Abstract
1436
Objectives: Aim to investigate the diagnostic value of 18F-fluorodeoxyglucose positron emission tomography /computed tomography (18F-FDG PET/CT) in detecting thoracic aortic prosthetic graft infection.
Methods: From October 2014 to November 2018, 24 patients (22 male; age 52±14 years) with suspected thoracic VPGI were retrospectively enrolled. 18F-FDG PET/CT scan was performed after open or endovascular thoracic aortic prosthetic reconstruction. Graft infection was defined as the presence of clinical or biochemical signs of graft infection with positive blood cultures or based on a combination of clinical, biochemical, and imaging parameters (other than PET scan data). All other grafts were deemed uninfected. Diagnostic efficiency of PET/CT for VPGI was assessed using a new 5-point (1-5) visual grading score (VGS score), values ≥3 was considered as positive graft infection, while score values ≤2 as negative. Maximum standard uptake value (SUVmax) and maximum tissue to background ratio (TBRmax) was calculated. Optimal SUVmax and TBRmax cut-offs identifying positive infected graft were determined by receiver-operating characteristic curve (ROC) anaysis. Differences between infected and uninfected thoracic aortic grafts were compared using t-test and chi-square test. In addition, 18 of 24 patients performed dual-time point 18F-FDG PET/CT imaging (DTPI) with delayed imaging in assessing thoracic aortic graft infection. Grading of image quality and SUVmax of the aortic graft on initial image (iSUVmax), delayed image (dSUVmax) and percentage of SUVmax change as retention index (RI) between initial and delayed images were recorded.
Results: 16 infected and 8 uninfected VPG patients were identified. By using VGS≥3, 21% of the uninfected and 79% of the infected grafts were judged as positive infected graft (P<0.01). The diagnostic sensitivity, specificity, and accuracy of VGS analysis for detecting VPGI was 94%, 50%, and 79%, respectively (P<0.05). Focal 18F-FDG uptake was noted in 25% of the uninfected and 81% of the infected grafts (P<0.05). Calculated optimal cut-offs of positive infected graft were: SUVmax=8.5, with 88% sensitivity/63% specificity vs TBRmax=5.3, with 81% sensitivity/ 88% specificity. Area under curve (AUC) of infected and uninfected grafts is 0.715 (0.496-0.878) and 0.859 (0.657-0.966) (P>0.05). The parameters of iSUVmax, dSUVmax, and RI did not differ between infected and noninfected grafts (P>0.05), but the delayed imaging did improve the PET image quality (P<0.05).
Conclusions: PET characteristics of focal uptake, high value of SUVmax and TBRmax by 18F-FDG PET/CT were major signs of VPGI. 18F-FDG PET/CT can be used to clarify the precise extent and location of infected graft and guide further therapy, it would be highly valued in patients with multiple prosthetic grafts implanted. Delayed imaging could improve the imaging quality, but it appears no improvement of the diagnostic efficacy.The findings warrant further prospective investigations.