Motion-compensated Image Reconstruction Vs Postreconstruction Correction in Respiratory-Binned SPECT with Standard and Reduced-Dose Acquisitions
Chao Song,Yongyi Yang,Wenyuan Qi,Miles N. Wernick,P. Hendrik Pretorius,Michael A. King
DOI: https://doi.org/10.1002/mp.12932
2018-01-01
Abstract:PurposeCardiac 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. MethodsWe 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. ResultsThe 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. ConclusionsRespiratory 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.