Interpolated Average CT for Attenuation Correction in PET—A Simulation Study

Previously, we proposed using an interpolated average CT (IACT) method for attenuation correction (AC) in positron emission tomography (PET), which is a good, low-dose approximation of cine average CT (CACT) to reduce misalignments and improve quantification in PET/CT. This study aims to evaluate the performance of IACT for different motion amplitudes. We used the digital four-dimensional (4-D) extended cardiac-torso phantom (XCAT) to simulate maximum of 2, 3, and 4 cm respiratory motions. The respiratory cycle was divided into 13 phases, with average activity and attenuation maps to represent 18F-fluorodeoxyglucose (18F-FDG) distributions with average respiratory motions and CACT, respectively. The end-inspiration, end-expiration, and midrespiratory phases of the XCAT attenuation maps represented three different helical CTs (i.e., HCT-1, HCT-5, and HCT-8). The IACTs were generated using: 1) 2 extreme+11 interpolated phases (IACT 2o ); 2) 2 phases right after the extreme phases+11 interpolated phases (IACT 2s); 3) 4 original+9 interpolated phases (IACT 4o). A spherical lesion with a target-to-background ratio (TBR) of 4:1 and a diameter of 25 mm was placed in the base of right lung. The noise-free and noisy sinograms with attenuation modeling were generated and reconstructed with different noise-free and noisy AC maps (CACT, HCTs, and IACTs) by Software for Tomographic Image Reconstruction, respectively, using ordered subset expectation maximization(OS-EM) with up to 300 updates. Normalized mean-square error, mutual information (MI), TBR, image profile, and noise-contrast tradeoff were analyzed. The PET reconstructed images with AC using CACT showed least difference as compared to the original phantom, followed by IACT 4o, IACT 2o , IACT 2s, HCT-5, HCT-8, and HCT-1. Significant artifacts were observed in the reconstructed images using HCTs for AC. The MI differences between IACT 2o and IACT 4o /CACT were <0.41% and <2.17%, respectively. With a slight misplacement of the two extreme phases, IACT 2s was still comparable to IACT 2o with MI difference of <2.23%. The IACT is a robust and accurate low-dose alternate to CACT.