Interpolated CT for attenuation correction on respiratory gating cardiac SPECT/CT — A simulation study

Purpose: Respiratory gated four-dimensional (4D) single photon emission computed tomography (SPECT) with phase-matched CT reduces respiratory blurring and attenuation correction (AC) artifacts in cardiac SPECT. This study aims to develop and investigate the effectiveness of an interpolated CT (ICT) method for improved cardiac SPECT AC using simulations. Methods: We used the 4D XCAT phantom to simulate a population of ten patients varied in gender, anatomy, ^99mTc-sestamibi distribution, respiratory patterns, and disease states. Simulated 120 SPECT projection data were rebinned into six equal count gates. Activity and attenuation maps in each gate were averaged as gated SPECT and CT (GCT). Three helical CTs were simulated at end-inspiration (HCT-IN), end-expiration (HCT-EX), and mid-respiration (HCT-MID). The ICTs were obtained from HCT-EX and HCT-IN using the motion vector field generated between them from affine plus b-spline registration. Projections were reconstructed by OS-EM method, using GCT, ICT, and three HCTs for AC. Reconstructed images of each gate were registered to end-expiration and averaged to generate the polar plots. Relative difference for each segment and relative defect size were computed using images of GCT AC as reference. Results: The average of maximum relative difference through ten phantoms was 7.93 ± 4.71%, 2.50 ± 0.98%, 3.58 ± 0.74%, and 2.14 ± 0.56% for noisy HCT-IN, HCT-MID, HCT-EX, and ICT AC data, respectively. The ICT showed closest defect size to GCT while the differences from HCTs can be over 40%. Conclusion: We conclude that the performance of ICT is similar to GCT. It improves the image quality and quantitative accuracy for respiratory-gated cardiac SPECT as compared to conventional HCT, while it can potentially further reduce the radiation dose of GCT.