Interpolated average CT for PET attenuation correction in different lesion characteristics.

Abstract

Previously we proposed using an interpolated average computed tomography (IACT) method as a low-dose alternate of cine average computed tomography (CACT) for PET attenuation correction (AC). This study aims to evaluate its performance for thoracic lesions with different characteristics in simulations and clinical patients. We used the XCAT phantom to simulate noisy fluorine-18 fluorodeoxyglucose (18F-FDG) distribution with respiratory motion amplitudes of 2 and 3 cm. Average activity and attenuation maps represented static PET and CACT, respectively. IACT was generated by the end-inspiration and end-expiration phases of the attenuation maps (HCT-in and HCT-ex) using a deformable registration method. Spherical lesions with diameters of 10 and 20 mm with four target-to-background ratios (TBRs) were simulated at four different locations individually, including the lower left lung, lower right lung, middle right lung, and upper right lung. Five patients with a total of six thoracic lesions were recruited. They were scanned 1 h after 315-480 MBq F-FDG injection. Simulated and clinical PET sinograms were reconstructed with AC using (i) CACT, (ii) IACT, and (iii) helical computed tomography (HCTs). The TBRs and mean standardized uptake value were analyzed. Significant artifacts were observed in PET HCTs from visual assessment. For both simulation and clinical study, PET IACT was more similar to PET CACT in terms of TBRs and mean standardized uptake value. The differences between CACT/IACT and HCTs were more significant for lesions located at the lower lungs. The IACT is a robust and low-dose AC method for improved thoracic lesion localization and quantitation for a wide range of lesion characteristics.