Michalak Gregory, Grimes Joshua, Fletcher Joel, Halaweish Ahmed, Yu Lifeng, Leng Shuai, McCollough Cynthia
Department of Radiology, Mayo Clinic, Rochester, Minnesota 55905.
Siemens Medical Solutions, Malvern, Pennsylvania 19355.
Med Phys. 2016 Jan;43(1):513. doi: 10.1118/1.4939128.
The purpose of this study was to evaluate, over a wide range of phantom sizes, CT number stability achieved using two techniques for generating dual-energy computed tomography (DECT) virtual monoenergetic images.
Water phantoms ranging in lateral diameter from 15 to 50 cm and containing a CT number test object were scanned on a DSCT scanner using both single-energy (SE) and dual-energy (DE) techniques. The SE tube potentials were 70, 80, 90, 100, 110, 120, 130, 140, and 150 kV; the DE tube potential pairs were 80/140, 70/150Sn, 80/150Sn, 90/150Sn, and 100/150Sn kV (Sn denotes that the 150 kV beam was filtered with a 0.6 mm tin filter). Virtual monoenergetic images at energies ranging from 40 to 140 keV were produced from the DECT data using two algorithms, monoenergetic (mono) and monoenergetic plus (mono+). Particularly in large phantoms, water CT number errors and/or artifacts were observed; thus, datasets with water CT numbers outside ±10 HU or with noticeable artifacts were excluded from the study. CT numbers were measured to determine CT number stability across all phantom sizes.
Data exclusions were generally limited to cases when a SE or DE technique with a tube potential of less than 90 kV was used to scan a phantom larger than 30 cm. The 90/150Sn DE technique provided the most accurate water background over the large range of phantom sizes evaluated. Mono and mono+ provided equally improved CT number stability as a function of phantom size compared to SE; the average deviation in CT number was only 1.4% using 40 keV and 1.8% using 70 keV, while SE had an average deviation of 11.8%.
The authors' report demonstrates, across all phantom sizes, the improvement in CT number stability achieved with mono and mono+ relative to SE.
本研究的目的是在广泛的体模尺寸范围内，评估使用两种生成双能计算机断层扫描（DECT）虚拟单能图像的技术所实现的CT值稳定性。
使用单能（SE）和双能（DE）技术在DSCT扫描仪上扫描横向直径为15至50 cm且包含CT值测试物体的水模。SE管电压为70、80、90、100、110、120、130、140和150 kV；DE管电压对为80/140、70/150Sn、80/150Sn、90/150Sn和100/150Sn kV（Sn表示150 kV束流用0.6 mm锡滤波器滤波）。使用单能（mono）和单能加（mono+）两种算法从DECT数据生成能量范围为40至140 keV的虚拟单能图像。特别是在大型体模中，观察到了水CT值误差和/或伪影；因此，水CT值超出±10 HU或有明显伪影的数据集被排除在研究之外。测量CT值以确定所有体模尺寸下的CT值稳定性。
数据排除通常仅限于使用管电压小于90 kV的SE或DE技术扫描大于30 cm的体模的情况。在评估的大范围体模尺寸中，90/150Sn DE技术提供了最准确的水背景。与SE相比，Mono和Mono+作为体模尺寸的函数，同样提高了CT值稳定性；使用40 keV时CT值的平均偏差仅为1.4%，使用70 keV时为1.8%，而SE的平均偏差为11.8%。
作者的报告表明，在所有体模尺寸下，相对于SE，Mono和Mono+在CT值稳定性方面都有改善。
