Development of a myocardial phantom and analysis system toward the standardization of myocardial SPECT image across institutions

Objective We developed a novel myocardial phantom and analysis program to standardize using a quantitative index to objectively evaluate the image quality. We aimed to reveal whether our proposed phantom and analysis program are suitable for image standardization. Methods An evaluation system of myocardial image based on technical grounds (EMIT) phantom was developed to standardize the image quality of myocardial SPECT and was constructed with the lung and myocardium in the thorax phantom; the myocardial phantom included five normal areas and eight defective areas with four defects in size (5, 10, 15, and 20 mm) and four defects in thickness (10, 7.5, 5, and 2.5 mm). Therefore, this phantom was appropriate to simultaneously simulate eight different defects and normal myocardium. The %rate value, calculated using the region of interest method, and the %count value, calculated from the profile method, were automatically analyzed to evaluate myocardial defects. The phantom was validated using difference in count levels and filter parameters compared with those in previously reported models. Results The average %count of eight defects by 0.3, 0.4, 0.5, and 0.6 cycles/cm were 56.8, 47.4, 44.3, and 43.4 %, respectively, whereas the %count for 0.3 cycles/cm was significantly higher than that for 0.5 and 0.6 cycles/cm. The uniformity between full- and half-time images was 16.5 ± 4.2 and 18.7 ± 5.5 % for integral uniformity and 3.4 ± 1.2 and 3.4 ± 1.3 % for differential uniformity, respectively, revealing a significant difference in integral uniformity between the two acquisition times. Visual differences in defects were evident in full-time images between 0.30 and 0.50 cycles/cm, and defect detectability of the myocardial image at 0.30 cycles/cm was poor. Normal myocardial thickness widened in comparison with images at 0.50 cycles/cm. Compared with full-time myocardial image at the same cut-off frequency, the half-time myocardial image demonstrated inhomogeneous distribution and thickness of the normal myocardium. Conclusion We developed a new phantom and program to standard image quality among multicenter for myocardial SPECT. The EMIT phantom and quantitative indices were useful for evaluating image quality. The physical characteristics of the image quality, including defects and uniformity, were properly measured by this method.