Graffy Peter M, Lee Scott J, Ziemlewicz Timothy J, Pickhardt Perry J
1 Department of Radiology, University of Wisconsin School of Medicine & Public Health, E3/311 Clinical Science Center, 600 Highland Ave, Madison, WI 53792-3252.
AJR Am J Roentgenol. 2017 Sep;209(3):491-496. doi: 10.2214/AJR.17.17853. Epub 2017 Jun 22.
Radiologists interpreting body CT scans may be the first to identify osteoporosis and associated vertebral fractures. This study correlates L1 trabecular attenuation measurements with prevalent vertebral body fractures in older adults undergoing routine CT.
Mean L1 trabecular attenuation was measured at thoracoabdominal CT in 1966 consecutive adults (983 men and 983 women) 65 years and older (mean age, 74.1 ± 6.6 [SD] years). Sagittal reconstructions and lateral scouts were analyzed for moderate or severe thoracolumbar vertebral compression fractures according to the Genant semiquantitative assessment method. The diagnostic performance of L1 attenuation for the evaluation of prevalent vertebral fractures was assessed, including ROC curve analysis.
A total of 162 (8.2%) individuals (mean age, 78.3 years; 66 men, 96 women) had at least one moderate or severe vertebral fracture. The mean L1 attenuation was 70.2 HU among patients with a prevalent fracture, whereas it was 132.3 HU among patients without fracture (p < 0.001). The prevalence of moderate or severe vertebral compression fractures was 32.5% when L1 attenuation was ≤ 90 HU. Prevalence increased to 49.2% with L1 attenuation of ≤ 50 HU. ROC curve analysis determined an optimal threshold of 90 HU (sensitivity = 86.9%, specificity = 83.9%), with a corresponding AUC of 0.895. The odds ratio of having a moderate or severe vertebral compression fracture was 31.9 for L1 attenuation ≤ 90 HU (95% CI, 20.2-50.5; p < 0.001).
Patients with moderate or severe vertebral compression fractures have significantly lower L1 attenuation values than patients who do not. L1 attenuation ≤ 90 HU may represent an optimal threshold for determining risk for osteoporotic vertebral fractures.
解读身体CT扫描的放射科医生可能是最先识别骨质疏松症及相关椎体骨折的人。本研究将L1小梁衰减测量值与接受常规CT检查的老年人中普遍存在的椎体骨折进行关联。
在1966名65岁及以上(平均年龄74.1±6.6[标准差]岁)的连续成人(983名男性和983名女性)的胸腹CT上测量L1小梁平均衰减。根据Genant半定量评估方法,对矢状面重建图像和侧位定位像进行分析,以确定是否存在中度或重度胸腰椎椎体压缩骨折。评估L1衰减对评估普遍存在的椎体骨折的诊断性能,包括ROC曲线分析。
共有162名(8.2%)个体(平均年龄78.3岁;66名男性,96名女性)至少有一处中度或重度椎体骨折。有普遍骨折的患者L1平均衰减为70.2HU,而无骨折患者为132.3HU(p<0.001)。当L1衰减≤90HU时,中度或重度椎体压缩骨折的患病率为32.5%。当L1衰减≤50HU时,患病率增至49.2%。ROC曲线分析确定最佳阈值为90HU(敏感性=86.9%,特异性=83.9%),相应的AUC为0.895。L1衰减≤90HU时发生中度或重度椎体压缩骨折的比值比为31.9(95%CI,20.2-50.5;p<0.001)。
中度或重度椎体压缩骨折患者的L1衰减值显著低于无骨折患者。L1衰减≤90HU可能是确定骨质疏松性椎体骨折风险的最佳阈值。