Whole Body Dynamic Perfusion Imaging of Bone Tumors Using Ultra-fast Digital Photon Counting ¹⁸F-NaF PET/CT

Objectives: Next-generation digital photon counting PET/CT systems incorporate improved system sensitivity and time of flight timing resolution enabling ultra-fast PET acquisition times. Previous work has assessed the use of 9 second per bed acquisitions for clinical applications enabled by optimized reconstruction. We have since applied these short acquisition times to a whole body dynamic ¹⁸F-NaF protocol, where imaging from the time of injection covers multiple bed positions in rapid succession. This imaging may give new insight into tissue perfusion and uptake characteristics. We assessed the feasibility of quantification based on this whole body dynamic acquisition methodology using ¹⁸F-NaF bone imaging.

Methods: Twenty prostate cancer patients with suspicion of recurrence were imaged for this initial study. Patients were injected with an average dose of 5.2 mCi ¹⁸F-NaF. Imaging began at the time of injection and covered 6 bed positions, acquiring 10 seconds per bed position for 15 consecutive whole body sweeps, until approximately 30 minutes post-injection. All imaging was performed on a digital photon counting system (Philips Vereos). 3D regions of interest were placed over the a portion of the liver, blood pool, any present suspected bone tumor, areas of degenerative bone changes, and a region of healthy bone, recording SUV_mean and SUV_max for each of the 15 acquisitions.

Results: Whole body dynamic acquisitions were readily feasible, producing excellent image quality. Quantitative assessment provided detailed insight into tissue perfusion characteristics which was very robust across all patient acquisitions. In the liver, an initial perfusion peak was seen during the 1^st and 2^nd whole body sweeps with an average SUV_mean of 4.0 (range 2.6- 5.5) followed by a gradual decrease in activity thereafter. In the blood pool, the SUV_mean was high, averaging 18.6 (range 10.2 - 50.0) on the first acquisition as the bolus passed through in the vascular phase. The 2^nd acquisition showed much lower activity, averaging 5.7 decreasing gradually to 2.0 on the last acquisition. The healthy bone regions showed minimal uptake throughout the whole body dynamic acquisition time, the SUV_mean at about 30 minutes post-injection was on average 28% of the value measured on eventual static images at 75 minutes post-injection. In suspected bone tumors, the uptake curves were more variable, increasing gradually over the whole body dynamic phase, on average reaching 41% of the eventual SUV_max at 75 minutes post-injection.

Conclusions: Next-generation digital PET technology will allow for the development of new imaging protocols, such as ultra-fast imaging extended to dynamic whole body acquisitions. We demonstrated the feasibility of performing repeated, fast acquisitions over a wide axial field of view, as well as the quantitative robustness of such a protocol. Dynamic imaging over a longer field of view gives detailed insight into perfusion and uptake characteristics in a variety of tissues and anatomic regions Research Support: ODSA TECH 13-060