Objective: Radionuclide imaging with fluorodeoxyglucose (FDG) and positron emission tomography (PET) has been proposed for the identification of vascular graft infection; however, its accuracy has not been determined. We performed this prospective study to compare the usefulness of FDG-PET in the assessment of vascular graft infection relative to computed tomography (CT).
Methods: FDG-PET was performed for 33 consecutive patients with a suspected arterial prosthetic graft infection. The PET images were then assessed visually in terms of the density of uptake. In cases with positive uptake, the pattern of accumulation was also defined, such as focal or diffuse uptake. We compared the diagnostic efficiency of PET with contemporaneous CT in detection of infection of the arterial prosthetic graft.
Results: On the basis of the surgical, microbiological, and clinical follow-up findings, the aortic grafts were considered infected in 11 patients and not infected in 22 patients. Although the sensitivity of PET (91%) was higher than that of CT (64%), its specificity (64%) was lower than that of CT (86%). When focal uptake was set as the positive criterion in FDG, the specificity and positive predictive value of PET for the diagnosis of aortic graft infection improved significantly to 95% (P < .05 for both).
Conclusions: Although both techniques are useful in evaluation of patients with suspected aortic graft infection, using the characteristic FDG uptake pattern described previously as a diagnostic criterion made the efficacy of FDG superior to that of CT in the diagnostic assessment of patients with suspected aortic graft infection.