The Impact of the Xspect Reconstruction Algorithms on the Recovery Coefficients Value for Small Tumors: A Phantom Study With 177Lu Publisher Pubmed



Ebrahimifard A^{1, 2, 3} ; Rajabi H¹ ; Geramifar P⁴ ; Luster M³ ; Yousefi BH³ ; De Nijs R^{2, 5} Show Affiliations
Source: Nuklearmedizin - NuclearMedicine Published:2024

Abstract

Aim Quantifying small tumors is still a challenge due to the partial volume effect (PVE). Although iterative reconstruction had promising results with a better recovery coefficient (RC), it suffers from the PVE. RC values typically depend on the reconstruction method, which may affect on 177Lu quantifying. In this study, we investigated the effect of different reconstruction methods on RC values for 177Lu SPECT/CT images. Methods A water-filled cylindrical polymethylmethacrylate (PMMA) phantom without background radioactivity with two solid cylinders which represent the bone and soft tissue equivalent was used for SPECT/CT imaging. We placed syringes with different volumes containing 177Lu in the vicinity of the cylinders as tumors. The iterative reconstructions of xSPECT, xSPECT Bone, and Flash3D was performed on SPECT/CT images.The effect of the reconstruction algorithms on the RC values was determined by calculating the radioactivity concentration in volumes of interest which were drawn manually and by a threshold method. Results The xSPECT Bone reconstruction has shown an RC value of approximately 100%, while the other reconstructions typically show an RC value of 93%. The RC values for all the reconstructions were decreased for smaller tumors. The smallest tumor showed a minimum RC between 52% and 72% compared to the largest tumor. Conclusion In this study, it was possible to obtain an RC value close to 100% when the xSPECT Bone reconstruction algorithm was used. The xSPECT results were slightly lower RC values compared to the Flash3D algorithm. In using standard RC curves for tumors close to the bone, special care has to be taken since RC values relative to bone are higher than in tissue. © 2024 Georg Thieme Verlag. All rights reserved.