UC Davis

The emergence of long axial field-of-view (FOV) total-body positron emission tomography (PET) systems has enabled a broad range of new applications previously not feasible due to the limited axial coverage. The 40-fold increase in system sensitivity across the entire adult human (compared to shorter, conventional PET systems) of the 194 cm long EXPLORER total-body PET scanner has enabled 1) reduction of scan time, 2) reduction of radiation dose, 3) improved dynamic range of radiotracer, 4) improved image quality, 5) system-level kinetic modeling of the human body, and more. However, the technical challenges stemmed from total-body PET systems are not thoroughly investigated. The global quantification of radioactivity in humans is not well studied since prior PET systems did not possess the necessary axial coverage to encompass the entire living human. The absolute quantification of regional tissue radioactivity concentration is a critical parameter of interest across various clinical and research applications and can be affected by a complex interplay of factors including scanner calibration, data correction, and image reconstruction. The increase in axial acceptance angle offered by total-body PET systems, while improves system sensitivity, can also make scatter correction more challenging. In addition, the high data rates caused by the large axial extent of the system impose significant computational and data storage burdens on the supporting infrastructure.

The goal of this work is to establish the quantitative performance baseline for long axial FOV PET systems using the EXPLORER total-body PET scanner, as well as to develop a computational foundation to address the high data rates imposed by such systems as well as its derivatives. To that end, we comprehensively assessed the quantitative accuracy of the EXPLORER scanner using a wide variety of phantoms as well as in healthy humans. Our results overall indicated that the quantitative performance achieved with the EXPLORER scanner was uniform across the axial FOV and provided the accuracy necessary to support a wide range of imaging applications spanning from low-dose studies to dynamic imaging with commonly used image reconstruction frame lengths.

Next, we investigated the relationship between 18F-fluorodeoxyglucose (18F-FDG) image signal-to-noise ratio (SNR) and noise-equivalent count rate (NECR) in total-body PET using the EXPLORER scanner. To estimate the complex scatter distributions observed in total-body PET systems, we developed a Monte Carlo scatter correction framework for total-body PET that utilizes continuous water density materials as the attenuation input. We discovered that the use of time-of-flight NECR (TOF-NECR) as a count rate performance metric over conventional non-TOF NECR at radioactivity levels beyond peak non-TOF NECR may be more suitable for assessing the count rate performance of PET systems with TOF capabilities.

To address the massive single event data rates imposed by total-body PET systems, we developed a high-performance, software-based coincidence processor capable of processing EXPLORER list-mode single event data in near real-time using several computational nodes.

Finally, we developed a simulation framework and performed a scanner sensitivity study of the NeuroEXPLORER (NX) scanner, the next-generation dedicated brain PET/CT scanner. The simulation results showed that the increase in solid angle coverage of the NX scanner over both the HRRT-D and Biograph Vision scanners can lead to approximately a 2-fold increase in peak system sensitivity, as well as an approximate 5-fold increase in total system sensitivity.

Overall, this work took a first dive into tackling new questions in total-body PET that were not examined in detail previously and established the quantitative and computational foundations for future total-body PET research and its derivatives requiring quantitatively accurate and high-throughput capabilities.