Whole-body parametric PET/CT imaging of the total distribution volume using a new reversible delayed Logan model

The widely used reversible Logan Model (LM) requires complete dynamic tissue data to estimate the total distribution volume (Vt), which restricts Vt imaging to the PET scanner’s axial field-of-view (FOV). This study validates a novel reversible delayed Logan Model (DLM) for whole-body (WB) Vt imaging, requiring only late steady-state tissue data, to complement the irreversible Patlak model for WB parametric imaging. DLM was validated using data from 50 patients scanned with [18F]FDG (FDG) and 10 patients scanned with [68Ga]Ga-PSMA-11 (Ga-PSMA) using a 70-minute dynamic WB (D-WB) PET acquisition protocol on a short-axial FOV (SAFOV) PET/CT scanner. Organs in the chest region were segmented and analyzed using an image-derived input function (IDIF). For FDG, the use of a population-based input function (PBIF) was tested. Region-based kinetic analysis and voxel-based Vt imaging were generated with LM (0–70 min post injection (p.i.)) and DLM (50–70 min p.i.). For comparisons, region-based LM analysis was considered the reference. We found good agreement between Vt images generated by DLM and the reference region-based LM with organ-specific biases less than 12% and 25% for FDG and Ga-PSMA, respectively. Parametric images generated by LM had higher biases. The Vt images generated by LM and DLM showed similar characteristics, but DLM images had fewer artifacts and extended across the WB range. Furthermore, biases below 2% were observed when utilizing the PBIF, which allows for short, clinically feasible D-WB protocols. DLM allows quantitative WB Vt imaging using both SAFOV and long-axial FOV (LAFOV) PET systems. Clinically feasible D-WB PET examinations (20-min for SAFOV, and 10-min for LAFOV) with PBIF produce excellent Vt image quality.