Abstract
2041
Objectives Our template-based method quantifies SPECT myocardial perfusion defect extent (PDE) without the use of normal databases; however, it is currently limited by extensive calculations. In the interest of clinical feasibility, we examine the trade-off between accuracy and processing time as the algorithm is simplified.
Methods Our method uses the patient’s SPECT image to create a 3D digital template of his/her healthy heart. This template is then projected, reconstructed and sampled into the bulls-eye map domain. A ratio of the patient and template images produces a final corrected image in which a weighted threshold is applied to identify perfusion defects. Theoretically, the template projector should include a scatter component; however, this leads to lengthy calculations. In an attempt to optimize our method, we analyzed the performance of 4 projectors of decreasing complexity: P+S+N, P+S, P+N and P, where P, S, and N represent primary photons, scattered photons, and Poisson noise, respectively. Eight simulated phantoms (2 healthy, 6 with perfusion defects), 4 physical thorax phantoms (1 healthy, 3 with perfusion defects), and 14 randomly selected patient studies were analyzed. All studies used GE`s Infinia Hawkeye4 SPECT/CT system following standard cardiac protocol. PDE results obtained from Cedars Sinai’s QPS software were considered the gold standard for the patient studies.
Results In all cases, the differences in mean PDE error were small (less than 2%), yet large differences were observed in processing time: the P method required less than 1 minute versus 2 hours for the P+S method. The addition of N did not considerably improve the PDE estimates.
Conclusions The template-based technique leads to accurate, fast, and patient-specific PDE measurements when the modeling of only primary photons is included