7 The clinical utility of PET-CT in the mangement of infective endocarditis

Introduction The latest ESC guidelines on Infective Endocarditis (IE) included the use of PET-CT as an additional diagnostic criterion, particularly for prosthetic valve endocarditis, and several studies showed enhanced sensitivity for the modified Duke criteria. This audit was designed to assess if there was incremental clinical benefit of PET CT imaging in patients suspected of having IE. Methods A two year retrospective audit was performed on consecutive patients diagnosed with IE, or had a PET-CT for investigation for IE in a large teaching hospital. The modified Duke criteria were used to establish a likelihood of IE and final diagnosis of IE was made based on all clinical information on discharge. The PET-CT scans were classified as having uptake at the site of interest, showed metastatic spread from endocarditis or were negative for either. Results The audit identified 58 patients with a diagnosis IE or who had a PET-CT for investigation of IE. The mean patient age was 66 years and the in-hospital mortality was 14%. In all of these cases apart from 1 the final diagnosis was infective endocarditis, either clinically or pathologically (after valve replacement). Of these 11 patients (19%) had a PET-CT as part of their work up during their inpatient stay. Table 1 provides baseline demographics and clinical details of our cohort. For the patients that had a PET CT the mean age was younger at 55 years compared to 68 years (p value 0.003). As expected, the percentage of patients with prosthetic valves was also higher in the PET CT group at 55% versus 13% (P value 0.006). Otherwise there was no significant differences between the two groups. Table 2 outlines the results of the 11 patients who had a PET CT scan as part of their diagnostic work up with 7 patients (64%) having a prosthetic valve (all AVR). All patients had a TOE and this showed a vegetation in 6 patients (50%), with 4 patients also having an abscess. In 5 cases the scans showed increased FDG activity; 2 scans with peri-aortic valve uptake, 2 scans showing increased aortic mural uptake of unknown significance and one scan showing an embolic phenomenon. In 6 cases there was no increased FDG uptake in the mediastinum/emboli. However, in 5 of these cases the final diagnosis was IE, including 3 who had surgical valve replacement and a pathological diagnosis. There was one case that was a true negative result, resulting in the cessation of antibiotic treatment. Overall, there were 3 true positive results, 1 true negative and 7 false negative results. Conclusion In this study almost 20% of the cohort underwent PET CT scanning, often in the most clinically challenging and complex cases where the diagnosis is uncertain and guiding surgical referral. Of these only three results were positive for peri-valvular uptake or embolic spread. Two patients had increased uptake in the ascending aorta which was of unknown significance, but was not counted as a positive scan result. The results in our cohort were not as helpful as reported in other international series. Despite the fact that our numbers were small and used in highly selected cases, they suggest caution in the use of PET-CT in IE. This is particularly pertinent for negative scan results, as 3 of our patients with no cardiac uptake on PET-CT had IE diagnosed pathologically. This highlights the importance of including all the clinical and imaging information in the decision-making algorithm for IE.