Ono Takaaki, Ichikawa Hajime, Kato Taiki
Department of Radiology, Toyohashi Municipal Hospital.
Nihon Hoshasen Gijutsu Gakkai Zasshi. 2022 Aug 20;78(8):838-845. doi: 10.6009/jjrt.2022-1237. Epub 2022 Jul 6.
The purpose of this study was to evaluate the detectability of the observers and motion detection software for blurred portable chest radiographies.
The chest phantom radiographies of various blur sizes were obtained by moving the phantom using 4° slope. The phantom was moved in two directions, vertical (the upper and lower parts of the chest phantom were parallel to rails) and horizontal (the left and right parts of the chest phantom were parallel to rails). Six observers performed receiver operating characteristic (ROC) studies on blurred images. We used the results to compare detectability for vertical and horizontal blur by ROC analysis, and calculated sensitivity and specificity. In addition, the motion detection software was enabled during image acquisition, and the detectability was compared with that of the observers.
The average of the area under the ROC curve for the detection of blur in the vertical and horizontal directions for the observers were 0.918 and 0.943, respectively, and no significant difference was found depending on the direction. The motion detection software performed better than the observers in most of the sensitivity and specificity scores.
The motion detection software could be a useful support tool for motion detection in portable chest radiography.
本研究旨在评估观察者及运动检测软件对模糊的便携式胸部X光片的检测能力。
通过以4°倾斜度移动胸部体模来获取不同模糊程度的胸部体模X光片。体模在两个方向上移动，垂直方向（胸部体模的上下部分与轨道平行）和水平方向（胸部体模的左右部分与轨道平行）。六名观察者对模糊图像进行了接受者操作特征（ROC）研究。我们利用结果通过ROC分析比较垂直和水平模糊的检测能力，并计算敏感性和特异性。此外，在图像采集期间启用运动检测软件，并将其检测能力与观察者的进行比较。
观察者检测垂直和水平方向模糊的ROC曲线下面积平均值分别为0.918和0.943，未发现因方向不同而存在显著差异。在大多数敏感性和特异性评分中，运动检测软件的表现优于观察者。
运动检测软件可能是便携式胸部X光摄影中运动检测的有用支持工具。
