By Michael H. Crawford, MD, Editor

SYNOPSIS: A multicenter, retrospective, observational study of patients undergoing 18F-fluorodeoxyglucose PET/CT for prosthetic valve endocarditis or other indications showed that if certain obvious confounders are excluded, this imaging technique offers a high degree of accuracy for diagnosing prosthetic valve endocarditis, especially if performed early in the course of the disease.

SOURCES: Swart LE, Gomes A, Scholtens AM, et al. Improving the diagnostic performance of 18F-fluorodeoxyglucose positron-emission tomography/computed tomography in prosthetic heart valve endocarditis. Circulation 2018;138:1412-1427.

Bittencourt MS, Blankstein R. More evidence supporting fluorodeoxyglucose position emission tomography for diagnosing prosthetic valve infective endocarditis. Circulation 2018;138:1428-1430.

The use of 18F-fluorodeoxyglucose (FDG) PET/CT imaging has shown promise for detecting prosthetic valve endocarditis (PVE), but several confounders such as the use of surgical adhesives or glue have reduced specificity. The use of quantitative parameters in addition to visual evaluation can overcome some of these limitations. The European Association of Nuclear Medicine Research Ltd (EARL) has proposed standardized calibration and reconstruction methods so that the results are interpreted in a uniform fashion between medical imaging centers.

The purpose of the Swart et al study was to evaluate the diagnostic performance of FDG PET/CT for identifying PVE in a cohort of patients in whom PVE was suspected compared to prosthetic valve patients undergoing FDG PET/CT for other reasons, using the EARL criteria. This retrospective, observational study took place across six centers in the Netherlands and included all patients who received a prosthetic valve and underwent FDG PET/CT between 2010 and 2016 and had at least a one-year follow-up. If there were multiple tests conducted for a patient with suspected PVE, only the first was included. The final diagnosis of PVE was made by expert consensus after one year of follow-up. There were 160 prosthetic valve patients who underwent FDG PET/CT for suspicion of PVE and 77 prosthetic valve patients with other indications for FDG PET/CT.

After one year of follow-up, 80 patients were diagnosed with PVE. Visual assessment for PVE was confounded significantly by the use of surgical adhesives or glue and low C-reactive protein (CRP) values due to prolonged antibiotic use, but not recent prosthetic valve implantation. When patients with significant confounders were excluded, the sensitivity and specificity for diagnosing PVE were 91% and 95%, respectively. When the quantitative EARL criteria were used in those without confounders, sensitivity and specificity were 100% and 91%, respectively. The authors concluded that both visual and quantitative FDG PET/CT imaging offers high diagnostic accuracy for PVE if conducted early and the use of surgical adhesives is considered.


PVE is notoriously difficult to diagnose. The standard transesophageal echo (TEE) and blood cultures approach has reduced sensitivity for diagnosing PVE compared to such usage in native valve endocarditis. However, the stakes are higher, as late diagnosis often leads to poor outcomes. The European Society of Cardiology (ESC) 2015 guidelines concluded that there were enough data on FDG PET/CT to include its use more than three months after valve replacement for suspected PVE.

The Swart et al study expands the database on FDG PET/CT use in suspected PVE, particularly regarding the effect of confounders that could reduce the accuracy of the test. First, they did not find that scans performed in the first three months after valve implantation were compromised by false-positive results. Thus, the authors could not support the ESC guidelines prohibition against FDG PET/CT use in the first three months after implantation. Other, more recent studies also support this conclusion. Second, Swart et al found that if FDG PET/CT was performed late in the course of PVE, when inflammation has quieted down (CRP < 40 mg/dL), that false-negative studies occurred more often. However, early use of FDG PET/CT in PVE cases with only valve vegetations and no annular inflammation can be negative since FDG PET/CT only identifies inflammation. Thus, Swart et al suggested that FDG PET/CT be used early in potential PVE cases, especially if the TEE is negative for vegetations. Third, the use of surgical adhesives for valve replacement created false-positive results. This knowledge is important for accurate interpretation of the scans. Fourth, Swart et al found that strict adherence to at least 24 hours of a low-carbohydrate diet and a 12-hour fast before the test to reduce myocardial uptake of FDG enhanced the interpretation of the scans. Finally, the authors showed that a quantitative approach comparing standardized uptake values at the valve with those of the blood pool in the descending aorta were highly sensitive and specific when patients with obvious confounders were eliminated. This new quantitative information using standardized protocols at all institutions is promising, but requires further evaluation. Nevertheless, it has been widely adopted in Europe.

There were limitations to this study. The most obvious is its retrospective, observational design; however, it is unlikely anyone will conduct a randomized, controlled trial that includes these critically ill patients. There probably was some selection bias in who underwent a PET/CT. Also, not all patients were suitable for adhering to a low-carbohydrate diet and a 12-hour fast. Even though myocardial uptake of FDG was not always suppressed, Swart et al did not find this to be a significant issue with scan interpretation. Strengths of the study included the use of two blinded readers and the inclusion of a non-PVE control group. Further, interobserver agreement was very high, but these were highly experienced nuclear medicine physicians. Considering the caveats, when PVE is suspected and the initial TEE is negative, it is reasonable to proceed to FDG PET/CT before the blood culture results return. PVE is a condition with high morbidity and mortality. Early diagnosis could alter the course of the disease. This should be proven in prospective outcome studies, but such investigations will be difficult to conduct.