Combinations of two imaging parameters to improve bone mineral density (BMD) assessment in patients with lumbar degenerative diseases

Abstract

Abstract Purpose To explore whether combining the Hounsfield unit (HU) values and vertebral bone quality (VBQ) scores can improve the BMD assessment in patients with lumbar degenerative diseases. Methods The HU values were measured by CT image, and VBQ scores were calculated by lumbar MRI image. The correlations of the opportunistic imaging parameters to the lowest T-scores were analyzed. Receiver-operating characteristic curve (ROC) analysis was used to evaluate the accuracy in detecting osteoporosis. Finally, the specificity and sensitivity of different combined methods of the HU values and VBQ scores in the diagnosis of osteoporosis were compared. Results Patients with osteoporosis had the lowest HU values and the highest VBQ scores. The correlation coefficients between the VBQ scores and the T-scores were smaller than HU values (L1 HU value: 0.702; average HU value:0.700; L1 VBQ score: -0.413; VBQ score: -0.386). The areas under the curve (AUCs) of the HU values were greater than those of the VBQ scores, and the AUCs of the L1 VBQ score were similar to the VBQ score (L1 HU value: 0.850; average HU value:0.857; L1 VBQ score: 0.704; VBQ score: 0.673). When combining the two imaging parameters in series, the specificity of the detection of osteoporosis was improved (L1 HU value and L1 VBQ score: 87.3%; Average HU value and VBQ score: 85.9%). When combining the two imaging parameters in parallel, the sensitivity of the detection of osteoporosis was improved (L1 HU value or L1 VBQ score: 88.1%; Average HU value or VBQ score: 91.5%). Conclusions Combinations of the HU values and VBQ scores could improve the diagnostic performance of osteoporosis. In addition, considering the same diagnostic performance but easier measurement, parameters at the single-segment level were recommended to assist in the diagnosis of osteoporosis.