBQ scores weakly correlated with QCT-derived BMD ( P =0.03, r =-0.27) and T scores ( P =0.04, r =-0.26).

Conclusion: This study attempted to further validate a previously developed MRI-based BMD scoring system against QCT-derived measurements. VBQ score was found to be a significant predictor of osteoporosis and could differentiate between healthy and osteoporotic vertebrae.

MeSH terms

  • Absorptiometry, Photon / methods
  • Bone Density*
  • Humans
  • Lumbar Vertebrae / diagnostic imaging
  • Lumbar Vertebrae / surgery
  • Magnetic Resonance Imaging
  • Osteoporosis* / diagnostic imaging
  • Retrospective Studies