The relevance of 18F-fluorodeoxyglucose positron emission tomography/computed tomography imaging in diagnosing prosthetic graft infections post cardiac and proximal thoracic aortic surgery

OBJECTIVES: Diagnosis of prosthetic graft infection after cardiac and proximal aortic surgery is a challenge. Besides technical considerations, redo surgery is associated with substantial morbidity and mortality. Therefore, an accurate diagnosis is mandatory. We report on our experience with hybrid 18-fluorodeoxyglucose positron emission tomography (18F-FDG PET)/computed tomography (CT) imaging, which is increasingly used to diagnose infections in the detection of graft infection after cardiac surgery. METHODS: Twenty-six patients who underwent 18F-FDG PET/CT imaging after cardiac surgery between February 2010 and September 2014 for suspected graft infection were retrospectively analysed (81% male, age 54.3 +/- 13.7 years). PET/CT imaging was performed 36.5 +/- 70.5 (0.5-300) months after surgery. 2 patients (8%) had undergone aortic valve replacement (concomitant ascending and proximal arch replacement in 1), 1 (4%) aortic root reconstruction, 9 (35%) aortic root replacement (concomitant partial arch in 4, arch replacement and postoperative TEVAR in 1), 2 (8%) ascending aortic and partial arch replacement and 2 (8%) ascending aortic replacement along with frozen elephant trunk. In 10 (38%), more than one previous cardiac surgical procedure had been performed. Maximum standardized uptake values (SUVmax) were obtained for all patients. If the patients were reoperated on, the final diagnosis was derived from intraoperative findings and/or microbiological results. Otherwise, the longest clinical follow-up available served as a reference. RESULTS: Conventional CT was positive for infection in 13 cases (50%). In 22 (85%), PET was indicative of infection (SUVmax 10.5 +/- 4.1). PET did not only confirm true-positive CT results in all but 1 case; in almost 30%, it provided substantial additional diagnostic information in comparison with CT alone. Receiver operating characteristic analysis identified an SUVmax of 7.25 to achieve maximum sensitivity (89%) and specificity (100%) in prediction of infection. Twelve patients (46%) required redo surgery for graft infection; in 1 additional patient (4%), sternal re-fixation was necessary. Furthermore, 2 patients had to be reoperated on for torn-out anastomosis and paraprosthetic perfusion (8%). CONCLUSIONS: PET provides functional data, confirms a CT diagnosis and may even increase diagnostic sensitivity in comparison with CT alone in selected cases. Specificity can be compromised by postoperative changes or chronic inflammatory reactions induced by the graft. CT and/or echocardiography should remain the first diagnostic step in case of a suspected infection because of their broad and fast availability. If confirmation is needed or diagnosis is not achievable using conventional methods, PET might be chosen as the next modality to gain additional information in experienced centres.