FormalPara Key Summary Points

Fluorodeoxyglucose positron emission tomography computed tomography (FDG PET/CT) has high sensitivity and specificity in the diagnosis of prosthetic valve endocarditis.

FDG PET/CT has low sensitivity but high specificity for the diagnosis of native valve endocarditis.

FDG PET/CT has a promising role in the diagnosis of cardiac implantable electrical device and left ventricular assist device infections and in transcatheter aortic valve associated endocarditis.

Further studies are needed to evaluate the clinical impact of FDG PET/CT in cardiac infection.

Introduction

For many years, echocardiography was the exclusive imaging tool for evaluating impaired structural conditions of the heart, including infectious processes. Besides its clear advantages, echocardiography has limitations. The poor imaging quality due to echoic prosthetic devices, difficulties in demonstrating small processes, and inability to differentiate between infectious and non-infectious masses are major limitations. In addition, transesophageal echocardiography (TEE) is an invasive procedure, often withheld for elderly patients, who represent a growing population of patients with cardiac infections [1]. Fluorodeoxyglucose positron emission tomography computed tomography (FDG PET/CT) has the advantage of providing anatomic and functional information of such processes. In recent years, growing data have been published dealing with different aspects of FDG PET/CT in native and prosthetic valve infective endocarditis, cardiac implantable electrical device (CIED), and left ventricular assist device (LVAD) infections.

In this review, we discuss the available data evaluating FDG PET/CT in cardiac infections from recent years. We review the accumulating data of accuracy of FDG PET/CT in diagnosis of infective endocarditis (IE) as well as the added clinical value of this imaging tool and its impact on patient outcomes. We also review the utility of FDG PET/CT in growing groups of patients with specific conditions of cardiac infections such as transcatheter aortic valve implantation-IE (TAVI-IE), CIED, and LVAD infections. Also, we summarize the recommendations for myocardium preparation in cardiac infection imaging by FDG PET/CT and the current status of the utility of FDG PET/CT in cardiac infections according to guidelines from international professional societies.

A search was conducted in PubMed, google scholar, and articles’ references. The search term combined the terms fluorodeoxyglucose positron emission tomography, computed tomography, or FDG PET/CT with the terms infection, infectious disease, native and prosthetic infective endocarditis, cardiac implantable electrical device infection, or left ventricular assist device. We searched studies published in English and reviewed clinical studies other than case reports. The focus was on studies published since 2015. This article is based on previously conducted studies and does not contain any new studies with human participants or animals performed by the author.

Diagnostic Accuracy of FDG PET/CT in IE

First publications on FDG PET/CT in cardiac infections were case reports and small case series suggesting the potential role of FDG PET/CT in the diagnosis of IE [2]. Thereafter, It was followed by studies focusing on the yield of the FDG PET/CT in the diagnosis of extracardiac septic emboli [3, 4]. With the improvement in techniques of imaging of the heart and overcoming the physiological myocardial glucose uptake, more attention was given to the diagnosis of infectious processes involving the heart, including native valve IE (NVE) and prosthetic valve IE (PVE), CIED infection, and more recently, LVAD infection.

Most studies evaluating FDG PET/CT in cardiac infections were conducted in single centers leading to relatively small cohorts. The cohorts are heterogeneous as some included together all types of cardiac infections (NVE, PVE, CIED infection) [5,6,7], while others focused on a specific group of patients, mainly PVE [8,9,10]. Despite the differences in studies’ design, pre-scan protocol, and interpreting criteria used, accumulating data show relatively high sensitivity of FDG PET/CT in PVE (73–93%) compared with a low sensitivity in NVE (22–68%) [5,6,7, 11,12,13,14,15]. On the other hand, specificity is high in almost all studies for both NVE (97–100%) [7, 11, 14] and PVE (80–95%) [7,8,9,10].

Improving the Diagnostic Performance of FDG PET/CT

There are several factors that affect the diagnostic yield of FDG PET/CT in cardiac imaging. The majority of these factors are technical challenges related to physical and physiological issues such as the cardiac-pulmonary motions, sub-optimal spatial resolution, and long data acquisition times as well as the physiological myocardial glucose uptake. Technical improvements have overcome some of these factors in recent years. A recent meta-analysis that included 26 studies with 1358 patients evaluating the performance of FDG PET/CT in IE showed that in studies published after 2015, sensitivity was higher for all subtypes of IE compared with studies published before 2015 [12]. The increasing sensitivity over time seems to be highly related to technical improvement in FDG PET/CT machines, acquisition protocols, and pre-scan preparation for myocardial suppression techniques. In a prospective study that included 92 patients with suspected PVE with or without CIED infection, a higher sensitivity was observed when combining CT angiography (CTA) with FDG PET compared with non-enhanced CT with FDG PET, reaching a sensitivity of 91% [9]. In another study that evaluated FDG PET/CT in suspected PVE, a standardized semiquantitative measure of FDG uptake with a specific value improved the sensitivity to 100% without reducing the specificity [10]. In addition, excluding patients with low C-reactive protein or with a history of use of surgical adhesives for prosthetic valve implantation increased sensitivity and specificity. Other studies suggested different technical approaches for improving the accuracy of FDG PET/CT. For example, in a small study that included 41 patients with suspected PVE, improved specificity was achieved when FDG uptake was observed in both attenuation correction and non-attenuation correction PET images [16]. The authors emphasized the importance of the patient preparation and image interpretation methods on the FDG PET/CT yield. It is widely accepted that focal rather than diffuse, and heterogeneous rather than homogenous uptake are the main findings suggestive of true infection [17, 18]. Conversely, recent surgery and prolonged antibiotic duration before PET CT performance, commonly assumed to affect isotope imaging results, were rejected by several studies. One study that included 80 patients with suspected NVE, PVE, or CIED infection showed that there were no significant differences in the length of antibiotic therapy before FDG PET/CT performance between false negative and true positive results [5]. Similarly, in a previously mentioned study, recent valve surgery was not significantly associated with increased false positives [10].

Important factors that determine the performance of FDG PET/CT are the gold standard used for making the final diagnosis and the length of follow-up used for making the diagnosis. Some studies used the modified Duke criteria as the gold standard for diagnosis [7, 14], while others used a consensus by a multidisciplinary endocarditis team [5, 9, 10]. Alongside the advantage of the standardized definition of the Duke criteria, the low sensitivity of these criteria is a limitation [19]. On the other hand, consensus by a multidisciplinary team is a non-standardized and difficult to compare reference method. As the results of FDG PET/CT were known in real time, it was inevitably incorporated in the final diagnosis given by the endocarditis teams [5]. Length of follow-up for final diagnosis ranged from diagnosis on discharge of the index admission [7, 14] to 3 months [8], 4 months [16], 6 months [6], and up to 12 months [20].

Clinical Impact of FDG PET/CT on IE

Beyond the diagnostic value, it is of interest whether FDG PET/CT findings lead to changes in management and improved patients’ outcomes. This may be related to the detection of either cardiac or extracardiac findings by FDG PET/CT. In a recent large study that included 303 patients with both NVE and PVE/ascending aortic prosthesis infection (AAPI) who underwent FDG PET/CT, 76% (47/62) of PVE/AAPI cases who were initially classified as possible IE according to modified Duke criteria were reclassified as definite IE when adding FDG PET/CT results [7]. In this study, extracardiac infectious foci were identified in 47/129 (28%) patients. In a previous small study that focused on identification of extracardiac embolic events by FDG PET/CT in IE, treatment modification was reported in 14 (35%) of 40 cases with definite IE they included. The main modifications regarded antibiotic treatment prolongation, referral to surgical procedures, and even prevention of unnecessary device extraction. In another small cohort, FDG PET/CT led to treatment modification, due to diagnostic reclassification or extracardiac findings necessitating treatment, in five out of 20 patients with IE [21]. In another single-center retrospective study that reviewed data of 72 patients with suspected IE who underwent FDG PET/CT focusing on extracardiac findings, new findings detected first time by FDG PET/CT were reported in 17 (23.6%) out of 72 patients with IE, of which in 11 patients the findings were considered of clinical importance as they led to treatment adjustment [22]. The calculated number needed to perform FDG PET/CT to have a clinically significant finding in this setting with obvious selection bias was seven, the true number in non-selected cases of IE may be much higher. The overall detection rate of extracardiac metastatic foci of infection in IE in a meta-analysis that included 13 studies was 17% (34/198) [23]. The variability in detection of extracardiac infectious foci is related to type of IE, specific pathogen, and timing of performance of FDG PET/CT.

FDG PET/CT was evaluated for its prognostic value. In a prospective study that included 179 patients with suspected NVE and PVE who underwent FDG PET/CT, a significant association between positive FDG PET CT and adverse clinical outcome (death, unplanned heart surgery, and embolic events) was observed among patients with PVE [20]. Furthermore, the intensity of FDG valvular uptake was associated with new embolic events in both PVE and NVE [20].

FDG PET/CT in TAVI-IE

One of the specific conditions that poses a challenge on the sensitivity of the modified Duke criteria in general and on echocardiography specifically is TAVI-IE. The decreased sensitivity of echocardiography is a attributed to the location of vegetation and the acoustic shadow of the valve stent [24,25,26]. Recently, a few reports have suggested that FDG PET/CT in this setting can contribute to improved diagnosis [27, 28]. In a small cohort of 16 cases with suspected TAVI-IE, echocardiography showed findings compatible with endocarditis according to the modified Duke criteria in only five out of 10 definite IE cases, while FDG PET/CT was positive for IE in nine cases [27]. In another retrospective multicenter study that evaluated the change in diagnosis with adding FDG PET/CT and/or cardiac CTA in 30 patients with suspected TAVI-IE, the diagnosis had been changed (for both rejection and confirmation) in one-third of the patients [28]. Whether TAVI by itself may result in “normal” FDG uptake due to a non-infectious inflammatory reaction following the implantation procedure is a question that was targeted by a recent small study that compared the FDG uptake within 1 month following TAVI in 31 patients (control group) with 14 cases with suspected TAVI-IE [29]. FDG uptake was observed in 7 (22%) patients from the control group and in all seven cases with definite IE as well as in one case with rejected IE. Yet, the uptake representing true infection was focal and involved less than 25% of the valve circumference compared with wide uptake involving more than 50% of valve circumference in the control group and the rejected case of IE [29]. These findings and other open questions such as the duration of “physiological” uptake following TAVI and the prognostic value of FDG PET/CT in this setting need further investigation in larger studies.

FDG PET/CT in CIED Infection

The accuracy of FDG PET/CT for diagnosis of cardiac implantable electrical device (CIED) infections is variable and limited in comparison with IE. In general, studies included small cohorts. In a recent meta-analysis that included 14 studies with 492 patients overall, the pooled sensitivity and specificity were 83% and 89%, respectively. The results showed better diagnostic accuracy for pocket infection than lead infection [30]. In a more recent study that included 63 patients with suspected CIED infection, a very high specificity was reported. For lead infection, sensitivity and specificity of PET/CT for CIED-IE were 38.5% and 98.0%, respectively [31]. A large French multicenter study is planned, aiming to evaluate the accuracy of FDG PET/CT in the diagnosis of CIED infection (ENDOTEP), but its status is currently unknown [32, 33]. The true impact of using FDG PET/CT in cases with suspected CIED infection should be evaluated in order to identify the conditions where FDG PET/CT can make a difference in patient outcome.

FDG PET/CT in LVAD Infection

The use of FDG PET/CT in the diagnosis of left ventricular assist device (LVAD) infection has been evaluated in small cohorts. In a recent meta-analysis that included four studies with 119 FDG PET/CT scans performed in suspected LVAD infection, the pooled sensitivity and specificity were 92% and 83%, respectively [34]. These results mandate further validation in larger studies. In addition to its diagnostic yield, FDG PET/CT was associated with clinical outcome in a small retrospective study that included 35 patients with LVAD who underwent FDG PET/CT [35]. In 28 patients, the FDG PET/CT was performed for suspected infection. The extent of infection on the device as detected by PET/CT was associated with prognosis. Similar association was observed in another cohort of 57 patients, who underwent 85 FDG PET/CT scans, with a trend toward lower survival when FDG PET/CT showed involvement of all components of LVAD and thoracic lymph nodes [36]. If further validated in larger studies, prognostic stratification using FDG PET/CT should be included in the criteria for heart transplantation, where patients with widespread infection according to FDG PET/CT should be prioritized.

Myocardium Preparation Protocols

Quality of imaging of FDG PET/CT in cardiac infection depends on achieving optimal suppression of the physiological myocardial glucose uptake. FDG, as a glucose analogue, is normally taken up by the myocardial cells owing to their high metabolic activity. In order to restrict the FDG uptake in the heart to pathological processes as in inflammatory cells and bacteria, there is a need to reduce the normal glucose uptake. The latter can be achieved by diverting the myocardial metabolism to free fatty acid instead of glucose. Studies investigating FDG PET/CT in cardiac infections had applied different patient preparation protocols [12, 15, 23]. Recently, standardized recommendations by the European Association of Nuclear Medicine (EANM) and the European Association of Cardiovascular Imaging (EACVI) were established [17]. The main recommendations are a high-fat-enriched diet lacking carbohydrates for 12–24 h followed by a prolonged fasting period of 12–18 h, with or without the use of intravenously administered heparin of 50 IU/kg approximately 15 min prior to FDG injection [17].

FDG PET/CT in Guidelines of Cardiac Infections

International guidelines increasingly recommend FDG PET/CT in cardiac infections in recent years. The European Society of Cardiology (ESC) in 2015 suggested FDG PET/CT as an optional diagnostic test for possible or even rejected cases of PVE when a high clinical suspicion for IE exists [37]. The American Heart Association guidelines for infective endocarditis, endorsed by the Infectious Diseases Society of America (IDSA), published in 2015 as well, did not include FDG PET/CT among the recommended diagnostic imaging tests, awaiting further evidence [38]. The Heart Rhythm Society (HRS) guidelines from 2017 (endorsed by IDSA) recommend the use of FDG PET/CT when the diagnosis of CIED pocket or lead infection is doubtful (a weak recommendation based on limited data) [39]. The European Heart Rhythm Association (EHRA) guidelines from 2020, endorsed among others by the European Society of Clinical Microbiology and Infectious Diseases (ESCMID), recommend FDG PET/CT more strongly and include it as a major criterion for the diagnosis of CIED/IE. It is recommended strongly in suspected CIED-related IE with positive blood cultures and negative echocardiography; in all Staphylococcus aureus bacteremia among patients carrying a CIED; and for identification of unexpected embolic localizations and metastatic infections.

With the accumulating data, it seems that in the near future more infectious diseases societies will adopt FDG PET/CT as a diagnostic imaging modality in the diagnosis of cardiac infections. Differences may exist in the specific conditions and indications.

To summarize, FDG PET/CT is a good diagnostic tool for cardiac infections and may have added value in the following situations: (1) suspected PVE (including TAVI) that lack imaging proof (i.e., without typical findings in echocardiography). Negative TEE and FDG PET/CT highly exclude PVE. (2) In native valve with difficult to define echocardiographic findings when IE is in the differential diagnosis (due to the high specificity). Negative FDG PET/CT should not be used to exclude NVE. (3) Suspicion of clinically significant extracardiac septic foci in both NVE and PVE. (4) Suspicion of CIED infection with non-conclusive findings in echocardiography. (5) Suspicion of LVAD infection for confirming, localizing, and determining the extension of infection. Special attention should be given to maximize the patient pre-scan preparation for myocardial suppression of physiological FDG uptake. Focal and heterogeneous rather than diffuse and homogenous uptake are the main findings suggestive of true infection.

Additional research is still needed. Large multicenter prospective studies are crucial for validation of the current evidence of many of the issues discussed above. Beyond the diagnostic accuracy performance of FDG PE/CT in different IE/CIED infection, the question should be whether this modality usage can improve patient outcome. Therefore, studies should focus on the clinical impact of adding FDG PET/CT in the management of patients with suspected cardiac infection and the optimal timing for its performance. In addition, further evaluation of factors that affect the physiological myocardial suppression of FDG uptake and the cardiac-pulmonary motion-related disturbances in order to improve the current protocols is needed. Finally, could FDG PET/CT have a role in the follow-up of patients diagnosed with cardiac infection and potentially lead to individualized treatment decision-making?