Investigation of Respiratory Motion Compensation Techniques in Myocardial SPECT


W. Paul Segars, T.S. Lee, and B.M.W. Tsui


The purpose of the study is to evaluate two respiratory motion compensation methods for myocardial SPECT, respiratory gating and summing of motion corrected projections using motion tracking.  

Fig. 1.  Extent of RM artifacts for the two compensation methods using different gating schemes. Intensity ratio (IR) values approaching 1 indicate minimal artifacts.

Fig. 2.  Results from the two compensation methods using an optimal 3 gate scheme.


The 4D NCAT phantom with its realistic respiratory model was used to generate 96 NCAT phantoms equally spaced over a complete respiratory cycle modeling the activity distribution of Tc-99m Sestamibi. The heart included a lateral wall defect was set to move 2 cm longitudinally. The respiratory cycle was divided into different gating schemes (1, 3, 6,¡­24 gates) by summing combinations of phantoms. Projection data were generated for each gate including the effects of attenuation (A), collimator-detector response (D) and scatter (S). Poisson noise was added equivalent to that of a typical clinical study. In the first compensation method, the projection data from each gate were reconstructed using OS-EM (with ADS corrections). Bull¡¯s-eye polar plots were generated from the reconstructed image for each gate and summed into one plot for the gating scheme. In the second compensation method, motion information from the phantom was used to shift the projections of each gate so the heart position precisely matched that of the first gate. This technique can also be used without respiratory gating. The aligned projections for each gating scheme were then summed, reconstructed using OS-EM, and used to produce a bull¡¯s-eye plot. Regions-of-interest in the left ventricle (LV) were used to compare the bull¡¯s-eye plots produced from the respiratory gating and motion-tracking methods.


Respiratory motion (RM) artifacts were reduced with both compensation methods which produced comparable results, Fig. 1. The artifacts were reduced the most when going from 3 to 6 gates. The reduction was less when going from 6 to 8 gates. An optimal gating scheme using 3 gates was tested. The uniform gates were arranged so as to minimize the respiratory motion in each. Gates on the rising and falling edges of the respiratory volume curve were summed into one gate while the remaining two gates covered the transition from inspiration to expiration and vice versa. The optimal gating scheme was found to produce comparable results to those obtained using 6 and 8 gates, Fig. 2.


Respiratory compensation through uniform respiratory gating and summing the resulting polar plots from each gate is an effective way to reduce RM artifacts. Respiratory compensation through shifting of projections is also effective, but may be limited depending on the method chosen to automatically determine the amount to shift each projection. In this study, the amount of shift for each projection was known from the phantom and produced the best possible results. It may be possible to design an optimal 3 gate respiratory gating scheme that will produce results similar to those using a higher number of gates. 


NIH research grant R01HL68075