Yang Ding-zhuo, Shang Jia-yun, Song Wen-zhong, Chen Jian, Zhu Dan, Wang Wen-zhi
Department of Radiology, Huaxi Fourth Hospital, Sichuan University, Chengdu 610041, China.
Zhongguo Yi Xue Ke Xue Yuan Xue Bao. 2003 Jun;25(3):262-6.
To design and manufacture a reliable spine phantom used in the cross calibration and quality control of dual X-ray absorptiometry (DXA).
A hydroxyapatite quality control phantom was designed and made through three steps: solid water, bone phantom material and integration, then evaluated the phantom on four different types of DXA machines made by LUNAR company.
Among the four DXA densitometers, the Expert fan beam densitometer had the biggest accuracy errors of bone mineral density (BMD), bone mineral content(BMC) and area values while the other three one narrow fan beam (Prodigy) and two pencil beam densitometers-had small errors. Of the three indexes measured by all the machines, BMD error was the smallest (-15.4%-11.5%), with the Prodigy's BMD was most outstanding. BMD errors at the higher density ends were small, tend to be positive values while the errors at the lower ends were big, tend to be negative. In cross calibration, giving consideration to the differences between the both ends, it is better to use the regression equation to correct. The base line of precision error derived by scanning the phantom once a day for consecutive 25 days was better than that derived by scanning the phantom 25 times consecutively on the same day. As to precision error, the coefficient of variation (CV) of scanning-25 times-a day was the smallest (0.0043) while the CV of 12 adults measurements was the biggest (0.0078).
This phantom can be used in the quality control and cross-calibration of different types of DXA machines.
设计并制造一种用于双能X线吸收法（DXA）交叉校准和质量控制的可靠脊柱体模。
通过固体水、骨体模材料及整合三个步骤设计并制作羟基磷灰石质量控制体模，然后在LUNAR公司生产的四种不同类型的DXA机器上对该体模进行评估。
在这四台DXA骨密度仪中，Expert扇形束骨密度仪的骨矿物质密度（BMD）、骨矿物质含量（BMC）和面积值的准确性误差最大，而其他三台——一台窄扇形束（Prodigy）和两台笔形束骨密度仪——误差较小。在所有机器测量的三个指标中，BMD误差最小（-15.4%-11.5%），其中Prodigy的BMD最为突出。较高密度端的BMD误差较小，趋于正值，而较低端的误差较大，趋于负值。在交叉校准中，考虑到两端的差异，最好使用回归方程进行校正。通过连续25天每天扫描一次体模得出的精度误差基线优于同一天连续扫描体模25次得出的基线。至于精度误差，一天扫描25次的变异系数（CV）最小（0.0043），而12名成年人测量的CV最大（0.0078）。
该体模可用于不同类型DXA机器的质量控制和交叉校准。
