Abstract
1681
Objectives: The objective of this study is to investigate the respiratory motion effect on defect detection in Tc-99m Sestamibi myocardial perfusion SPECT (MPS) using a phantom population that realistically models patient variability.
Methods: Three respiratory amplitudes are included, i.e. no respiration, normal respiration, and large respiration. For each respiratory amplitude, six 3D NCAT phantoms, 3 males and 3 females, of different anatomies were generated. Inferior, anterior and left defects with 2 cm, 1.5 cm, 1 cm in size and contrast of 85%, 88.75%, 92.5% were inserted. Noise-free SPECT projections were simulated using an analytical projector. Poisson noise was then added to generate noisy realizations. Data were reconstructed using ML-EM and OS-EM for 1, 3, 5, 7, 10 iterations. Short-axis images containing the defect centroid were extracted then channelized Hotelling observer (CHO) was applied.
Results: The CHO results show that value of the area under the ROC curve (AUC) is significantly affected by the respiratory amplitude. For all defect contrasts studied, the optimal AUC values decrease with the increase of respiratory amplitude. With no respiration, the ranking of the optimal AUC value is anterior, inferior, then left defect. The AUC value of left defect drops more severely with increase in respiratory amplitude as compared to other defect locations. As the respiratory amplitude increases, smaller defect drops more quickly than larger ones. With large respiratory motion, the left defect has a lower AUC than that of the inferior defect.
Conclusions: We demonstrated that respiratory motion affects the defect detectability of MPS imaging. The development of optimization of the data acquisition and motion correction methods are needed to improve the defect detectability in MPS.
- Society of Nuclear Medicine, Inc.