Ponente
Descripción
Positron Emission Tomography (PET) relies on CT-based attenuation maps for accurate image reconstruction. The quality of these attenuation maps depends on CT acquisition parameters, particularly tube voltage and current. Higher tube currents improve image quality by reducing noise, but at the cost of higher radiation exposure. Conversely, reducing the current lowers the radiation dose but introduces more noise, which may affect PET quantification. Optimizing this balance is especially relevant in new long axial field-of-view PET systems, where higher sensitivity enables lower CT doses. This study evaluates how different CT dose levels affect PET image reliability, using the IMAS Total-Body PET system installed at La Fe Hospital in Valencia.
The experimental setup included the IMAS PET system, a GE Brightspeed CT scanner, and the NEMA IEC Body Phantom. The IMAS system employs semi-monolithic crystals, provides a 71.4 cm axial field of view, includes time-of-flight capabilities, and delivers spatial resolution below 4 mm uniformly across the field of view. CT scans were acquired at 80, 100, and 120 kV, with tube currents ranging from 10 to 120 mA in steps 10, producing 48 CT datasets. PET images were reconstructed with the OSEM algorithm using 20 iterations. Quantitative analysis of the CT images was performed with two regions of interest (ROIs): one covering the phantom background, excluding the lung insert, and one inside the lung insert. The mean and standard deviation of voxel values were extracted for each ROI. Additionally, representative CT dose levels were selected, and the corresponding PET images were evaluated according to the NEMA NU 2-2018 image quality protocol.
ROI analysis of the CT confirmed stable mean attenuation across tube currents and voltages, even at low doses. Noise increased in the background at low currents, especially at 80 kV, while the lung insert remained more stable, indicating that noise propagation affects uniform regions.
The analysis based of the PET images using the NEMA protocol, compared low, mid, and high CT dose levels. Contrast recovery improved with increasing sphere diameter, and was higher for mid and high dose acquisitions, while low-dose conditions consistently showed lower recovery values. Background variability decreased with sphere size but was always higher in the low-dose group than in the mid and high doses. Importantly, differences between mid and high dose conditions were minimal, suggesting that intermediate CT dose levels are sufficient to ensure robust PET image quality while avoiding unnecessary exposure.
In conclusion, the reliability of PET images in the IMAS system remains consistent across most CT dose levels, with mean attenuation preserved even at low doses. Still, very low currents, particularly at 80 kV, increase noise and reduce contrast recovery. NEMA analysis shows that mid-level doses perform similarly to high doses, offering the best balance between safety and diagnostic quality. The study confirms the feasibility of optimized low-dose PET/CT protocols.
