Moore C S, Wood T J, Saunderson J R, Beavis A W
Radiation Physics Department, Queen's Centre for Oncology and Haematology, Castle Hill Hospital, Hull & East Yorkshire Hospitals NHS Trust, Castle Road, Hull, HU16 5JQ, UK. Faculty of Science and Engineering, University of Hull, Cottingham Road, Hull, HU6 7RX, UK.
Phys Med Biol. 2015 Dec 7;60(23):9047-58. doi: 10.1088/0031-9155/60/23/9047. Epub 2015 Nov 5.
This work assessed the appropriateness of the signal-to-noise ratio improvement factor (KSNR) as a metric for the optimisation of computed radiography (CR) of the chest. The results of a previous study in which four experienced image evaluators graded computer simulated chest images using a visual grading analysis scoring (VGAS) scheme to quantify the benefit of using an anti-scatter grid were used for the clinical image quality measurement (number of simulated patients = 80). The KSNR was used to calculate the improvement in physical image quality measured in a physical chest phantom. KSNR correlation with VGAS was assessed as a function of chest region (lung, spine and diaphragm/retrodiaphragm), and as a function of x-ray tube voltage in a given chest region. The correlation of the latter was determined by the Pearson correlation coefficient. VGAS and KSNR image quality metrics demonstrated no correlation in the lung region but did show correlation in the spine and diaphragm/retrodiaphragmatic regions. However, there was no correlation as a function of tube voltage in any region; a Pearson correlation coefficient (R) of -0.93 (p = 0.015) was found for lung, a coefficient (R) of -0.95 (p = 0.46) was found for spine, and a coefficient (R) of -0.85 (p = 0.015) was found for diaphragm. All demonstrate strong negative correlations indicating conflicting results, i.e. KSNR increases with tube voltage but VGAS decreases. Medical physicists should use the KSNR metric with caution when assessing any potential improvement in clinical chest image quality when introducing an anti-scatter grid for CR imaging, especially in the lung region. This metric may also be a limited descriptor of clinical chest image quality as a function of tube voltage when a grid is used routinely.
本研究评估了信噪比改善因子（KSNR）作为胸部计算机X线摄影（CR）优化指标的适用性。此前一项研究的结果被用于临床图像质量测量（模拟患者数量 = 80），该研究中，四名经验丰富的图像评估人员使用视觉分级分析评分（VGAS）方案对计算机模拟胸部图像进行分级，以量化使用防散射格栅的益处。KSNR用于计算在物理胸部模型中测量的物理图像质量的改善情况。评估了KSNR与VGAS的相关性，该相关性是胸部区域（肺、脊柱和膈肌/膈后区域）的函数，也是给定胸部区域内X射线管电压的函数。后者的相关性由Pearson相关系数确定。VGAS和KSNR图像质量指标在肺区域无相关性，但在脊柱和膈肌/膈后区域显示出相关性。然而，在任何区域，相关性都不是管电压的函数；肺区域的Pearson相关系数（R）为 -0.93（p = 0.015），脊柱区域的系数（R）为 -0.95（p = 0.46），膈肌区域的系数（R）为 -0.85（p = 0.015）。所有结果均显示出强负相关，表明结果相互矛盾，即KSNR随管电压增加而VGAS随管电压降低。在为CR成像引入防散射格栅时，医学物理学家在评估临床胸部图像质量的任何潜在改善时应谨慎使用KSNR指标，尤其是在肺区域。当常规使用格栅时，该指标作为管电压函数对临床胸部图像质量的描述可能也有限。
