Optimization of acquisition time and reconstruction parameters for quantitative single-photon emission computed tomography/computed tomography using iodine-123 metaiodobenzylguanidine

Abstract

This study aimed to determine the optimal measurement conditions for accurate standardized uptake value (SUV) analysis of iodine-123 metaiodobenzylguanidine (¹²³I-MIBG) by examining the relationship between image convergence and quantitation. Single-photon emission computed tomography/computed tomography images were acquired using JS-10 and National Electrical Manufacturers Association (NEMA) body phantoms, with acquisition time per view varied (10, 30, 50, and 100 s/view). Image reconstruction was performed using three-dimensional-ordered subset expectation maximization, adjusting the product of subset and iteration (SI product; 60, 120, 180) and Gaussian filter parameters (8, 10, 12 mm). For the JS-10 phantom, we evaluated the dose linearity (DL), the recovery coefficient (RC) of individual rods, the scatter ratio (SR), and the coefficient of variation (CV). For the NEMA body phantom, we assessed the contrast-to-noise ratio (CNR) of the 17-mm-diameter hot sphere. We also evaluated the maximum and mean SUVs for all its hot spheres, and their relative standard error (RSE), using SUVs obtained at 100 s/view as reference. In the JS-10 phantom, the DL remained stable under all conditions. The RC decreased when the Gaussian filter was large and the SI product was small. A trade-off between the CV and the SR emerged, depending on the acquisition time and the SI product; optimal results were observed at 50 − 100 s/view and an SI product of 120 − 180. In the NEMA body phantom, contrast improved with acquisition times of ≥30 s/view, and the CNR increased as noise declined with longer acquisition times. At ≥50 s/view, variation in the maximum and mean SUVs decreased, with the RSE remaining below 5%. In conclusion, accurate SUV measurement with ¹²³I-MIBG requires an acquisition time of ≥50 s/view, an SI product of approximately 120, and a Gaussian filter of 10 − 12 mm. These findings provide a foundation for future studies comparing this method with the heart-to-mediastinum ratio, supporting its clinical application.