Textural features of 18F-fluorodeoxyglucose positron emission tomography scanning in diagnosing aortic prosthetic graft infection

1626 Objectives The clinical dilemma in suspected aortic graft infection (AGI) is how to obtain non-invasively a reliable proof of infection. 18F-fluorodeoxyglucose positron emission tomography scanning (18F-FDG PET) has been suggested to have a pivotal role in the detection of AGI, however visual assessment is not convincing. In addition to confirm the presence of infection, it is necessary to obtain information regarding the pattern of infection in order to select a proper strategy for reoperation. Therefore, developing a more reliable, non-invasive physiological approach to detect infected prostheses is required. We sought to investigate whether textural features can have a predictive value in diagnosing AGI. Methods A total of 30 patients with a history of aortic graft reconstruction and an 18F-FDG PET scan were retrospectively included from the database of the University Medical Center Groningen (UMCG). Patients were clinically suspected for AGI in case of undefined fever; a deep wound infection, an incision fistula or persisting high laboratory infection parameters. Sixteen patients where considered to have an AGI (group I). Positive cultures were found in 10 of these 16 patients during treatment (group Ia). In the other 6 patients, no positive microbiology cultures were found during follow up (group Ib). A control group was constructed out of 14 patients who underwent 18F-FDG PET for other reasons than suspected for infection (group II). We extracted 65 different textural features. 18F-FDG PET based textural features to characterize heterogeneity of 18F-FDG uptakes in the aortic prosthetic graft were measured for each patient. The volume of interest (VOI) was manually delineated on axial planes of the low-dose CT for attenuation correction to enclose three-dimensional coverage of the entire suspected prosthetic graft. An experienced nuclear medicine physician assessed the 18F-FDG PET images including the following features; maximal standardized uptake value (SUVmax), tissue-to-background ratio (TBR), FDG distribution patterns and Visual Grading Scale (VGS). Results Sixteen patients were suspected for AGI. Microbiology was obtained through percutaneous puncture in 9 of these 16 patients, and through surgery in the other 7 patients. Significant differences were found in the SUVmax between groups I and II (P Conclusions Textural features for the assessment of heterogeneity can play an important role in diagnosing AGI, but still needs validation and further refinement in changing clinical decision-making. Further research is needed to improve the texture features in diagnosing AGI.