Assessing lumbar vertebral bone quality: a methodological evaluation of CT and MRI as alternatives to traditional DEXA.

PURPOSE: The purpose of this study was to investigate the impact of various methods on the assessment of vertebral bone quality. METHODS: A consecutive series of 427 candidates for lumbar disc replacement with lumbar DEXA and MRI and/or CT scans were included. Two measurement techniques were used on CTs-a sagittal and axial. From axial images, the upper, mid, and lower portions of each vertebral body were measured. Four MRI vertebral bone quality (VBQ) calculations were generated using separate equations. RESULTS: All CT measures were highly correlated with each other, regardless of measurement or calculation method (range 0.925-0.995). Sagittal measurements were highly correlated with axial (r = 0.928, p < 0.001). CT values were correlated with DEXA (range 0.446-0.534). There was no benefit to measuring multiple axial images of each vertebral body vs. just midbody (r = 0.441 and 0.455, respectively). No MRI VBQ values were highly correlated with DEXA (r = - 0.103, p = 0.045). In receiver operating curve analysis, the area under the curve ranged from 0.539 to 0.558, indicating poor ability of VBQ to identify osteoporosis/osteopenia. CONCLUSION: CT produced values more closely related to DEXA, while MRI was less reliable for osteoporosis/osteopenia screening. On CT, there was no benefit to making multiple measurements for each vertebral body to calculate a composite. Measuring sagittal CT images produced values similar to axial and required less time. While assessing bone quality from existing images rather than getting an additional DEXA scan is appealing, the methods of measuring these images needs standardization to maximize their utility.