High-Performance Automated System For Analysis of Cardiac SPECT

描述（由申请人提供）：冠状动脉疾病（CAD）仍然是一个主要的公共卫生问题。它是美国男性和女性死亡的最大单一原因，占所有死亡人数的20%。虽然对冠心病有有效的医学和侵入性治疗，但它们的适当使用取决于对疾病的准确检测和对个体患者心脏风险的评估。门控心肌灌注SPECT （MRS）在这一过程中发挥了关键作用，提供了心肌灌注和心室功能的关键信息。2005年，美国有超过800万患者接受了核磁共振治疗。目前，MRS解释的标准方法是在应激和休息状态下对局部心肌灌注摄取的主观视觉评分。这种视觉方法耗时，受观察者之间的可变性的影响，并且在异常检测和估计其大小方面可能是次优的。我们的目标是开发一种全自动的MRS计算机系统，它将在诊断CAD和预测心脏事件方面超越经验丰富的人类读者的表现。这种高水平的性能将通过应用新的图像处理技术，提高图像质量，以及所有可用图像数据的自动区域集成来实现。具体而言，我们的目标是：1)开发灌注量化增强技术，2)开发衰减校正MRS量化新技术，3)通过与大型多中心研究中多位专家的视觉评估进行比较，验证最终集成系统的诊断性能，并通过对大型结果数据库的回顾性分析进行预后验证。新系统将有能力从成像伪影中区分真正的异常，并将检测细微的缺陷。我们假设，新系统将能够检测CAD和预测结果，如心脏死亡比最好的视觉分析。这种发展将产生深远和直接的影响，因为这种新的准确性和自动化水平的MRS可以在国内和国际上广泛复制。这项工作将提高MRS检测的效率，并由于CAD的更准确诊断和更好地选择适当的治疗而节省大量的成本。在休息和应激状态下的心肌灌注（心肌血流）成像使医生能够检测疾病并预测美国每年数百万患者的风险，但目前受限于视觉解释的需要，这取决于医生的经验。研究人员建议开发并验证一种自动化、高度精确和客观的计算机系统，该系统在解释这些图像方面甚至比经验丰富的医生表现更好，从而通过更好地选择需要治疗的患者来挽救更多的生命，同时也节省了时间和成本。

DESCRIPTION (provided by applicant): Coronary artery disease (CAD) continues to be a major public health problem. It is the single greatest cause of death for men and women in the US, accounting for 20% of all deaths. While there are effective medical and invasive therapies for CAD, their appropriate use is dependent on accurate detection of the disease and evaluation of cardiac risk in individual patients. Gated myocardial perfusion SPECT (MRS) has played a critical role in this process, providing key information about myocardial perfusion and ventricular function. Over 8 million patients underwent MRS in the US in 2005. Currently, the standard method for MRS interpretation is subjective visual scoring of regional myocardial uptake of perfusion at stress and rest. This visual approach is time-consuming, suffers from inter-observer variability, and is potentially sub-optimal in the detection of abnormalities and estimation of their magnitude. We aim to develop a fully automated computer system for MRS that will surpass the performance of experienced human readers in diagnosing CAD and in predicting cardiac events. This high level of performance will be accomplished by the application of new image processing techniques, improvement of image quality, and automatic regional integration of all available image data. Specifically, we aim to: 1) develop enhanced techniques for perfusion quantification, 2) develop new techniques for quantification of attenuation corrected MRS, and 3) validate performance of the final integrated system diagnostically by comparison to the visual evaluation by multiple experts in a large multi-center study and prognostically by retrospective analysis of a large outcome database. The new system will have the ability to distinguish true abnormalities from imaging artifacts and will detect subtle defects. We hypothesize that the new system will be able to detect CAD and predict outcomes such as cardiac death better than the best attainable visual analysis. Such development will have far-reaching and immediate consequences since this new level of accuracy and automation for MRS can be widely reproduced nationally and internationally. This work will result in increased efficiency of MRS testing and large cost savings due to more accurate diagnosis of CAD and better selection of appropriate treatment. Imaging of myocardial perfusion (heart muscle blood flow) at rest and stress allows physicians to detect disease and predict risk in millions of patients in the US each year, but it is currently limited by the need of visual interpretation, which is dependent on doctor's experience. The investigators propose to develop and validate an automated, highly-accurate and objective computer system which will outperform even experienced physicians in interpreting these images and consequently allow a greater number of lives saved by better selection of patients needing treatment and also resulting in time- and cost-savings.