CT-Based Attenuation Correction on the FLEX Triumph Preclinical PET/CT Scanner

Abstract

Positron Emission Tomography (PET) has emerged as a valuable molecular imaging modality for quantitative measurement of biochemical processes in vivo in the clinical and preclinical imaging domains. However, PET imaging suffers from various physical degrading factors including photon attenuation, which can be corrected using CT-based attenuation correction (CTAC) on combined PET/CT scanners. The attenuation map is calculated by converting CT numbers derived from low-energy polyenergetic x-ray spectra to linear attenuation coefficients at 511 keV. Generation of accurate attenuation maps is crucial for reliable attenuation correction of PET data and hence is a prerequisite for accurate quantification of biological processes. In this study, we implemented the CTAC procedure on the FLEX Triumph™ preclinical PET/CT scanner and evaluated tube voltage dependence for different kVps (40, 50, 60, 70, and 80). The quantitative impact of both bilinear and quadratic based energy-mapping methods on linear attenuation coefficients, attenuation maps and corrected PET images was assessed at different CT tube voltages. Attenuation maps were calculated from CT images of a cylindrical polyethylene phantom containing different concentrations of K2HPO4 in water. Correlation coefficients and best regression fit equations were calculated for both methods. Phantom and rodent PET/CT images were used to assess improvements in image quality and quantitative accuracy. It was observed that the slopes of the bilinear calibration curves for CT numbers greater than 0 HU increase with increasing tube voltage. In addition, higher correlation coefficients were obtained for the quadratic compared to the bilinear energy-mapping method. Tube voltage of 70 kVp produced the smallest relative error and higher correlation coefficient compared to other tube voltages. For low concentrations of K2HPO4, the mean relative difference (in %) between theoretical and calculated attenuation coefficients when using bilinear and quadratic energy-mapping methods are 1.39 ± 1.9 and 1.33 ± 1.8, respectively. They are 2.78 ± 1.3 and 2.5 ± 1.3, respectively, for high concentrations of K2HPO4 . As expected, higher activity concentrations were obtained for PET after attenuation correction. The increased PET signal for mouse tissues ranged between 21 and 31% for bilinear energy-mapping and between 21.8 and 35% for quadratic energy-mapping, whereas these varied from 40 to 51% and from 41 to 56%, respectively, for rat tissues. For biological tissues having a high atomic number such as bone, the quadratic energy-mapping method produced slightly improved results compared to the bilinear energy-mapping method. Phantom and rodent PET studies were successfully corrected for photon attenuation using the developed CTAC procedure.