OPTIMIZATION OF ACTIVITY CONCENTRATION AND RECONSTRUCTION PARAMETERS OF OSEM AND FBP METHODS IN BRAIN SPECT IMAGING

Abstract

Brain SPECT imaging is a nuclear medicine study which can detect blood flow and activity in the brain of the patients with neurological and psychiatric disorders. The major problem of brain SPECT imaging is that when the administered activity is low, scanning time is short or counts rate is low, the image quality was reduced and difficult for diagnosis. To compensate for this problem, the optimal reconstruction parameters can be applied to improve the image quality in terms of contrast, noise quantitatively and visual scoring on brain SPECT images. The purpose of this study was to determine the optimal reconstruction parameters of different activity concentration in brain SPECT images by using Hoffman 3-D brain phantom. Three different activities of 99mTcO4 solution, 55.5-MBq (1.5 mCi)-low activity, 111-MBq (3 mCi) -normal activity, and 165.5-MBq (4.5 mCi)-high activity had been inserted in Hoffman 3D brain phantom for three acquisitions and reconstruct using OSEM with various update numbers and full width at half maximum (FWHM) of Gaussian filter and FBP with various cut-off frequencies and Butterworth filter were applied. The percent contrast and noise of gray and white matter were calculated to determine optimal reconstruction parameters in brain SPECT imaging for quantitative measurement and visual scoring from two nuclear medicine physicians for qualitative analysis. The optimal parameters for 3D-OSEM method in low activity concentration (46 kBq/cc) were 8-iteration and 8-subsets (64-iterative updates), percent contrast and noise were 66.00 % and 14.60 and score of 12, for normal activity concentration (92kBq/cc) were 10-iteration and 8-subsets (80-iterative updates), percent contrast and noise were 78.00 % and 14.00 and score of 13, for high activity concentration (138 kBq/cc) were 12-iteration and 8-subsets (96-iterative updates), percent contrast and noise were 84.00 % and 13.60 and qualitative score of 14 respectively. FWHM of Gaussian filter 5-mm and Chang’s attenuation coefficient of 0.12cm-1 were fixed for each activity concentration. For FBP reconstruction, low activity concentration (46 kBq/cc), the optimal parameters were 0.35 cycles/pixel, order 10, percent contrast and noise 52.92 % and 11.80 and score of 10. For normal activity concentration (92 kBq/cc) was 0.45 cycles/pixel, order 10, percent contrast and noise were 62.19 % and 11.00 and qualitative score of 10, for high activity concentration (138 kBq/cc), the optimal parameters were 0.45 cycles/pixel, order 10, percent contrast and noise 68.61 % and 10.70 and qualitative score of 13 respectively. Post processing filter of Butterworth filter and Chang’s attenuation coefficient of 0.14cm-1 were applied for each activity. In conclusion, the image quality had been determined to obtain the optimal image reconstruction parameters of OSEM and FBP on the low, normal and high activity concentration for Hoffman brain phantom. This study showed that image quality can be improved by optimizing reconstruction parameters especially in low activity concentration to reduce the patient dose. OSEM reconstruction method is the best choice for low counts statistic. These optimizing parameters can be implemented in routine clinical studies.