Improving Performance of FDG PET/CT for Prosthetic Valve Endocarditis
Study Questions:
What is the diagnostic accuracy of 18F-fluorodeoxyglucose (FDG) positron emission tomography/computed tomography (PET/CT) imaging for prosthetic valve endocarditis?
Methods:
The authors conducted a prospective multicenter study of patients with prosthetic valves referred for FDG PET/CT. All patients were fasted for a minimum of 6 hours, although some had also been on a low carbohydrate diet and/or treated with intravenous unfractionated heparin. Scans were visually graded for quality, and poor quality scans were excluded. Images were interpreted visually as well as quantitatively using standardized image reconstruction techniques qualified by the European Association of Nuclear Medicine Research. Additional scans of patients referred for non-infectious reasons were retrospectively identified and used as negative controls. Regression was used to identify causes associated with false-positive and false-negative scans.
Results:
A total of 160 patients with prosthetic valves and suspicion of endocarditis were prospectively recruited, and an additional 77 patients with prosthetic valves who underwent FDG PET/CT for non–endocarditis-related reasons were retrospectively included as negative controls. Visual assessment by an expert reader had a sensitivity of 74%, specificity of 91%, positive predictive value of 89%, and negative predictive value of 78%. By contrast, echocardiography had a sensitivity of 65% and specificity of 85%. Low systemic inflammatory activity, defined as C-reactive protein <40 mg/L, was associated with false-negative scans, while use of surgical adhesives was associated with false-positive scans. Recent valve implantation was not associated with false-positive scans. Excluding patients with these confounders led to a sensitivity of 91% and specificity of 95%. When standardized reconstructions are used and confounders are absent, a ratio of standardized uptake value in the blood to the valve and surrounding tissue of ≥2.0 had 100% sensitivity and 91% specificity.
Conclusions:
FDG PET/CT is a highly accurate tool for the identification of prosthetic valve endocarditis. Standardized, quantitative protocols may further improve accuracy.
Perspective:
FDG PET/CT is emerging as a highly useful tool for management of suspected endocarditis in patients with prosthetic valves. The added value of this study compared to the prior literature is: 1) larger size, 2) multicenter imaging suggesting generalizability, 3) standardization of imaging protocols, and 4) careful consideration of confounders that could limit diagnostic accuracy. The identification of surgical glue and low inflammatory activity as reasons for decreased diagnostic accuracy are important advances over prior literature that should be accounted for in clinical practice. Application of an algorithm developed from these data on clinical outcomes should be tested in a prospective randomized trial.
Keywords: Cardiac Surgical Procedures, Diagnostic Imaging, Echocardiography, Endocarditis, Fluorodeoxyglucose F18, Heart Valve Diseases, Heart Valve Prosthesis, Image Processing, Computer-Assisted, Inflammation, Nuclear Medicine, Positron-Emission Tomography, Tomography, Emission-Computed, Single-Photon
< Back to Listings