PET/CT with 18F-Fluordesoxyglucose in patients with suspected endovascular prosthesis infection

Infection of large vessel prostheses is a rare but critical complication. The aim of this work is to assess the impact of PET/CT with 18F-Fluordesoxyglucose (PET-FDG) on the diagnosis of infection in our environment.

Thirty-five patients (38 scans) were evaluated for suspected prosthetic infection. A qualitative analysis was performed taking into account the distribution of the radiopharmaceutical, categorizing the studies as positive or negative for infection. Those with focal or multifocal deposits along the vascular prosthesis were considered positive, and negative if a homogeneous and diffuse distribution over the whole prosthesis was observed, or a total absence of uptake. A semi-quantitative analysis was performed using SUVmax and average SUV values, and a metabolic index was calculated (SUVmax of the graft / average SUV of the normal vascular pool).

The PET-FDG study was positive in 20 patients, with a diagnostic accuracy of 84%. The 38 PET-FDG scans performed showed positive capture patterns (focal in 6, multifocal in 15, diffuse in 4) and negative pattern in the remaining 13. The sensitivity, specificity, positive and negative predictive values obtained for the PET-FDG were 95%, 89%, 90% and 94%, and for the AngioTC study 50%, 73%, 73% and 50%, respectively. The area values under the ROC curve were as follows: for the AngioTC 0.642 (not significant), and for the SUVmax values of 0.925 (p<0.005), average SUV of 0.922 (p<0.005) and for the metabolic index of 0.917 (p<0.005).

The PET-FDG proves to be a tool with high diagnostic accuracy in the infection of vascular prosthesis, both visual analysis according to patterns and semi-quantitative.