Abstract
Purpose: To early identify abnormal lesions by applying the (18)F-FDG PET dynamic modeling approach for discharged patients recovering from COVID-19. Methods: Seven discharged COVID-19 patients (COVID-19 group), twelve healthy volunteers (control group 1), and eight cancer patients with normal pulmonary function (control group 2) were prospectively enrolled. Control group 1 completed static (18)F-FDG PET/CT only; COVID-19 group and control group 2 completed 60-min dynamic (18)F-FDG PET/CT. Among COVID-19 group and control group 2, the uptake of FDG on the last frame (at 55-60 min) of dynamic scans was used for static analysis. Prior to performing scans, COVID-19 patients provided negative real-time Reverse Transcription-Polymerase Chain Reaction (rRT-PCR) of SARS-CoV-2, normal lung functions test, and normal laboratory test. Organ-to-liver standard uptake ratio (OLR, i.e. SUV(max) (evaluated organ)/ SUV(max) (liver)) from conventional static data and Patlak analysis based on the dynamic modeling to calculate the (18)F-FDG net uptake rate constant (K(i)) were performed. Results: Compared to the control groups, COVID-19 patients at two to three months after discharge still maintained significantly higher K(i) values in multiple organs (including lung, bone marrow, lymph nodes, myocardium and liver), although results for regular OLR measurements were normal for all discharged COVID-19 patients. Taking the image of lung as an example, the differences of SUV(max) images between COVID-19 group and control group were hard to distinguish. In contrast, a high (18)F-FDG signal of the lung among the COVID-19 group was observed for K(i) images. Conclusion: The K(i) from (18)F-FDG PET/CT dynamic imaging quantification might contribute to identifying residual lesions for COVID-19 survivors. Trial registration: The trial is registered with ClinicalTrials.gov, number NCT04519255 (IRB-approved number, K52-1).