Investigation of respiratory motion correction with and without motion correction of attenuation map in cardiac SPECT/CT in presence of non-rigid motion, an NCAT study

Joyoni Dey
W. P. Segars
P. Hendrik Pretorius
Michael A. King


Respiratory motion combined with "respiratory creep" of the heart can introduce false cooling or affect the appearance of defects and thereby impact the diagnostic accuracy of cardiac PET or SPECT perfusion images. In this work we investigate the differences in apparent imaging agent localization induced in emission images when the attenuation maps used for attenuation correction (from transmission or CT imaging) are misaligned with the patient anatomy during emission imaging due to variations in respiratory state. We first obtain various NCAT attenuation maps acquired at different stages of the respiratory cycle (end-expiration, end-inspiration, the mid-way map, and the average transmission map). We investigated use of these different attenuation maps to correct for attenuation in MLEM reconstruction of several variants of the NCAT phantom which included both with and without non-rigid motion between liver and heart. We tested these cases with and without emission motion correction. For most cases as long as we performed emission motion correction the false cooling artifacts induced by respiratory motion were greatly diminished and the appearance of the slices did not change significantly based on which of the attenuation maps we employed. In particular the quantitative RMS error with respect to the corresponding motion-less reference case reduced by factor of 0.93 to 0.47 times with emission correction versus without correction. However for some non-rigid cases there was were remaining cooling artifacts when we used the end-expiration map. © 2011 IEEE.