The diagnostic value of [18F]FDG-PET/CT in detecting septic thrombosis in patients with central venous catheter-related Staphylococcus aureus bacteremia

Abstract

Septic thrombosis often complicates Staphylococcus aureus bacteremia (SAB) in patients with a central venous catheter. Currently there is no reference standard for diagnosis. We describe the diagnostic value of [18F]FDG-PET/CT imaging in a patient cohort and the potential contribution of quantitative measurements in detecting septic thrombosis. We selected patients with catheter-related SAB from our institutional database (2013–2020). The contribution of [18F]FDG-PET/CT on clinical diagnosis of septic thrombosis was evaluated. Standardized Uptake Values (SUV) were measured and compared with a composite reference standard (clinical signs, initial [18F]FDG-PET/CT result, Multidisciplinary Team (MDT) meeting outcome) to identify a cut-off value for detecting septic thrombosis. We identified 93 patients with a catheter-related SAB. Quantitative measurements were possible for 43/56 patients in whom a [18F]FDG-PET/CT scan was performed. Septic thrombosis was clinically diagnosed in 30% (13/43) of the cases. In 85% of these cases, significant [18F]FDG-PET/CT uptake at the site of the thrombus was the deciding factor for diagnosis of septic thrombosis during the MDT meeting. All mean SUV’s of thrombotic lesions were higher in patients with clinically proven septic thrombosis compared to patients in whom this diagnosis was rejected (p < 0.001). A SUVpeak thrombus/SUVmean blood ratio of 1.6 (AUC-ROC value 0.982) as cut-off to differentiate between septic thrombosis and non-septic thrombosis had a sensitivity of 92% (95% CI 64–100) and specificity of 89% (95% CI 65–99). An algorithm was designed to guide diagnosis of septic thrombosis. Quantitative [18F]FDG-PET/CT-derived parameters seem helpful to differentiate between septic and non-septic thrombosis in patients with catheter-related SAB.