@article {Yao1758, author = {Shulin Yao and Baixuan Xu and Zheng Qian and Yun Dong and Hongdi Li and Wentao Zhu}, title = {A dynamic protocol for Patlak analysis with clinically practical patient management}, volume = {59}, number = {supplement 1}, pages = {1758--1758}, year = {2018}, publisher = {Society of Nuclear Medicine}, abstract = {1758Objectives: A major burden for dynamic analysis is the excessive long scan duration. Concerning patient comfortableness, a clinically practical protocol should minimize the dynamic scan duration. We proposed a protocol which allows patients to leave the bed after a short PET-only scan, and undergo a second dynamic PET/CT scan later. This protocol only requires patients to remain rigid for ~40min instead of \>80min in routine protocols, and yields consistent diagnostic information. Methods: several patients underwent 0~80min 18F-FDG PET/CT scans, where the 40~80min frame after steady state was used to compute the Patlak parameters. The difference between the proposed and "gold standard" protocol was that the former was more patient friendly and had much shorter scan time. For the "gold standard", input function was derived from reconstructed frames since T=0. For the proposed protocol, only a short frame (~5min) was collected before T=40min and the patients could indeed leave the bed afterwards. MLAA was used to reconstruct this frame, after which the image-derived blood sample was fed to a population-based method to fit the input function for T=0~40min. Finally the Patlak parameters for the proposed method was estimated and the deviation from the "gold standard" was measured. Results: \> the mean estimated integrals of blood input function for T=0~40min yielded a 6.2\% relative bias from respective ground truth, due to the mixed error from MLAA and population based fitting. The 5min acquisition was of merit as it reduced the integral error by 3\% approximately compared with not acquiring the frame. One ROI were drawn in the left ventricle and another one drawn in the liver for each patient. The final Patlak slope images yielded mean relative difference 0.5\% and 3.4\% respectively, which were much less significant compared with error of the integral of input function and were both below 5\%. Conclusions: we provided a clinically practical protocol to perform Patlak analysis. Even with not-so-trivial error in the blood input function for T=0~40min, the error propagated into the final Patlak parameters was non-significant. The proposed protocol may be used in clinic in a routine manner to reduce total scan time, improve patient management, and maintain quantitative accuracy.}, issn = {0161-5505}, URL = {https://jnm.snmjournals.org/content/59/supplement_1/1758}, eprint = {https://jnm.snmjournals.org/content}, journal = {Journal of Nuclear Medicine} }