Assessment of the Impact of CT Respiratory Motion on PET SUV Quantification through Virtual Imaging 📝

Author: Ehsan Abadi, Darrin Byrd, Paul E. Kinahan, Katie Marie Olivas, Ehsan Samei 👨‍🔬

Affiliation: Duke University, Center for Virtual Imaging Trials, Duke University, Clinical Imaging Physics Group, Department of Radiology, Duke University Health System, University of Washington 🌍

Abstract:

Purpose: To evaluate the impact of respiratory motion during CT acquisitions on PET image quantification using an integrated PET-CT simulation pipeline.
Methods: A validated CT simulator (DukeSim) and a PET simulator (SimSET) were integrated to form a PET-CT simulation pipeline which facilitates consistent I/O, relatable coordinates, and a streamlined workflow for multi-modality virtual imaging trials. The pipeline enables simulating PET-CT acquisition and reconstruction processes with computational phantoms as inputs. For this study, we simulated acquisitions of four-dimensional extended cardiac-torso (XCAT) phantoms with respiratory motion. PET emission data was simulated for a five-minute, free-breathing scan and reconstructed with a TOF-OSEM algorithm (34 subsets, 4 iterations). The reconstruction was done with four different CT attenuation correction (CTAC) maps: inspiration breath-hold, expiration breath-hold, and free-breathing (six and sixty phases). Simulated CT projections were acquired in helical mode (120 kVp, 100 mA, 0.50 second gantry rotation time, 1.0 pitch). Mean and maximum SUVs of the liver and blood pool were measured and compared to ground truth using consistent regions of interest.
Results: Respiratory motion during simulated CT aquisitions affected the recorded SUVmean and SUVmax values. For inspiration CTAC, SUVmean/SUVmax were 2.26/3.79 and 4.50/6.87 for liver and blood pool, respectively. Corresponding values were 3.35/4.82 and 4.73/7.18 for expiration CTAC, 2.83/4.32 and 4.57/6.95 for free-breathing CTAC with six phases, and 3.30/4.74 and 4.70/7.16 for free-breathing CTAC with sixty phases.
Conclusion: This study demonstrated that respiratory motion during CT acquisition can markedly affect SUV quantification, at levels that influence patient management. This highlights the importance of optimizing and standardizing CT acquisition protocols for PET-CT to minimize and make consistent the effect of respiratory motion. The integrated PET-CT simulation platform and virtual imaging trial technology enable systematic evaluation of protocol impacts, offering insights to refine imaging techniques and improve the accuracy of PET-based quantification and disease characterization.

Back to List