---

Background: Butterworth, Gaussian, Hamming, Hanning, and Parzen are commonly used SPECT filters during filtered back-projection (FBP) reconstruction, which greatly affect the quality and size accuracy of image. Materials and Methods: This study involved a cardiac phantom in which 1.10 cm thick cold defect was inserted into its myocardium wall and filled with 4.0 μCi/ml (0.148 MBq/ml) 99mTc concentration. The cardiac insert was then put into a cylindrical tank which was filled with 1.2 μCi/ml (0.044 MBq/ml) 99mTc concentration as background. 272 combinations of filter parameters were selected from those filters and applied on image. The measurements of count in myocardium, background, and defect regions of interest (ROI) were performed on each filtered image. Those measurements were then used to calculate contrast, signal-to-noise ratio (SNR), and defect size. For every filter, each criterion was graded (1 to 100) and then summed at their specific setting for total comparison. Results: The results show that, the different filter types produced myocardial image with different contrast, SNR, and defect size. For contrast and SNR, Gaussian filter was the best, while Parzen filter was the best in producing accurate defect size. However, Butterworth filter was found the best for trade off between contrast, SNR, and defect size accuracy. Conclusion: Selection of filter should consider the type of analysis, whether qualitatively or quantitatively. Qualitative analysis depends on image quality which is denoted by high contrast and SNR, thus Gaussian filter was suggested. Instead, the Butterworth filter was suggested for quantitative analysis as it is greatly dependent on both, image quality and size accuracy. Iran. J. Radiat. Res., 2009 6 (4): 173-182

---

Background: We investigated whether the lungheart ratio parameter (LHR) can be used to identify the optimum cut off frequency for Butterworth filter in 99mTc myocardial SPECT imaging. Materials and Methods: This study involved a cardiac phantom system consisting of cardiac insert in which 1.10 cm cold defect was inserted into its myocardium wall and filled with 4.0 μCi/ml (0.148 MBq/ml) 99mTc concentration. The cardiac insert was then put into a cylindrical tank which filled with six different 99mTc concentrations as background. Thus, six target background concentrations ratios (T/B) were carried out. The LHR was determined for every SPECT raw image obtained corresponding to each T/B. Then, 130 different combinations of filter parameters from Butterworth filter were utilized to reconstruct each SPECT raw image. The determination of count in myocardium, background, and defect regions of interest (ROI) were performed for every reconstructed image. All the count values were then used to calculate contrast, signal-to-noise ratio (SNR), and defect size. Each criterion was graded (1 to 100) and then summed together to obtain total grade. The optimum cut off frequency for each LHR was determined from the total grade. The relation between optimum cut off frequency for Butterworth filter and LHR was established using linear regression. Results: There were good relationship between the optimum Butterworth cut off frequency and LHR (R2 = 0.864, p < 0.01). The optimal cut off frequency correspond to the change in LHR can be expressed by the equation: Optimum cut off frequency = 0.715*LHR + 0.227. Conclusion: This study suggests that the optimum cut off frequency for Butterworth filter should be determined by referring to LHR in each patient study. Iran. J. Radiat. Res., 2010 8 (1): 17-24