Technical evaluation of delayed-phase SUV estimation using four-phase whole-body dynamic ¹⁸F-FDG PET/CT: a phantom and clinical study

Four-phase whole-body dynamic PET/CT (4Dy) enables evaluation of temporal changes in FDG uptake. We evaluated SUV changes under 4Dy conditions using phantom data and a novel approach for estimating delayed-phase SUV directly from early four-phase SUV changes in pancreatic cancer without an additional delayed-phase acquisition. A NEMA PET Body Phantom was used to measure SUVmax and SUVpeak in six hot spheres (10-37 mm). The clinical study included 32 patients with pancreatic cancer. Linear, exponential, and logarithmic models were compared using early four-phase SUV data, and the best-fitting logarithmic model was used to estimate delayed-phase SUVmax and SUVpeak. Associations and agreement between measured and estimated values were evaluated using Pearson correlation and Bland-Altman analyses. Measured SUVmax and SUVpeak generally increased with sphere diameter and with phase progression. Clinically, both increased across the early 4Dy phases. The logarithmic model showed the highest coefficient of determination. For measured versus estimated delayed-phase SUV, R² was 0.67 for SUVmax and 0.93 for SUVpeak. Bland-Altman mean differences were - 0.42 for SUVmax and - 0.56 for SUVpeak, with narrower 95% limits of agreement for SUVpeak. Early four-phase 4Dy imaging allowed evaluation of temporal SUV changes under phantom and clinical conditions. Logarithmic approximation showed potential for delayed-phase SUV estimation, with SUVpeak showing higher correlation and better agreement than SUVmax. This approach may provide complementary delayed-phase SUV information while reducing the need for additional delayed-phase imaging, potentially shortening examination time and reducing patient burden.