Abstract
T4
Introduction: A circular orbit is appropriate for uniformity of myocardial accumulation, while a neighboring elliptical orbit is appropriate for lesion detection, making it challenging to achieve both. The purpose of this study was to propose a novel acquisition orbit that positions the heart at the center of rotation to achieve both uniformity and lesion contrast, and to assess its image quality.
Methods: Single-photon emission computed tomography with computed tomography scanning of an anthropomorphic cardiac phantom was performed with circular, neighboring elliptical (NE), and cardiac-centered circle (CCC) orbits. CCC orbit was achieved with a 10 cm spacer placed in front of the phantom based on the NE orbit. Reconstructions were performed with filtered back projection (FBP), ordered subset expectation maximization with resolution recovery (OSEMRR) with or without attenuation and scatter correction. A count profile curve of short-axis images was obtained by circumferential and linear profile using DRIP Version 3.0.2.0 (PDRadiopharma, Tokyo, Japan). Circumferential profiles were set up in three slices: the center of the defect slice (defect slice) and homogeneous slices displaced 7 slices to apex (apex slice) and base (base slice), respectively. Lesion contrast, uniformity of normal accumulation, and count distortion were calculated from circumferential profiles. Normal myocardial uniformity was defined as %CV and calculated by dividing a standard deviation by an average count of the apex slices.
Results: The lesion contrast was highest in NE orbit in FBP (0.55), slightly lower in CCC orbit in FBP (0.51) and markedly lower in circular orbit than in NE orbit in FBP (0.45). Uniformity and count distortion were the best for CCC orbits in OSEMRR. NE orbit in FBP exhibited a clear low accumulation artifact, around the 11 o'clock position.
Conclusions: CCC orbit achieved uniformity and reduced count distortion while maintaining lesion contrast. Although CCC orbit provides excellent image quality regardless of reconstruction method, we believe it is particularly suitable for combination with OSEMRR reconstruction method.