UNLABELLED: The clinical performance of the Duke Endocarditis Service criteria to establish the diagnosis of infectious endocarditis (IE) can be improved through functional imaging procedures such as radiolabeled leukocytes ((99m)Tc-hexamethylpropyleneamine oxime [HMPAO]-labeled white blood cells [WBC]). METHODS: We assessed the value of (99m)Tc-HMPAO-WBC scintigraphy including SPECT/CT acquisitions in a series of 131 consecutive patients with suspected IE. Patients with permanent cardiac devices were excluded. (99m)Tc-HMPAO-WBC scintigraphy results were correlated with transthoracic or transesophageal echocardiography, blood cultures, and the Duke criteria. RESULTS: Scintigraphy was true-positive in 46 of 51 and false-negative in 5 of 51 cases (90% sensitivity, 94% negative predictive value, and 100% specificity and positive predictive value). No false-positive results were found, even in patients with early IE evaluated within the first 2 mo from the surgical procedure. In 24 of 51 patients with IE, we also found extracardiac uptake, indicating septic embolism in 21 of 24. Despite the fact that septic embolism was found in 11 of 18 cases of Duke-definite IE, most of the added value from the (99m)Tc-HMPAO-WBC scan for decision making was seen in patients in whom the Duke criteria yielded possible IE. The scan was particularly valuable in patients with negative or difficult-to-interpret echocardiographic findings because it correctly classified 11 of 88 of these patients as having IE. Furthermore, 3 patients were falsely positive at echocardiography but correctly negative at (99m)Tc-HMPAO-WBC scintigraphy: these patients had marantic vegetations. CONCLUSION: Our results demonstrate the ability of (99m)Tc-HMPAO-WBC scintigraphy to reduce the rate of misdiagnosed cases of IE when combined with standard diagnostic tests in several situations: when clinical suspicion is high but echocardiographic findings are inconclusive; when there is a need for differential diagnosis between septic and sterile vegetations detected at echocardiography; when echocardiographic, laboratory, and clinical data are contradictory; and when valve involvement (especially of a prosthetic valve) needs to be excluded during febrile episodes, sepsis, or postsurgical infections.