Optimization of iterative image reconstruction in PET/CT imaging procedures: A Patient-centered approach

Moustafa, Mohamed S.; Elsayed, Tamer M.; Kany, Amr M.I.; El-maghraby, Sherif

doi:10.21608/egyjnm.2025.356891.1113

Optimization of iterative image reconstruction in PET/CT imaging procedures: A Patient-centered approach

Document Type : Original Article

Authors

¹ Department of Physics, Faculty of Science, Al-Azhar University, Cairo, Egypt.

² Oncology and Nuclear Medicine Department, Kasr Al-Aini Hospital, Cairo University, Egypt.

10.21608/egyjnm.2025.356891.1113

Abstract

Aim. The goal of this study is to optimize image quality metrics including Signal-to-Noise Ratio (SNR), Contrast-to-Noise Ratio (CNR), and Lesion-to-Background Ratio (LBR) while maintaining consistent Standardized Uptake Value (SUV(g/ml)) measures using patient and phantom data. The optimized reconstruction parameters were compared with the default reconstruction parameter sets of the General Electric Discovery IQ PET/CT (GE-IQ) and the United Imaging PET/CT (uMI550) scanners.
Results. Image quality analysis demonstrated wide variations in SUVmax, SUVmean, SNR, CNR, and LBR across different reconstruction parameters in both scanners. In GE-IQ scanner, parameter sets utilizing optimized Z-axis weighting, specific iteration and subset, reduced Full Width at Half Maximum (FWHM) consistently outperformed other approaches, including the default parameter set, by delivering higher SUV values and improved image quality. Similarly, in the uMI550 scanner, a parameter set incorporated tighter FWHM, increased subset configurations, and a smoothing filter demonstrated superior image quality and diagnostic accuracy. Overall, the optimized parameter sets on the GE-IQ and uMI550 scanners exhibited a pronounced enhancement in image quality metrics, indicating its superior effectiveness for diagnostic applications, when compared to default parameter sets that recommended by the vendor of each scanner.
Conclusions. The findings underscore the potential of optimizing reconstruction techniques in PET/CT imaging. Parameter set 39 (Z-axis filter = heavy, iteration = 2, subsets = 12, FWHM = 5.0 mm) in GE-IQ and parameter set 8 (Filter = smoothing1, iteration = 2, subsets = 20, FWHM =3.0 mm) in uMI550 scanners demonstrated superior performance, offering promising avenues for enhancing patient outcomes and advancing medical imaging practices.

Keywords