DESIGN CONSIDERATIONS FOR AN ULTRA-LONG AXIAL FOV PET/CT SCANNER

Objectives: Ultra-long axial field-of-view (FOV) PET/CT scanners have great potential in research and clinical applications, given their high sensitivity and capability to image simultaneously most organs of interest. Because of the complexity and cost of such devices, a trade-off needs to be found between different design parameters, such as axial coverage, accepted opening angle for lines-of-response (LORs), time coincidence window, and the performance of the scanner, such as spatial resolution, sensitivity, peak noise equivalent count rate, as well as site related factors. A wide acceptance angle provides high sensitivity, potentially at the cost of axial spatial resolution degradation. A wide coincidence time window ensures the collection of all true coincidence events at the cost of high randoms count rate. In addition, as more oblique LORs are accepted, body attenuation increases, reducing the amount of true coincidences, while increasing the random and scatter fractions. Furthermore, the adoption of ultra-long axial FOV for research and routine clinical applications will include financial factors. The aim of this work was to evaluate the relationship between several design parameters and their impact on the performance of an ultra-long axial FOV scanner based on the Siemens Biograph Vision’s technology, and to understand how not to inflate the scanner cost without clinical benefits.

Methods: The analysis included Monte Carlo simulations (GATE), employed to model geometrical phantoms, as well as whole-body patient data of different weights from the Siemens Biograph Vision. The patient data were forward projected into a simulated ultra-long axial FOV scanner with a similar detector design as the Vision scanner. In addition, we included data from the PennPET Explorer as a reference to compare trends with our results. Data from the UPenn Explorer scanner spans a range of 70 cm (3 rings) to 140 cm (6 rings) axial length.

Results: We observed that the NEMA sensitivity almost doubled for an axial length of 100 cm when the acceptance angle increased from 20^o (equivalent Biograph Vision coverage - 26 cm axial FOV) to 52^o (full coverage). We also observed the same effect in the forward-projected patient data, in agreement with the first subjects acquired on the 3-ring PennPET Explorer. From the forward-projected patient data, we found that a time coincidence of 5 ns allowed for the collection of almost 100% of the true coincidences. Based on the PennPET Explorer data, axial spatial resolution degraded by about 10% when all oblique LORs were included [1].

Conclusions: An ultra-long axial FOV PET/CT scanner with 70-140 cm length would offer simultaneous coverage of most organs of interest with significant sensitivity increase compared to current clinical scanners. A 5 ns time coincidence window is sufficient in typical patient studies to collect all true coincidences, while limiting the amount of random coincidences. References: [1] Karp, J., Viswanath, V., Geagan, M.J., et al. PennPET Explorer: Design and Preliminary Performance of a Whole-body Imager. Journal of Nuclear Medicine. 2020 61:136-143.