Reinforced physiology-informed learning for image completion from partial-frame dynamic PET imaging

Dynamic positron emission tomography(PET) imaging using ${}^{18}$F-FDG typically requires over an hour to acquire a complete time series of images. Therefore, reducing dynamic PET scan time is crucial for minimizing errors caused by patient movement and increasing the throughput of the imaging equipment. However, shortening the scanning time will lead to the loss of images in some frames, affecting the accuracy of PET parameter estimation. In this paper, we proposed a method that combined physiology-informed learning with time-implicit neural representations for kinetic modeling and missing-frame dynamic PET image completion. Based on the two-tissue compartment model, three types of constraint terms were constructed for network training, including data terms, boundary terms, and reinforced physiology residual terms. The method works effectively without the need for specific training datasets, making it feasible even with limited data. Three commonly used scanning schemes were defined to verify the feasibility of the proposed method and the performance was evaluated based on simulation data and real rat data. The best-performing scheme was selected for detailed analysis of PET images and parameter maps on datasets of four human organs obtained from Biograph Vision Quadra. Our method outperforms traditional nonlinear least squares (NLLS) fitting in both reconstruction quality and computational efficiency. The metrics calculated from different organs, such as the brain (SSIM $>$ 0.98) and the thorax (PSNR $>$ 40), show that the proposed network can achieve promising performance.