Introduction: The number of angular projections has a significant impact on the image quality in SPECT imaging. Commonly used numbers of projections include 32, 64, and 128, with the selection typically depending on the imaging protocol used in each institution. Purpose: The aim of this master's thesis was to determine how the number of angular projections affects the quality of SPECT/CT images. Methods: A NEMA IEC Body Phantom was filled with 99mTc and scanned eight times using 32 angular projections and eight times using 64 angular projections. For each scan, the activity ratio between the background and the spheres was varied. The images were processed using the Oasis software, employing three different reconstruction methods: FBP, IT and IT with attenuation correction using CT (AC). We measured the number of counts in all six spheres and compared contrast and the contrast-to-noise ratio (CNR) across the different reconstruction methods and between the 32 and 64 angular projection datasets. Descriptive statistics and Mann-Whitney U-test were performed using SPSS software. Results: The analysis showed that images obtained with 64 angular projections did not show a statistically significant improvement in contrast or CNR compared to those with 32 projections, across all three reconstruction methods. The best results were achieved with attenuation-corrected images, followed by iterative reconstruction, while filtered back projection produced the lowest image quality. Visual assessment did not reveal significant differences between 32 and 64 projections in any of the reconstruction methods. However, subjective evaluation suggested a slight preference for images obtained with 64 projections. Discussion and conclusion: It was observed that both contrast and CNR increased with larger sphere sizes and higher sphere-to-background activity ratios, regardless of whether 32 or 64 projections were used. The highest contrast and CNR were achieved with attenuation-corrected reconstructions, followed by iterative reconstructions, with filtered back projection performing the worst. While 64 angular projections did not statistically significantly improve image quality compared to 32 projections, the visual assessment indicated a slight advantage. Considering the reduced acquisition time and lower radiation exposure to the patient, the use of 32 angular projections can be recommended for routine diagnostic imaging. However, it is important that the number of projections is adapted to the specific imaging protocol and clinical requirements of each institution
|
