Accuracy of cancellous bone volume fraction measured by micro-CT scanning
Accuracy of cancellous bone volume fraction measured by micro-CT scanning
复制标题
DOI：
10.1016/s0021-9290(98)00176-6
复制
发表时间：
1999-03-01
影响因子：
2.4
通讯作者：
Hvid, I
中科院分区：
文献类型：
--
作者：
Ding, M;Odgaard, A;Hvid, I
体积分数是描述骨小梁微观结构的最重要的参数，可以很容易地从显微CT图像的三维重建中计算出来。这项研究试图量化这种测量的准确性。制备了160个人体松质骨标本，覆盖了大范围的体积分数（9.8-39.8%）。对试样进行显微CT扫描，并根据阿基米德原理确定体积分数作为参考。扫描后，使用扫描仪提供的算法（方法I）确定的各个阈值对所有micro-CT数据进行分割。发现体积分数与方法I存在显著偏差;回归线的y轴截距和斜率均与基于阿基米德的体积分数显著不同（p < 0.001）。新的个体阈值是基于体积分数到基于阿基米德的体积分数的校准（方法II）来确定的。将两种方法的平均阈值应用于20个随机选择的标本的分割。结果表明，使用方法I的平均阈值的体积分数被低估了4%（p = 0.001），而方法II的平均阈值产生准确的值。测量的精确度非常好。我们的数据表明，必须小心时，在生成3-D数据应用阈值，并可以使用一个固定的阈值，以获得可靠的体积分数数据。该固定阈值可以从样本子组的基于阿基米德的体积分数确定。不同材料之间的阈值可能不同，因此应在进行研究系列时确定。(C)1999 Elsevier Science Ltd.保留所有权利。
Volume fraction, the single most important parameter in describing trabecular microstructure, can easily be calculated from three-dimensional reconstructions of micro-CT images. This study sought to quantify the accuracy of this measurement. One hundred and sixty human cancellous bone specimens which covered a large range of volume fraction (9.8-39.8%) were produced. The specimens were micro-CT scanned, and the volume fraction based on Archimedes' principle was determined as a reference. After scanning, all micro-CT data were segmented using individual thresholds determined by the scanner supplied algorithm (method I). A significant deviation of volume fraction from method I was found; both the y-intercept and the slope of the regression line were significantly different from those of the Archimedes-based volume fraction (p < 0.001). New individual thresholds were determined based on a calibration of volume fraction to the Archimedes-based volume fractions (method II). The mean thresholds of the two methods were applied to segment 20 randomly selected specimens. The results showed that volume fraction using the mean threshold of method I was underestimated by 4% (p = 0.001), whereas the mean threshold of method II yielded accurate values. The precision of the measurement was excellent. Our data show that care must be taken when applying thresholds in generating 3-D data, and that a fixed threshold may be used to obtain reliable volume fraction data. This fixed threshold may be determined from the Archimedes-based volume fraction of a subgroup of specimens. The threshold may vary between different materials, and so it should be determined whenever a study series is performed. (C) 1999 Elsevier Science Ltd. Ail rights reserved.