Medical Imaging Research Center, Department of Electrical and Computer Engineering, Illinois Institute of Technology, Chicago, IL, 60616, USA.
Department of Radiology, University of Massachusetts Medical School, Worcester, MA, 01655, USA.
Med Phys. 2018 Jul;45(7):2991-3000. doi: 10.1002/mp.12932. Epub 2018 May 13.
Cardiac perfusion images in single-photon emission computed tomography (SPECT) can suffer from respiratory motion blur. We investigated a reconstruction approach for correcting respiratory motion in respiratory-binned acquisitions and assessed the benefit of this approach in both standard dose and reduced dose.
We modeled the acquired data from different respiratory bins by a joint probability distribution which was parameterized with respect to a common reference bin. The acquired data from all the respiratory bins were then utilized simultaneously for determining the source distribution in the reference bin using maximum a posteriori (MAP) estimation. We evaluated this approach with simulated imaging data and ten sets of clinical acquisitions, and compared it with a postreconstruction motion correction approach developed previously. We quantified the accuracy of the reconstruction results both at standard dose and with imaging dose reduced by 50% and 75%, respectively.
The proposed motion-compensated reconstruction (MCR) approach led to improved reconstruction of the myocardium in terms of both noise level and LV wall resolution. Compared to traditional acquisition (without motion correction), the proposed approach reduced the mean squared error of the image intensity in the myocardium by 27.59%, 20.59%, and 12.05% at full, half-, and quarter dose, respectively; the LV resolution, quantified by the full width at half-maximum (FWHM), was improved by 17.34%, 14.35%, and 12.95% at full, half-, and quarter dose, respectively; in addition, the proposed approach also improved the perfusion defect detectability at both full dose and reduced dose. Furthermore, with motion correction, the reconstruction results obtained at half-dose were comparable to that obtained at full dose without correction. Similar improvements were also demonstrated in the clinical acquisitions at different dose levels.
Respiratory motion correction in perfusion SPECT can improve the reconstruction of the myocardium at both standard and reduced dose. At half-dose, the results obtained with motion correction are comparable to that of traditional reconstruction obtained at full dose. MCR can be more accurate than postreconstruction correction.
单光子发射计算机断层扫描（SPECT）中的心脏灌注图像可能会受到呼吸运动模糊的影响。我们研究了一种用于校正呼吸门控采集运动的重建方法，并评估了该方法在标准剂量和降低剂量下的益处。
我们通过一个联合概率分布来对来自不同呼吸门的采集数据进行建模，该分布针对一个共同的参考门进行参数化。然后，我们同时利用所有呼吸门的采集数据，使用最大后验（MAP）估计来确定参考门中的源分布。我们使用模拟成像数据和十组临床采集来评估这种方法，并将其与以前开发的一种后重建运动校正方法进行比较。我们分别在标准剂量和降低 50%和 75%的成像剂量下，量化了重建结果的准确性。
与传统采集（无运动校正）相比，所提出的运动补偿重建（MCR）方法在噪声水平和左心室壁分辨率方面都改善了心肌的重建。与传统采集（无运动校正）相比，该方法分别将心肌图像强度的均方误差降低了 27.59%、20.59%和 12.05%；左心室分辨率（以半最大值全宽（FWHM）表示）提高了 17.34%、14.35%和 12.95%；此外，该方法还提高了在全剂量和降低剂量下的灌注缺陷检测能力。此外，使用运动校正，在半剂量下获得的重建结果与未经校正的全剂量下获得的结果相当。在不同剂量水平的临床采集中也显示出了类似的改善。
在灌注 SPECT 中进行呼吸运动校正可以改善标准剂量和降低剂量下的心肌重建。在半剂量下，使用运动校正获得的结果与全剂量下传统重建获得的结果相当。MCR 可以比后重建校正更准确。
