Evaluation of PET standard uptake values dependence on imaging protocols for dual time point imaging

Dual time point FDG PET imaging is often used to differentiate malignant lesions from benign lesions and inflammatory processes. Standard uptake values, SUVs, for malignant lesions usually increase over the time, where SUVs over the time for benign or inflammatory lesions are likely to be stable or decreased. Very often, to save time, a second time point imaging is done only for a limited area (one or two bed positions vs. VVB), and to achieve higher resolution, specially for head and neck lesions, imaging protocol with finer pixel size is used. However, SUV accuracy and reproducibility is affected by selected acquisition and image reconstruction parameters. The aim of this study is to evaluate the effects of different PET imaging protocols on SUVs. The evaluation was done using NEMA/IEC PET Phantom with six hot spheres and cold central cylinder (5.1 cm dia), and PET Esser phantom with four hot cylinders and 3 cold cylinders filled with water, air and bone. Simulated lesion to background activity concentration ratios were: 2.7:1 for large spheres (37 and 28 mm dia.) and 25mm cylinder; 7:1 for smaller spheres (22, 17,13 and 10 mm dia.) and three cylinders (16, 12 and 8 mm dia.). The evaluation was done on the Siemens Biograph PET/CT LSO 3D system. All studies were acquired and processed using Abdomen WB, Head and Neck, and Brain clinical protocols, with single and multiple bed imaging. SUVs were calculated using the average and the maximum count density from ROIs placed in the center of the structure of interest The differences in average SUVs measured from abdominal and head & neck protocols ranged from 7% for larger spheres to 40% for smaller spheres, and from 30-99% for abdominal and head protocols. The differences in average SUVs measured from single bed acquisition and acquisition with multiple (overlapping) beds ranged from 2 to 10%, The large variations in the calculated SUVs as a function of selected imaging protocols, indicate that imaging protocol have to be closely matched in order to distinguishing between benign and malignant disease based on SUVs from dual time point imaging