O.N. Diagnostics LLC, Berkeley, CA, USA.
Department of Mechanical Engineering, University of California, Berkeley, CA, USA; Department of Bioengineering, University of California, Berkeley, CA, USA.
Bone. 2017 Oct;103:325-333. doi: 10.1016/j.bone.2017.07.029. Epub 2017 Aug 1.
Patient-specific phantomless calibration of computed tomography (CT) scans has the potential to simplify and expand the use of pre-existing clinical CT for quantitative bone densitometry and bone strength analysis for diagnostic and monitoring purposes. In this study, we quantified the inter-operator reanalysis precision errors for a novel implementation of patient-specific phantomless calibration, using air and either aortic blood or hip adipose tissue as internal calibrating reference materials, and sought to confirm the equivalence between phantomless and (traditional) phantom-based measurements. CT scans of the spine and hip for 25 women and 15 men (mean±SD age of 67±9years, range 41-86years), one scan per anatomic site per patient, were analyzed independently by two analysts using the VirtuOst software (O.N. Diagnostics, Berkeley, CA). The scans were acquired at 120kVp, with a slice thickness/increment of 3mm or less, on nine different CT scanner models across 24 different scanners. The main parameters assessed were areal bone mineral density (BMD) at the hip (total hip and femoral neck), trabecular volumetric BMD at the spine, and vertebral and femoral strength by finite element analysis; other volumetric BMD measures were also assessed. We found that the reanalysis precision errors for all phantomless measurements were ≤0.5%, which was as good as for phantom calibration. Regression analysis indicated equivalence of the phantom- versus phantomless-calibrated measurements (slope not different than unity, R≥0.98). Of the main parameters assessed, non-significant paired mean differences (n=40) between the two measurements ranged from 0.6% for hip areal BMD to 1.1% for mid-vertebral trabecular BMD. These results indicate that phantom-equivalent measurements of both BMD and finite element-derived bone strength can be reliably obtained from CT scans using patient-specific phantomless calibration.
基于患者的 CT 扫描无模校准有可能简化并扩展现有临床 CT 在定量骨密度和骨强度分析方面的应用，以便于诊断和监测。在本研究中，我们使用空气以及主动脉血液或髋部脂肪组织作为内部校准参考材料，对一种新的基于患者的无模校准方法进行了操作者间重新分析精度误差的量化，并试图确认无模和（传统）基于模体的测量之间的等效性。对 25 名女性和 15 名男性（平均年龄 67±9 岁，范围 41-86 岁）的脊柱和髋部 CT 扫描进行了分析，每位患者每个解剖部位各有一次扫描，由两位分析员分别使用 VirtuOst 软件（O.N. Diagnostics，加利福尼亚州伯克利）进行分析。扫描时使用的管电压为 120kVp，层厚/层距小于或等于 3mm，扫描设备为 24 台不同型号的 9 种 CT 扫描仪。评估的主要参数包括髋部（全髋和股骨颈）的面积骨密度（BMD）、脊柱的小梁体积 BMD、通过有限元分析得出的椎体和股骨强度；还评估了其他体积 BMD 测量。我们发现，所有无模测量的重新分析精度误差均≤0.5%，与模体校准相当。回归分析表明，模体校准与无模校准的测量结果等效（斜率不等于 1，R≥0.98）。在所评估的主要参数中，两种测量方法之间（n=40）的配对均值差异无统计学意义，范围从髋部面积 BMD 的 0.6%到中胸段小梁 BMD 的 1.1%。这些结果表明，使用基于患者的无模校准可以从 CT 扫描中可靠地获得与模体等效的 BMD 和基于有限元分析的骨强度测量值。
