[18F] Sodium Fluoride Dose Reduction Enabled by Digital Photon Counting PET/CT for Evaluation of Osteoblastic Activity

Abstract

The aim of the study was to assess the quality and reproducibility of reducing the injected [¹⁸F] sodium fluoride ([¹⁸F]NaF) dose while maintaining diagnostic imaging quality in bone imaging in a preclinical skeletal model using digital photon counting PET (dPET) detector technology. Beagles (n = 9) were administered three different [¹⁸F]NaF doses: 111 MBq (n = 5), 20 MBq (n = 5), and 1.9 MBq (n = 9). Imaging started ≃45 min post-injection for ≃30 min total acquisition time. Images were reconstructed using Time-of-Flight, ultra-high definition (voxel size of 1 × 1 × 1 mm³), with 3 iterations and 3 subsets. Point spread function was modeled and Gaussian filtering was applied. Skeleton qualitative and quantitative molecular image assessment was performed. The overall diagnostic quality of all images scored excellent (61%) and acceptable (39%) by all the reviewers. [¹⁸F]NaF SUV_mean showed no statistically significant differences among the three doses in any of the region of interest assessed. This study demonstrated that a 60-fold [¹⁸F]NaF dose reduction was not significantly different from the highest dose, and it had not significant effect on overall image quality and quantitative accuracy. In the future, ultra-low dose [¹⁸F]NaF dPET/CT imaging may significantly decrease PET radiation exposure to preclinical subjects and personnel.

Keywords: PET/CT; [18F] sodium fluoride; bone imaging; bone metabolism; canine PET; digital photon counting; preclinical molecular imaging; sodium fluoride dose reduction.

Figure 1

Representative whole-body [¹⁸F]NaF PET maximum intensity projections (MIPs) REDCap surveys showing the same subject with standard (SD), low (LD), and ultra-low dose (ULD) from top to bottom. Subject scans were presented with three maximum intensity projections (MIPs), with 0°, 108°, and −90° angles (from left to right), with a gray level of 10.00. All images were reconstructed using ultra-high-definition (voxel volume = 1 × 1 × 1 mm³), 3 iterations, and 3 subsets. Point spread function (PSF) was modeled and Gaussian filtering was applied. SD and LD were retrospectively list-mode clipped accordingly to simulate the same count density as ULD.

Figure 2

Representative whole-body [¹⁸F]NaF PET/CT uptake fusion showing the metabolic bone activity. Standard dose [SD; (A)] and ultra-low dose [ULD; (B)] dorsal (left) and sagittal (right) images.