From Siemens Healthcare, An der Laende 1, 91301 Forchheim, Germany (T.N., B.S.); Institute of Diagnostic and Interventional Radiology (M.E., N.S., H.A., A.E.) and Department of Rheumatology (D.F., O.D.), University Hospital Zurich, University of Zurich, Raemistrasse 100, 8091 Zurich, Switzerland.
Radiology. 2021 Jan;298(1):147-152. doi: 10.1148/radiol.2020202767. Epub 2020 Nov 3.
Background Bone mineral density (BMD) could be derived from CT localizer radiographs and could potentially enable opportunistic osteoporosis screening. Purpose To assess the accuracy and precision of BMD measurement using two localizer radiographs obtained with energy-integrating detector CT and a single localizer radiograph obtained with photon-counting detector CT. Materials and Methods A calibration phantom and a porcine phantom with lumbar vertebrae were imaged with a dual-energy x-ray absorptiometry (DXA) scanner, a clinical energy-integrating detector CT scanner, and a prototype photon-counting detector CT scanner. Two localizer radiographs at different combinations of tube voltages were obtained with energy-integrating detector CT, and one localizer radiograph was obtained with photon-counting detector CT using different energy thresholds. BMD was calculated for all three approaches and compared with the known specifications in the calibration phantom. In the animal phantom, BMDs from both CT systems were compared with those from the DXA scanner (the reference standard). Accuracy was defined as the measurement error of BMD (ΔBMD), and precision was defined as the coefficient of variation (in percentage). Radiation doses were estimated. Nonparametric tests were applied. Results In the calibration phantom, ΔBMD was smaller with both CT systems compared with the DXA scanner (both < .05). ΔBMD ranged from -5% to -1.8% for DXA, from -2.3% to -1.7% for energy-integrating detector CT, and from -1.6% to 1.6% for photon-counting detector CT. Precision (range, 0.3%-2.8%) was high for both CT systems. In the animal phantom, ΔBMD ranged from -0.6% to 0.1% for energy-integrating detector CT and from -0.1% to 0.6% for photon-counting detector CT, with no significant differences between CT systems ( = .65). The dose-area product in the animal phantom was 4.6 cGy ∙ cm for DXA, 3.5-11.5 cGy ∙ cm for energy-integrating detector CT, and 7.2-11.2 cGy ∙ cm for photon-counting detector CT, depending on tube voltage and energy threshold combination. Conclusion Experimental evidence suggests that bone mineral density measurements are accurate and precise using two localizer radiographs at different tube voltages from energy-integrating detector CT and a single localizer radiograph with different energy thresholds from photon-counting detector CT. © RSNA, 2020 See also the editorial by Pourmorteza in this issue.
背景 骨密度（BMD）可从 CT 定位片获得，这可能使机会性骨质疏松症筛查成为可能。目的 评估使用两种能量积分探测器 CT 获得的两个定位片和一种使用光子计数探测器 CT 获得的单个定位片测量 BMD 的准确性和精密度。材料与方法 使用双能 X 射线吸收法（DXA）扫描仪、临床能量积分探测器 CT 扫描仪和原型光子计数探测器 CT 扫描仪对校准体模和带有腰椎的猪体模进行成像。使用能量积分探测器 CT 以不同的管电压组合获得两个定位片，使用不同的能阈获得一个光子计数探测器 CT 定位片。对所有三种方法进行 BMD 计算，并与校准体模中的已知规格进行比较。在动物体模中，比较两种 CT 系统与 DXA 扫描仪（参考标准）的 BMD。准确性定义为 BMD 的测量误差（ΔBMD），精密度定义为变异系数（以百分比表示）。估计了辐射剂量。应用了非参数检验。结果 在体模校准中，与 DXA 扫描仪相比，两种 CT 系统的 ΔBMD 均较小（均<0.05）。DXA 的 ΔBMD 范围为-5%至-1.8%，能量积分探测器 CT 的 ΔBMD 范围为-2.3%至-1.7%，光子计数探测器 CT 的 ΔBMD 范围为-1.6%至 1.6%。两种 CT 系统的精密度（范围为 0.3%-2.8%）均较高。在动物体模中，能量积分探测器 CT 的 ΔBMD 范围为-0.6%至 0.1%，光子计数探测器 CT 的 ΔBMD 范围为-0.1%至 0.6%，两种 CT 系统之间无显著差异（=0.65）。动物体模中的剂量面积产物为 DXA 的 4.6 cGy ∙ cm、能量积分探测器 CT 的 3.5-11.5 cGy ∙ cm 和光子计数探测器 CT 的 7.2-11.2 cGy ∙ cm，具体取决于管电压和能阈组合。结论 实验证据表明，使用能量积分探测器 CT 以不同的管电压获得两个定位片和使用光子计数探测器 CT 以不同的能阈获得单个定位片，可以准确且精确地测量骨密度。
