Patient Motion Detection and Compensation in SPECT

描述（由申请人提供）：患者运动是可能限制诊断成像准确性的伪影的始终存在的潜在原因。该问题对于诸如SPECT和PET之类的成像模态以及其中模态之间的患者对准是必要的多模态成像尤其重要，其中SPECT和PET要求患者在延长的时间段内保持不动。我们的假设是，整合到迭代重建相结合的信息，从视觉跟踪系统（VTS）和成像方式本身将导致一个强大的运动校正策略。VTS是指处理患者表面上的回射标记的立体图像以提供独立于临床成像系统的运动信息源的计算系统。将针对使用这种组合方法的心脏灌注SPECT成像研究补偿的患者运动类型是刚体运动（RBM）、非刚体运动（非RBM）、呼吸运动（RM）、心脏向上蠕动（由RM变化引起）和顺序多模态成像研究之间的运动。注意，我们以不同的方式校正身体运动和RM，因此我们将它们区分为运动估计的一部分。在这一竞争性更新中，我们建议扩展我们目前的VTS，以包括多模态成像、发射数据一致性和基于有限元建模（FEM）的非RBM校正的应用。志愿者的MRI将用于建立外部标记运动与体内心脏运动之间的关系。MRI还将提供不同运动状态下胸部的高分辨率解剖结构，并结合测量的标记运动，以指导创建逼真的SPECT模拟，从而改进算法。初始临床试验将在患者-志愿者中使用重复静息心脏SPECT研究，其中第一次静息成像研究作为判断第二次研究（患者有意移动）校正成功的标准。我们的假设的成功的最终评估将是医生-观察者ROC研究比较冠状动脉疾病（CAD）的检测准确性与运动补偿的患者接受SPECT灌注成像与心导管检查的结果作为“金”标准。我们认为，我们提出的调查的意义不仅在于通过运动校正的心脏和非心脏SPECT，PET和多模态成像，提高诊断准确性的潜力，但也在提供一个更好的理解病人的运动成像。 公共卫生相关性：患者通常在SPECT和PET所需的延长的成像期间移动，以收集统计上足够数量的事件来形成用于做出准确诊断的图像。这样的运动可能会导致切片中的人为变化，这可能会误导医生。在多模态成像中，患者还可以在发射成像和用于形成衰减图的成像之间移动，并为发射成像中的功能变化提供解剖框架。这可能会导致两组图像之间的配准不良，从而严重影响临床判读。而且，在患者呼吸时采集的所有图像都遭受被成像的结构的呼吸运动模糊。该项目的目标是为所有形式的患者运动开发一种准确和强大的运动校正方法，该方法最小限度地改变当前的采集协议，以便可以在忙碌的诊所中使用。我们选择用于开发该方法的试验平台是心脏灌注SPECT成像，其不仅是当今最常见的SPECT成像程序，而且在提供冠状动脉疾病风险评估和分层所需的灌注和功能信息方面发挥关键作用。一旦该方法被完善用于心脏SPECT，它可以被调整用于其他SPECT和PET成像程序。拟议的研究还将更好地了解医学成像和治疗过程中的患者运动。

DESCRIPTION (provided by applicant): Patient motion is an ever-present potential cause of artifacts that can limit the accuracy of diagnostic imaging. The problem is especially significant for imaging modalities such as SPECT and PET, which require the patient to remain motionless for protracted periods of time, and multimodality imaging where patient alignment between the modalities is essential. Our hypothesis is that integrating into iterative reconstruction combined information from a visual-tracking-system (VTS) and the imaging modalities themselves will result in a robust motion-correction strategy. By VTS we mean a computational system that processes stereo-images of retro-reflective markers on the patient surface to provide a source of motion information that is independent of the clinical imaging system(s). The types of patient motion for which compensation will be investigated for cardiac perfusion SPECT imaging with this combined approach are rigid-body motion (RBM), non-rigid-body motion (non-RBM), respiratory motion (RM), cardiac upward creep (which is caused by changes in RM), and motion between sequential multi-modality imaging studies. Note we correct body-motion and RM differently, thus we differentiate between them as part of motion estimation. In this competitive renewal we propose to extend our current VTS to include application with multimodality imaging, emission data-consistency, and finite-element-modeling (FEM) based correction of non-RBM. MRI of volunteers will be employed to establish the relationship between external-marker motion and the motion of the heart within the body. MRI will also provide high-resolution anatomy of the chest at different motion states in combination with measured-marker motion to guide the creation of realistic SPECT simulations for refinement of algorithms. Initial clinical testing will make use of repeat-rest cardiac-SPECT studies in patient-volunteers where the first rest imaging study serves as the standard for judging the success of correction of the second study during which the patients intentionally move. The ultimate assessment of the success of our hypothesis will be physician-observer ROC studies comparing the detection accuracy of coronary artery disease (CAD) with and without motion compensation for patients undergoing SPECT perfusion imaging with the results of cardiac catheterization serving as the "gold" standard. We believe the significance of our proposed investigations lies not only in the potential to improve diagnostic accuracy through motion-correction of both cardiac and non-cardiac SPECT, PET, and multimodality imaging but also in affording a better understanding of patient motion during imaging. PUBLIC HEALTH RELEVANCE: Patients commonly move during the protracted imaging period required by SPECT and PET to collect a statistically sufficient number of events to form images for making an accurate diagnosis. Such motion can cause artifactual changes in the slices which may mislead the physician. In multi-modality imaging patients also may move between emission imaging and imaging used to form attenuation maps and provide an anatomical framework for the functional changes in emission imaging. This can cause a misregistration between the two sets of images which can significantly impact clinical interpretation. Also all images acquired while patients are breathing suffer from respiratory motion blurring of the structures being imaged. The goal of this project is to develop an accurate and robust motion-correction methodology for all forms of patient motion which minimally alters current acquisition protocols so that it can be employed in a busy clinic. The test-bed we have chosen for development of this methodology is cardiac perfusion SPECT imaging which is not only the single most frequent SPECT imaging procedure performed today, but also plays a critical role in providing perfusion and functional information needed for risk assessment and stratification regarding coronary artery disease. Once the methodology is perfected for cardiac SPECT it can be adapted for use with other SPECT and PET imaging procedures. The proposed research will also provide a better understanding of patient motion during medical imaging and treatment.