Comparison of L1 CT-attenuation and cortical thickness in predicting osteoporosis by opportunistic CT

Abstract

BACKGROUND: In vertebrae, the amount of cortical bone has been estimated at 30–60%, but 45–75% of axial load on a vertebral body is borne by cortical bone. OBJECTIVE: To compare the role of L1 CT-attenuation and cortical thickness in predicting osteoporosis by opportunistic CT and explore cortical thickness value in osteoporosis. METHODS: We collected data of 94 patients who underwent DXA and thoracic and/or abdominal CT to demonstrate an entire L1 for other indications in routine practice. Patients were divided into three groups according to T-score: osteoporosis, osteopenia, or normal. CT-attenuation value and cortical thickness of L1 were measured. ANOVA analysis was utilized to analyze CT-attenuation and cortical thickness among the three groups. Sensitivity, specificity, and area under the curve (AUC) predicting low BMD were determined using ROC. Pearson correlations were employed to describe relationship between L1 BMD and CT-attenuation value, BMD, as well as cortical thickness. RESULTS: The mean cortical thickness was 0.83±0.11, 0.72±0.10, and 0.64±0.09 mm for normal, osteopenia, and osteoporotic subgroups, respectively. A statistically significant difference was observed in cortical thickness and CT-attenuation value among these three subgroups. A mean CT-attenuation value threshold of > 148.7 yielded 73.0% sensitivity and 86.0% specificity for distinguishing low BMD from normal with an AUC = 0.83. Pearson correlation analysis indicated that BMD was positively correlated with CT-attenuation (r = 0.666, P < 0.001) and cortical thickness (r = 0.604, P < 0.001). CONCLUSIONS: L1 CT-attenuation and cortical thickness measured on opportunistic CT can help predict osteoporosis. Compared with cortical thickness, CT-attenuation is a more sensitive and accurate index for distinguishing low BMD from normal.