Effect of respiratory motion on plaque imaging in the mouse using Tc-99m labeled Annexin-V


W. Paul Segars, Y. Wang, and B.M.W. Tsui


Transgenic mouse models bred to develop atherosclerotic plaques in the aorta are currently being used to investigate microSPECT imaging techniques to differentiate stable and vulnerable plaques. Respiratory motion is a major factor that can cause blurring and artifacts in the resulting microSPECT images. We investigate the effects of respiratory motion on microSPECT imaging of atherosclerotic plaques in the mouse using a new, realistic 4D digital mouse phantom developed in our laboratory.  

Fig. 1.  (Left) Mouse heart phantom with a plaque modeled in the aortic arch. (Right) Volume renderings (fused transmission and emission images) showing reduced plaque contrast in the case with an average respiratory motion.


The anatomy of the phantom was based on magnetic resonance imaging (MRI) data from the Duke Center for In Vivo Microscopy and includes a realistic respiratory model based on respiratory-gated MRI data of a normal mouse from the University of Virginia. Plaques of varying sizes were modeled in the phantom as 3D surfaces on the interior wall of the aortic arch. Voxelized phantom sets, modeling the distribution of attenuation coefficients and radioactivity concentrations in the organs, were then generated with and without an average respiratory motion. The radioactivity concentrations were set to model a typical Tc-99m labeled Hynic-Annexin-V study. Three different activity uptake ratios (relative to the background) were simulated for the plaques: 9 to 1, 18 to 1, and 36 to 1. Different levels of respiratory motion were also simulated with the heart moving 0.5, 1.0, and 2.0mm with the diaphragm. Emission projection data were generated from the voxelized phantoms using a realistic pinhole SPECT projection model. Poisson noise was added to the projections equivalent to a typical small animal imaging application. The projection data were reconstructed into 80x80x80 arrays with a pixel width and slice thickness of 0.3 mm. The contrast and signal to noise ratio of the plaques in the resulting images was then assessed to determine the effect of the respiratory motion.


The measured uptake in the plaques was found to decrease significantly with increasing levels of respiratory motion, Fig. 1. The reduction in plaque contrast was as high as 60% in the reconstructed images with an average cardiac respiratory motion of 2mm.  


We conclude that respiratory motion contributes significantly to blurring and artifacts in microSPECT plaque imaging in small animals. The 4D digital mouse phantom is a useful tool to assess the effects of respiratory motion, to investigate optimal imaging parameters, and to design respiratory gating schemes for improved microSPECT imaging of plaques.


NIH Research Grant R01EB00168