Static and dynamic reconstruction for gated cardiac SPECT

Myocardial perfusion imaging by electrocardiography (ECG) gated single-photon emission computed tomography (SPECT) is a very useful diagnostic tool for coronary artery disease and is frequently performed in clinical nuclear medicine. However, the effectiveness of gated SPECT imaging is often hampered by increased noise due to gating. In the first part of this work, the gated images are treated collectively as a spatio-temporal signal and reconstructed by a regularized expectation-maximization (EM) algorithm based on the maximum a posteriori (MAP) criterion. Compensation for degradation factors such as attenuation and scattering is taken into account in the reconstruction algorithm. For the evaluation study, we used the SIMIND Monte Carlo simulation package to simulate Tc-99m labeled Sestamibi imaging based on the NURBS-based cardiac-torso (NCAT) phantom. Our experimental results demonstrated that use of motion-compensated gate temporal regularization can lead to effective noise reduction in the reconstructed images, and that use of attenuation and scatter correction can greatly improve the reconstruction accuracy.; While the tracer distribution is assumed to be static in traditional gated SPECT, dynamic SPECT aims to obtain information about the tracer kinetics in the heart. However, the heart is usually treated as motionless in dynamic studies. The second part of this work is proposed to unify gated SPECT and dynamic SPECT into a single imaging method, called dynamic gated SPECT. Without changing the gated acquisition protocol, we aim to reconstruct a dynamic image sequence for each gate. To solve this severely underdetermined inverse problem, we propose to use dynamic shape constraints to regulate the tracer dynamic behavior and apply gate regularization as in our static gated SPECT reconstruction. To demonstrate the proposed method, we simulated gated cardiac perfusion imaging using both the 4D gated mathematical cardiac-torso (gMCAT) phantom and the NCAT phantom with Tc-99m labeled Teboroxime as the imaging agent. Our results show that the proposed method can produce an image sequence revealing information about both cardiac motion and tracer kinetics, and that reconstruction with motion-compensated gate regularization can yield more accurate reconstruction of dynamic images than reconstruction with spatial smoothness alone.