Department of Biomedical Engineering, Case Western Reserve University, Cleveland, OH, 44106, USA.
Department of Radiology, Case Western Reserve University, 10900 Euclid Avenue, Cleveland, OH, 44106, USA.
Int J Cardiovasc Imaging. 2021 Mar;37(3):767-774. doi: 10.1007/s10554-020-02072-4. Epub 2020 Oct 28.
Our goal is to assess the ability of physicians to detect coronary calcifications in dual energy chest X-rays processed by a previously developed advanced algorithm. Because the chest X-ray is the most common imaging procedure, because the presence of coronary calcium provides proof of coronary artery disease, and because adherence to therapy can improve health, successful detection could positively impact healthcare for a large number of patients. Both dual energy chest and corroborative CT calcium score images were acquired. Dual energy images were processed with the advanced techniques, including sliding organ registration, so as to enhance coronary calcifications in two-shot dual energy acquisitions. We performed ROC to determine physicians' ability to detect coronary calcifications. Since detection might be easier with heavier calcifications, we used various Agatston score cut-points for determining cases actually positive with calcification in the ROC analysis. In many cases, coronary calcifications were made more visible with the advanced processing as compared to conventional processing. At an Agatston cut-point of 300, coronary calcifications were detected with AUC = 0.85. There were marginal effects on detection performance found with increased X-ray exposure, nearby Agatston cut-point values, and coronary artery territory. Coronary calcifications can be detected in dual energy chest X-rays. The ability to detect disease compares very favorably to other accepted screening methods (e.g., X-ray mammography). As the chest X-ray is an already ordered procedure, there is an opportunity to detect a very large number of persons with coronary artery disease at zero or low cost.
我们的目标是评估医生在经过先前开发的先进算法处理的双能胸部 X 光片中检测冠状动脉钙化的能力。由于胸部 X 光是最常见的成像程序，由于冠状动脉钙的存在提供了冠状动脉疾病的证据，并且由于坚持治疗可以改善健康状况，因此成功的检测可以对大量患者的医疗保健产生积极影响。我们同时获取了双能胸部和补充 CT 钙评分图像。双能图像使用先进的技术进行处理，包括滑动器官配准，以便在双能采集的两次拍摄中增强冠状动脉钙化。我们进行了 ROC 分析，以确定医生检测冠状动脉钙化的能力。由于较重的钙化可能更容易检测，因此我们在 ROC 分析中使用了各种 Agatston 评分切点来确定实际上存在钙化的病例。在许多情况下，与常规处理相比，先进的处理可以使冠状动脉钙化更明显。在 Agatston 切点为 300 时，冠状动脉钙化的 AUC 为 0.85。随着 X 射线曝光、附近的 Agatston 切点值和冠状动脉区域的增加，检测性能有了一定的改善。在双能胸部 X 光片中可以检测到冠状动脉钙化。与其他公认的筛查方法（例如 X 射线乳房 X 光检查）相比，检测疾病的能力非常出色。由于胸部 X 光是已经预定的程序，因此有机会以零成本或低成本检测到大量患有冠状动脉疾病的人。
