Accuracy of PET/CT in the Diagnosis of Vascular Graft Infections

Abstract

Introduction - Vascular graft infections (VGI) have an incidence of 1 to 6 %. Positron emission tomography/computed tomography with 18F-fluorodeoxyglucose (FDG-PET/CT) has emerged as a new tool to replace nuclear scintigraphy. However, its clinical reliability is still unknown. The aim of this study was to evaluate the diagnostic accuracy of FDG-PET/CT in a population with suspected VGI and to validate the quantitative value of diagnostic parameters. Methods - This is a retrospective cohort study. FDG-PET/CT was performed in 37 patients with suspected VGI between 2013 and 2017. A reliable culture was obtained in 25 of them and were finally included in the study. VGI was considered proven only in case of a positive culture (blood, peri-graft collections or graft samples. Diagnostic accuracy was assessed comparing culture results (gold standard) with FDG-PET/CT values. We assessed maximal standardized uptake value (SUVmax), tissue-to-background ratio (TBR), and visual parameters (FDG distribution patterns and visual grading scale) in the final confirmation of the diagnosis of AGI. Receiver operating characteristics (ROC) curves were used to assess the ability of SUVmax and TBR to identify the presence and absence of VGI. Results - The mean age was 64.32 ± 16.75 years, 72% were males. Cultures were positive in 20/25 patients. FDG-PET/CT was considered positive in 19/25 patients. Surgery was performed in 13/25 patients. Mortality was 24% (6/25 patients). The most frequent microorganism was Staphylococcus aureus (9/13 patients). All these patients were followed with long-term antibiotic treatment until negativization of tests including FDG-PET / CT. Mean SUVmax was 6.4 ± 3.41 and TBR 2.14 ± 0.5 in all cases. For positives cultures the mean SUVmax was 7.34 ± 3.3 with a TBR 3.73 ±1.7. In the ROC analysis, the area under the curve for SUVmax was 0.56 (95% CI: 0.30-0.84) and for TBR was 0.53 (95% CI:0.27-0.78). A SUVmax cut off point of 4.5 yielded a 91,7% Sensitivity (Sens) and 50% Specificity (Spec). A TBR cut off point of 2.9 yielded 58% Sens and a 60% Spec with a positive predictive value of 73.68% and negative predictive value of 83.33%. Conclusion - FDG-PET/CT may be a useful tool in the diagnosis and surveillance of VGI. Nevertheless, the accuracy of quantitative parameters is jeopardized by the uncertainty of gold standard based on microbiology data. Further and larger registers with standardized measurements are needed to establish a more reliably parameters, necessary to take difficult decisions in this kind of patients. References-Sah BR, Husmann L, Mayer D, Scherrer A, Rancic Z, Puippe G, Weber R, Hasse B; VASGRA Cohort. Diagnostic performance of 18F-FDG-PET/CT in vascular graft infections. Eur J Vasc Endovasc Surg. 2015 Apr;49(4):455-64.-Saleem BR, Berger P, Vaartjes I, de Keizer B, Vonken EJ, Slart RH, de Borst GJ, Zeebregts CJ. Modest utility of quantitative measures in (18)F-fluorodeoxyglucose positron emission tomography scanning for the diagnosis of aortic prosthetic graft infection. J Vasc Surg. 2015 Apr;61(4):965-71.