Vous êtes sur la page 1sur 11

Archives of Cardiovascular Disease (2013) 106, 52—62

Available online at



Imaging investigations in infective endocarditis:

Current approach and perspectives
Imagerie de l’endocardite infectieuse : approche actuelle et perspectives

Franck Thuny a,b,∗, Jean-Yves Gaubert c,d,

Alexis Jacquier c, Laetitia Tessonnier e,
Serge Cammilleri e, Didier Raoult c, Gilbert Habib a,b

Département de Cardiologie, Centre Hospitalier Universitaire Timone, Assistance
publique—Hôpitaux de Marseille, Aix-Marseille Université, 13005 Marseille, France
URMITE, CNRS-UMR 6236, Faculté de Médecine, Aix-Marseille Université, Marseille, France
Service de Radiologie Adultes, Centre Hospitalier Universitaire Timone, Assistance
publique—Hôpitaux de Marseille, Aix-Marseille Université, Marseille, France
Laboratoire de Radiologie Interventionnelle Expérimentale (LRIE) : EA 4264, Aix-Marseille
Université, Marseille, France
Service de Biophysique et de Médecine Nucléaire, Centre Hospitalier Universitaire Timone,
Assistance publique—Hôpitaux de Marseille, Aix-Marseille Université, Marseille, France

Received 11 July 2012; received in revised form 8 September 2012; accepted 17 September 2012
Available online 25 December 2012

KEYWORDS Summary Infective endocarditis is a serious disease that needs rapid diagnosis and accurate
Endocarditis; risk stratification to offer the best therapeutic strategy. Echocardiography plays a key role in
Imaging; the management of the disease but may be limited in some clinical situations. Moreover, this
Echocardiography; method is insensitive for very early detection of the infection and assessment of therapeutic
Computed response because it does not provide imaging at the molecular and cellular levels. Recently,
tomography; several novel morphological, molecular and hybrid imaging modalities have been investigated
Magnetic resonance in infective endocarditis and offer new perspectives for better management of the disease.
imaging; © 2012 Elsevier Masson SAS. All rights reserved.
Positron emission

Abbreviations: CIED, cardiovascular implantable electronic device; CT, computed tomography; FDG, fluorodeoxyglucose; IE, infective
endocarditis; MRI, magnetic resonance imaging; NPV, negative predictive value; PET, positron emission tomography; PPV, positive predictive
value; TEE, transoesophageal echocardiography; TTE, transthoracic echocardiography.
∗ Corresponding author. Département de Cardiologie, Hôpital de la Timone, Assistance publique—Hôpitaux de Marseille, Aix-Marseille

Université, boulevard Jean-Moulin, 13005 Marseille, France. Fax: +33 4 91 38 47 64.

E-mail address: franck.thuny@gmail.com (F. Thuny).

1875-2136/$ — see front matter © 2012 Elsevier Masson SAS. All rights reserved.
Imaging investigations in infective endocarditis 53

MOTS CLÉS Résumé L’endocardite infectieuse est une maladie grave qui nécessite un diagnostic rapide
Endocardite ; et une stratification du risque précise afin définir la meilleure stratégie thérapeutique.
Imagerie ; L’échocardiographie joue un rôle clé dans la prise en charge de la maladie mais peut présenter
Échocardiographie ; des limites dans certaines situations cliniques. De plus, cette méthode est peu sensible pour la
Tomodensitométrie ; détection très précoce de l’infection ainsi que pour l’évaluation de la réponse thérapeutique
Imagerie par car elle ne propose pas une imagerie au niveau cellulaire et moléculaire. Récemment, plusieurs
résonance nouvelles modalités d’imagerie morphologique, moléculaire et hybride ont été étudiées dans le
magnétique domaine de l’endocardite infectieuse et offrent de nouvelles perspectives pour une meilleure
nucléaire ; prise en charge de la maladie.
Tomographie par © 2012 Elsevier Masson SAS. Tous droits réservés.
émission de positons

Infective endocarditis: from Morphological imaging of endocardial

pathophysiology to imaging damage and its consequences
Infective endocarditis (IE) is a serious disease associated Echocardiography: the main imaging modality
with poor prognosis despite improvements in medical and
surgical therapies [1—4]. IE results in a complex pathogen- Echocardiography plays a key role not only in the diag-
esis that involves many host-pathogen interactions [5—7]. nosis of IE but also in the prognostic assessment and
Indeed, previous endocardial lesions can lead to the expo- follow-up under therapy and during surgery. Recently, rec-
sure of the underlying extracellular matrix proteins, local ommendations for the practice of echocardiography in
inflammation and then thrombus formation, which is termed IE have been published, to provide an update on the
‘non-bacterial vegetation’. In the case of bacteraemia, value and limitations of this technique in IE and to
valves with pre-existing sterile vegetations or tissues with define the optimal use of transthoracic echocardiogra-
minimal lesions can be colonized because of strong interac- phy (TTE) and transoesophageal echocardiography (TEE)
tions between the bacteria, platelets and endothelial cells [13].
via several bacterial surface proteins or plasma-bridging
molecules. This process leads to the recruitment of circu-
Echocardiography for the diagnosis of infective
lating inflammatory cells and the release of cytokines and
procoagulant factors, which contribute to the enlargement
of vegetations and the protection of bacterial pathogens Although IE may present with several very different initial
from host defences. Ultimately, valvular and perivalvu- symptoms, its diagnosis usually relies on the associa-
lar tissues are destroyed, thus increasing the risk of tion of clinical, microbiological and morphological criteria,
valve dysfunction, abscess formation and embolization [8]. which are included in the modified Duke classification
Moreover, in addition to embolic events, other extrac- [14]. By detecting several forms of endocardial damage,
ardiac life-threatening complications may occur, such as echocardiography remains an accurate method for pro-
infectious aneurysms and intracranial and visceral haemor- viding the major diagnostic criteria of IE. Knowledge of
rhages. the anatomical features of IE is fundamental in order to
Therefore, early and reliable diagnostic and risk strat- better understand, analyse and describe the echocardi-
ification strategies are critical to reduce delays to the ographic findings (Table 1). TTE is the initial technique
start of appropriate antimicrobial therapy and to identify of choice for investigation. A normal scan in low-risk
patients who require urgent valve surgery. As described in patients provides rapid non-invasive confirmation that the
recent international recommendations, echocardiography is diagnosis is unlikely. Moreover, TTE is better than TEE
a simple accurate method for detecting endocardial dam- for the detection of anterior cardiac abscesses and for
age in IE and helping in risk stratification [1,9]. However, haemodynamic assessment of valvular dysfunction. Because
echocardiography studies may be limited in some clinical of its higher sensitivity and specificity, TEE is recom-
situations [10—12] and this technique is insensitive for very mended [1,13] in cases of: negative TTE associated with
early detection of infection because it does not provide an high clinical suspicion; poor TTE quality; presence of pros-
imaging assessment of IE at the molecular and cellular lev- thetic valves or intracardiac device; and positive TTE
els. Recently, other morphological and molecular imaging (Fig. 1). In preliminary studies, three-dimensional TEE
strategies have emerged for the detection of endocardial provided incremental value over two-dimensional TEE in
involvement and extracardiac complications. its ability to accurately identify and localize vegetations
The aims of this review are to provide an update on the and to identify complications such as abscesses, per-
value of echocardiography in the management of IE, to dis- forations and ruptured chordae [15,16] (Fig. 2). Thus,
cuss the potential role of other imaging techniques and, echocardiography must be done rapidly and repeated once
finally, to consider the challenges and perspectives in the a week as soon as it is negative but the condition is
imaging investigations. suspected.
54 F. Thuny et al.

Table 1 Anatomical and echocardiographic definitions.

Surgery/necropsy Echocardiography
Vegetation Infected mass attached to an endocardial Oscillating or non-oscillating intracardiac mass
structure or on implanted intracardiac on valve or other endocardial structures, or on
material implanted intracardiac material
Abscess Perivalvular cavity with necrosis and Thickened, non-homogeneous perivalvular area
purulent material not communicating with with echodense or echolucent appearance
the cardiovascular lumen
Pseudoaneurysm Perivalvular cavity communicating with the Pulsatile perivalvular echo-free space, with
cardiovascular lumen colour Doppler flow detected
Perforation Interruption of endocardial tissue continuity Interruption of endocardial tissue continuity
traversed by colour Doppler flow
Fistula Communication between two neighbouring Colour Doppler communication between two
cavities through a perforation neighbouring cavities through a perforation
Valve aneurysm Saccular outpouching of valvular tissue Saccular bulging of valvular tissue
Dehiscence of a Dehiscence of the prosthesis Paravalvular regurgitation identified by
prosthetic valve TTE/TEE, with or without rocking motion of
the prosthesis
Adapted from Habib et al. [13] with permission.
TEE: transoesophageal echocardiography; TTE: transthoracic echocardiography.

and embolic events represent the three most frequent and

severe complications of IE. Echocardiography plays a key
role in the management of these complications by helping a
decision to be made regarding valve surgery and its optimal
Heart failure represents the main indication for valve
surgery in IE [19] and the operation is usually indicated in
an emerging (within 24 hours) or urgent (within a few days)
setting [1,9,20,21]. TEE allows identification of the mech-
anisms responsible for these complications, such as acute
valve regurgitation, valve obstruction or intracardiac fistula.
Moreover, TTE may provide criteria of poor haemodynamic
tolerance, such as torrential regurgitation, elevated left
and right filling pressures, pulmonary arterial pressures and
ventricular dysfunction. Even in the absence of clinical con-
gestive signs, the presence of these echocardiographic signs
suggests the need for valve surgery because the evolution
to heart failure usually is inevitable [1].
Perivalvular extensions are present in around 20% of cases
[2] and indicate valve surgery because they expose the
patient to risks of: heart failure by the occurrence of fistula
or prosthetic valve dehiscence; complete atrioventricular
Figure 1. Indications for echocardiography in the diagnosis and block by interruption of the cardiac conduction system; and
assessment of infective endocarditis. IE: infective endocarditis; persistence of the infection [22—24]. TEE is the technique
TTE: transthoracic echocardiography; TEE: transoesophageal echo- of choice for the diagnosis of perivalvular extension and
cardiography. its resulting complications but TTE seems better in case of
Adapted from Habib et al. [1] with permission. anterior abscess of the aortic annulus [25].
Embolic events are frequent and life-threatening
complications of IE, which are symptomatic in around
Echocardiography for risk stratification and 20—25% of cases [26—29] and silent (only detected by cere-
follow-up of infective endocarditis bral imaging) in almost 50% of cases [26,27]. These events
In addition to its role in diagnosing IE, echocardiography are factors for poor prognosis, especially in case of involve-
also has major prognostic value in IE in predicting death and ment of the cerebral circulation bed [29]. Echocardiography,
complications [17,18]. Heart failure, perivalvular extension especially TEE, is useful for the evaluation of embolic risk
Imaging investigations in infective endocarditis 55

Figure 2. Mitroaortic infective endocarditis explored by two-dimensional (2D) and three-dimensional (3D) transoesophageal echocardi-
ography (TEE). The 2D TEE (A) shows destruction of the cusps of the aortic valve with vegetations (white arrow) and perforation of the
basal portion of the anterior leaflet (red arrow), resulting in severe mitral regurgitation visualized by colour Doppler (B). The 3D TEE allows
better localization of the mitral perforation thanks to an atrial ‘en face’ view (C) and a left ventricle view (D).

at admission, by providing the maximal length, mobility and risk of systemic embolism compared with conventional ther-
location of vegetations. Indeed, large and highly mobile apy [34]. Although this result is of crucial importance, it
vegetations are associated with a higher risk of embolism was limited by the fact that it was obtained in a popu-
[17], especially in the mitral position [30]. This evaluation lation with a very low operative risk. Thus, we now have
must be performed very early in the course of IE because strong evidence that early surgery reduces embolic risk
the risk of embolic events remains high during the first but we need better risk stratification in order to eval-
week after diagnosis and initiation of antibiotics [17,31,32]. uate accurately the benefit-risk ratio of this procedure.
Therefore, valve surgery should be performed in an urgent Indeed, for a high embolic risk associated with a low or
setting when a large vegetation (> 10 mm) is present fol- intermediate predicted operative mortality (computed by
lowing one or more embolic episodes. In addition, when scoring systems), the benefit of early surgery would be
associated with other known predictors of a complicated greater.
course (heart failure, persistent infection under therapy, Intraoperative TEE is mandatory in patients operated on
abscess and prosthetic endocarditis), the presence of a large for IE; it provides the surgeon with a final anatomical evalua-
vegetation (> 10 mm) indicates an earlier surgical decision. tion of valvular and perivalvular damage, and is particularly
Finally, the decision to operate in the case of an isolated useful for assessing the immediate result of conservative
very large vegetation (> 15 mm) is more difficult and must be surgery, as well as in cases of complex perivalvular repair
specific for the individual patient. Surgery may be preferred [35]. Finally, echocardiography must be used for follow-
when a valve repair seems possible, particularly in mitral up of patients with IE under antibiotic therapy and after
valve IE [1,33]. Nevertheless, the prediction of embolism surgery, along with clinical follow-up. The number, type and
remains challenging and should take into account other timing of repeat examinations depend on clinical presen-
criteria, such as the type of microorganisms (Staphylococcus tation, type of microorganism and initial echocardiographic
aureus) and conditions associated with a prothrombogenic findings. After hospital discharge, the main complications
state (atrial fibrillation, diabetes, etc.). Recently, a ran- include recurrence of infection, heart failure, need for
domized trial demonstrated that early surgery in patients valve surgery and death. Thus, clinical and echocardiogra-
with large vegetations and significant valve dysfunction sig- phic periodic close follow-up (at 1, 3, 6 and 12 months) is
nificantly reduced the composite endpoint of death from mandatory during the first year after the end of antibiotic
any cause and embolic events by effectively decreasing the treatment [1].
56 F. Thuny et al.

Limitations of echocardiography in infective cases, the pseudoaneurysm was located close to the right
endocarditis coronary cusp, a location that is difficult to investigate by
TEE [38].
A negative echocardiogram may be observed in about 15% of
Thus, this imaging modality offers the possibility to
cases of IE. The imaging diagnosis may be particularly chal-
rapidly image the heart and other organs and thus to identify
lenging in some cases, such as with intracardiac devices,
both cardiac lesions and extracardiac complications, such
valvular prostheses, the presence of pre-existing severe
as embolic events, infectious aneurysms, haemorrhages and
lesions (mitral valve prolapse, degenerative lesions), very
septic metastases, which can modify the therapeutic strat-
small vegetations and abscesses or no vegetation. In addi-
egy (Fig. 3). Moreover, it provides an anatomical assessment
tion, the diagnosis may be difficult at the early stage of the
of the coronary bed, which is important in the preoperative
disease. Conversely, false diagnosis of IE may occur in other
evaluation [39]. CT seems to be especially useful in case of a
situations: for example, it may be difficult to differentiate
negative or inconclusive echocardiographic study. However,
between vegetations and thrombi, cusp prolapse, car-
contrast products should be used with caution in patients
diac tumours, myxomatous changes, Lambl’s excrescences
with renal failure or haemodynamic instability because of
or strands. Thus, in some situations, the echocardiogram
the risk of worsening renal impairment in combination with
remains negative or doubtful, even if it is performed by
antibiotic nephrotoxicity. In some cases, the indications for
expert hands and after a repeat examination [12,36]. Inno-
a CT scan might be limited to the brain and its arteries.
vations in the specialty of diagnostic strategy have emerged
Specific recommendations are needed to clearly define the
to resolve these issues through new imaging modalities. Mul-
appropriate situations where this modality should be used.
tislice computed tomography (CT) and magnetic resonance
imaging (MRI) might help to better identify both anatom-
ical intracardiac damages and extracardiac complications. Magnetic resonance imaging
Positron emission tomography (PET) and other molecular
Although multiple case reports on the use of MRI in identify-
imaging methods might provide imaging of the inflamma-
ing valvular and perivalvular damage in patients with IE have
tion and infection at molecular level and will be fused with
been published, this imaging modality seems to be of most
‘anatomical’ imaging.
value in the identification of silent cerebral complications
[26—28,41,42]. The role of cerebral MRI in the diagnosis
Multislice computed tomography and management of this disease has been defined in some
studies, which showed that systematic MRI could detect
Recent advances in the temporal and spatial resolution of subclinical cerebrovascular complications in about 50% of
multislice CT scanners allow high-resolution cardiac imag- patients [26—28].
ing. Currently, the major application of multislice CT is in In a single-centre study, Duval et al. described how the
the evaluation of coronary artery disease but it has been identification of brain damage by cerebral MRI modified
used also for heart valve disease, such as aortic stenosis their classification and management of 130 patients with
[37] and, more recently, in IE [38—40]. suspected or definite endocarditis. In this work, MRI iden-
In a small study of 37 consecutive patients with clini- tified cerebral lesions in 82% of cases. Solely on the basis
cally suspected IE, Feuchtner et al. found good results in of these MRI results, and excluding microhaemorrhages, the
detecting IE valvular and perivalvular damage using electro- diagnostic classification of 32% of the cases of non-definite
cardiogram (ECG)-gated 64-slice CT or dual-source CT. The endocarditis was upgraded to either definite or possible.
diagnostic performance of CT for the detection of evident Moreover, the therapeutic plans were modified for 18% of
abnormalities for IE compared with TEE was: sensitivity 97%, patients, including surgical plan modifications for 14% [27].
specificity 88%, positive predictive value (PPV) 97% and neg- The same investigators analysed the benefit of the addition
ative predictive value (NPV) 88% on a per-patient basis. In of abdominal MRI in these patients; they demonstrated that
a per valve-based analysis, the diagnostic accuracy for the both cerebral and abdominal MRI findings affected diagnosis,
detection of vegetations and abscesses/pseudoaneurysms but only cerebral MRI affected clinical management plans
compared with surgery was: sensitivity 96%, specificity [43].
97%, PPV 96%, NPV 97% and sensitivity 100%, specificity MRI offers the advantages of a non-ionizing imaging
100%, PPV 100%, NPV 100%, respectively, without signifi- modality but is limited by its lower spatial resolution and
cant differences compared with TEE. Although the small availability in comparison with current CT scans.
leaflet perforations were missed, CT provided more accurate
anatomical information regarding the perivalvular extent of Imaging of infectious aneurysms
abscesses/pseudoaneurysms than TEE [39]. Gahide et al.
found similar results in patients with aortic valve IE [40]. Infectious (mycotic) aneurysms result from septic arte-
As prosthetic valve IE represents one of the most diffi- rial embolism to the intraluminal space or vasa vasorum,
cult situations for echocardiographic studies, Fagman et al. or from subsequent spread of infection through the inti-
recently investigated the role of ECG-gated 64-slice CT in mal vessels. Because of the involvement of the muscular
the diagnosis of aortic prosthetic valve IE. In 27 patients, layer, infectious aneurysms are actually pseudoaneurysms
the authors showed that the strength of agreement between [44]. The intracranial arteries are the most frequent loca-
ECG-gated CT and TEE was good for abscess and dehis- tion, with a predilection for the distal branching points
cence, and moderate for vegetation. In comparison with of the middle cerebral artery. The reported frequency of
intraoperative findings, CT detected three additional pseu- infectious aneurysms of 2—4% is probably an underesti-
doaneurysms that were unnoticed by TEE. In two of these mation as some are clinically silent [45]. As the clinical
Imaging investigations in infective endocarditis 57

Figure 3. Cardiac and cerebral computed tomography (CT) scan in a patient with a mitroaortic infective endocarditis. This figure illustrates
that the CT scan allows rapid and accurate assessment of structural cardiac damage and extracardiac complications. The electrocardiogram-
gated cardiac CT scan (A) shows a calcified aortic valve with a large pseudoaneurysm just below the ostium of the left coronary artery (green
arrow) and an ‘aneurysm’ of the mitral valve anterior leaflet (red arrow). The cerebral CT scan (B) shows, after contrast injection, the
suspicion of a mycotic (infectious) aneurysm of the right middle cerebral artery (orange arrow), which was confirmed by three-dimensional
reconstruction of the angiography (C,D).

presentation is highly variable (focal neurological deficit, Imaging of actors involved in infective
headache, confusion, seizures) and the prognosis is poor, endocarditis pathophysiology
imaging should be performed to detect infectious aneurysms
in any case of IE with neurological symptoms [1]. CT or mag- Echocardiography, CT and MRI provide morphological imag-
netic resonance (MR) angiography may be used to detect ing without accurate information on the activity of IE. This
infectious aneurysms with similar sensitivity but no recent lack of functional data is a strong limitation, especially
large series has been performed to address this issue. A when the microbiological investigations are negative and
study showed that the sensitivities of CT scanning and MR in case of negative or doubtful structural lesions. More-
angiography were 94% and 86%, respectively, for the detec- over, these morphological imaging modalities are insensitive
tion of intracranial non-infectious aneurysms 5 mm or larger for very early diagnosis and accurate assessment of the
but only 57% and 35%, respectively, for aneurysms < 5 mm response to antibiotic therapy [47]. Hope comes from the
[46]. Thus, conventional angiography remains the gold imaging of actors involved in IE pathophysiology. The vegeta-
standard and should be performed when non-invasive tions are composed of microorganisms, inflammatory cells,
techniques are positive or negative with a remaining platelets and fibrin. Currently, as there are major difficul-
suspicion. ties in radiolabelling microorganisms at the site of infection,
58 F. Thuny et al.

Figure 4. Results of echocardiographic studies and 18 F-fluorodeoxyglucose positron emission tomography computed tomography (18 F-FDG
PET/CT) in a case of suspicion of aortic bioprosthetic valve infective endocarditis. The first transoesophageal echocardiography (TEE) (A)
showed a small doubtful thickening around the aortic bioprosthetic annulus (yellow arrow) in a patient with fever and negative blood
cultures. The second TEE (B), performed 8 days later, showed the development of a periprosthetic abscess (yellow arrow). 18 F-FDG PET/CT
performed the day after the first TEE showed hyperfixation around the aortic prosthesis (C). The patient underwent urgent valve surgery,
which confirmed the abscess. Of note, colonic hyperfixation was shown by 18 F-FDG PET/CT (green arrow), which allowed a polyp to be
revealed at the colonoscopy performed thereafter.

methods of conducting molecular imaging of the host’s reac- In a prospective study including 24 patients with definite
tion have emerged. These methods aimed to detect the IE, 18 F-FDG PET/CT detected a clinically occult embolism
inflammation and infection processes. Among them, 18 F- or metastatic infection in 28% of cases [50]. Three recent
fluorodeoxyglucose (18 F-FDG) PET/CT is the modality that works have investigated the role of 18 F-FDG PET/CT in CIED
has been most studied recently in the clinical setting. infections (Table 2).
Bensimhon et al. investigated the diagnostic value of
18 F-FDG PTE/CT in infective endocarditis 18
F-FDG PET/CT. The authors analysed 21 patients with sus-
F-FDG PET/CT is an imaging modality that allows mea- pected CIED infection and 14 controls free of infection. The
surement of metabolic activity within an organ, obtained final diagnosis was obtained either from bacteriological data
from the emission of positrons following disintegration of after device culture or by a 6-month follow-up according
the injected radioactive product. As the majority of the to modified Duke’s criteria. Sensitivity, specificity, PPV and
malignant cells have high glycolytic activity, this technique NPV were, respectively, 80%, 100%, 100% and 84.6% in the
is widely used for the diagnosis and follow-up of cancers. patient-based analysis; they were 100%, 100%, 100% and
Recently, it has been used for the identification of inflam- 100% for boxes but only 60%, 100%, 100% and 73% for leads.
matory and infectious processes because they also result None of the control patients was positive for CIED uptake
in significant FDG uptake by the inflammatory cells. Thus, [54].
F-FDG PET/CT has been used in the detection of vascu- In another study, Ploux et al. tested the role of 18 F-FDG
lar prosthetic infections [48,49]. Recently, numerous case PET/CT in the management of suspected CIED infections. By
reports and a few preliminary studies have shown much including 10 patients with a suspected CIED infection with
promise for this modality in the setting of IE. It was reported negative TEE signs, the authors showed that six patients had
in several cases that 18 F-FDG PET/CT was able to detect an increased FDG uptake. As a result of this finding, these
periprosthetic valve abscesses, while the first TTE and TEE patients subsequently underwent complete removal of the
studies were normal or doubtful (Fig. 4) [50—52]. This situa- implanted material. Cultures of the leads were positive in
tion represents almost 30% of cases [12]. Moreover, 18 F-FDG all six patients, confirming involvement of the leads in the
PET/CT is able to reveal the source of infection, which can infectious process. In the other four patients, the pacing
be a neoplasia (colonic cancer) in some cases. However, system was left in place without objective signs of active
for the time being, the role of 18 F-FDG PET/CT has been lead endocarditis during follow-up [55].
especially investigated in the detection of peripheral emboli Sarrazin et al. have made additional important contrib-
and the diagnosis of cardiovascular implantable electronic utions regarding the utility of 18 F-FDG PET/CT for suspected
device (CIED) infections [53]. CIED infection. In a study including 66 patients, they
Imaging investigations in infective endocarditis 59

Table 2 Diagnostic value of F-FDG PET/CT in cardiovascular implantable electronic device infections.
Study Material Sensitivity (%) Specificity (%)
Bensimhon et al. Pocket + leads 80 100
[54] Leads 60 100
Ploux et al. [55] Leads 100 93
Sarrazin et al. [56] Pocket + leads 89 86
CT: computed tomography; FDG: fluorodeoxyglucose; PET: positron emission tomography.

Table 3 Potential indications for imaging modalities other than echocardiography in infective endocarditis.
Imaging modality Potential indications
Cerebral CT scan In all patients (with or without neurological symptoms) at admission, to detect: cerebral
infarctions; cerebral haemorrhages; infectious aneurysms
Abdominal CT scan In patients with a difficult diagnostic situation or difficult surgical decision, and in
patients with uncontrolled infection, to detect: visceral infarctions; visceral abscesses;
infectious aneurysms
Cerebral MRI Same indications as cerebral CT scan. MRI is preferable when available because it
provides better sensitivity for the detection of small cerebral infarctions and
haemorrhages without ionizing radiation and injection of nephrotoxic contrast products
Cerebral angiography In patients with: cerebral haemorrhages; clinical suspicion of infectious aneurysm but
negative CT or MR angiography; presence of infectious aneurysm on CT scan or MR
PET/CT scan Suspicion of prosthetic valve or pacemaker/defibrillator leads endocarditis with negative
or doubtful echocardiography
CT: computed tomography; MR: magnetic resonance; MRI: magnetic resonance imaging; PET: positron emission tomography.

confirmed the relative good sensitivity (89%) and specificity performances when vascular prosthetic infection was sus-
(86%) of 18 F-FDG PET/CT interpreted by a qualitative visual pected [62], its value in IE should be demonstrated.
score. Moreover, they provided evidence that inflammation Better knowledge of the host/pathogen interaction offers
accompanying acute pocket surgery does not result in false new perspectives in the field of functional imaging of IE.
positive imaging, thus extending the applicability of this Indeed, by labelling other actors involved in this unique
technique to suspected early device infection. The authors interaction some investigators have developed novel imag-
also demonstrated the ability to distinguish deep pocket ing methods from animal models.
infection, which implies device infection and the necessity Because vegetations are platelet-fibrin clots in which
for device extraction, from superficial infection, which can platelet proaggregant activity is enhanced by bacterial
be treated with antibiotics alone [56]. colonization, Rouzet et al. investigated the ability of radio-
Although the preliminary results of 18 F-FDG PET/CT in labelled (Tc-99m) annexin V, to provide in vivo functional
the diagnosis and the management of IE are encouraging, imaging of platelet activation in an animal endocardi-
several questions remain. What is its value in native and tis model (rabbits and rats). Annexin V is a ligand of
prosthetic valve IE? Is it cost-effective? What is the impact phosphatidylserines exposed by activated platelets and
of prolonged prior antibiotic therapy? What is the impact of apoptotic cells. This experimental study showed that this
hyperglycaemia? More experience, especially with use of the technique can provide functional imaging of vegetations as
new ECG-gated PET/CT, is necessary before strong recom- well as embolic events [63].
mendations can be made regarding the use of this technique. In a mouse model, Panizzi et al. developed a new method
for imaging S. aureus IE. Because S. aureus is able to induce
blood coagulation via staphylocoagulase, the investigators
Other functional imaging modalities created a fluorescent prothrombin-based probe that could
Gallium-67-, indium-111- and technecium-99m-HMPAO- be activated in vivo by the enzyme. When injected into
labelled leucocyte scintigraphy is another option for imaging mice with IE, fluorescence molecular tomography combined
the infection that has been used in clinical practice [57—61]. with CT revealed the presence of vegetations with a high
Unlike 18 F-FDG PET/CT, this method is more specific for specificity and a high signal over background. Moreover, they
infection but is more time-consuming (24 hours) and offers generated a new prothrombin-based probe for PET-CT imag-
poor image quality. Although this modality showed good ing, which showed similar results. Finally, this innovative
60 F. Thuny et al.

Figure 5. Imaging perspectives for improvement of the management of infective endocarditis (IE). 2D: two-dimensional; 3D: three-
dimensional; CT: computed tomography; MRI: magnetic resonance imaging; PET: positron emission tomography; TEE: transoesophageal
echocardiography; TTE: transthoracic echocardiography.

method allowed visualization of the response to the antibi- Cardiology (ESC). Endorsed by the European Society of Clinical
otic treatment [64]. Microbiology and Infectious Diseases (ESCMID) and the Interna-
tional Society of Chemotherapy (ISC) for Infection and Cancer.
Eur Heart J 2009;30:2369—413.
Conclusions and perspectives [2] Selton-Suty C, Celard M, Le Moing V, et al. Preeminence
of Staphylococcus aureus in infective endocarditis: a 1-year
population-based survey. Clin Infect Dis 2012;54:1230—9.
Imaging of IE remains a diagnostic challenge because echo-
[3] Thuny F, Beurtheret S, Mancini J, et al. The timing of surgery
cardiography has several limitations, which can impact on influences mortality and morbidity in adults with severe com-
patient prognosis. Novel imaging modalities are emerging plicated infective endocarditis: a propensity analysis. Eur
and offer hope of better management of the disease and thus Heart J 2011;32:2027—33.
a reduction in mortality. Some of these methods provide a [4] Thuny F, Grisoli D, Collart F, et al. Management of
better morphological evaluation of the intra- and extracar- infective endocarditis: challenges and perspectives. Lancet
diac damage; others allow visualization of the inflammation 2012;379:965—75.
and infection at the molecular level. Although more stud- [5] Benoit M, Thuny F, Le Priol Y, et al. The transcriptional pro-
ies are necessary to clearly define the indications for each gramme of human heart valves reveals the natural history of
method, the new modalities should be included in the diag- infective endocarditis. PLoS One 2010;5:e8939.
[6] Moreillon P, Que YA, Bayer AS. Pathogenesis of streptococ-
nostic criteria for IE. The future in imaging of IE will be
cal and staphylococcal endocarditis. Infect Dis Clin North Am
multimodal (Table 3, Fig. 5). 2002;16:297—318.
[7] Que YA, Moreillon P. Infective endocarditis. Nat Rev Cardiol
Disclosure of interest [8] Thuny F, Habib G, Le Dolley Y, et al. Circulating matrix met-
alloproteinases in infective endocarditis: a possible marker of
The authors declare that they have no conflicts of interest the embolic risk. PLoS One 2011;6:e18830.
concerning this article. [9] Baddour LM, Wilson WR, Bayer AS, et al. Infective endo-
carditis: diagnosis, antimicrobial therapy, and management
of complications: a statement for healthcare professionals
References from the Committee on Rheumatic Fever, Endocarditis, and
Kawasaki Disease, Council on Cardiovascular Disease in the
Young, and the Councils on Clinical Cardiology, Stroke, and
[1] Habib G, Hoen B, Tornos P, et al. Guidelines on the prevention,
Cardiovascular Surgery and Anesthesia, American Heart Associ-
diagnosis, and treatment of infective endocarditis (new ver-
ation: endorsed by the Infectious Diseases Society of America.
sion 2009): the Task Force on the Prevention, Diagnosis, and
Circulation 2005;111:e394—434.
Treatment of Infective Endocarditis of the European Society of
Imaging investigations in infective endocarditis 61

[10] Evangelista A, Gonzalez-Alujas MT. Echocardiography in infec- during infective endocarditis: a prospective multicentre study.
tive endocarditis. Heart 2004;90:614—7. Eur Heart J 2007;28:1155—61.
[11] Habib G, Derumeaux G, Avierinos JF, et al. Value and limitations [30] Anderson DJ, Goldstein LB, Wilkinson WE, et al. Stroke loca-
of the Duke criteria for the diagnosis of infective endocarditis. tion, characterization, severity, and outcome in mitral vs aortic
J Am Coll Cardiol 1999;33:2023—9. valve endocarditis. Neurology 2003;61:1341—6.
[12] Hill EE, Herijgers P, Claus P, et al. Abscess in infective endo- [31] Dickerman SA, Abrutyn E, Barsic B, et al. The relationship
carditis: the value of transesophageal echocardiography and between the initiation of antimicrobial therapy and the inci-
outcome: a 5-year study. Am Heart J 2007;154:923—8. dence of stroke in infective endocarditis: an analysis from
[13] Habib G, Badano L, Tribouilloy C, et al. Recommendations for the ICE Prospective Cohort Study (ICE-PCS). Am Heart J
the practice of echocardiography in infective endocarditis. Eur 2007;154:1086—94.
J Echocardiogr 2010;11:202—19. [32] Steckelberg JM, Murphy JG, Ballard D, et al. Emboli in infective
[14] Li JS, Sexton DJ, Mick N, et al. Proposed modifications to the endocarditis: the prognostic value of echocardiography. Ann
Duke criteria for the diagnosis of infective endocarditis. Clin Intern Med 1991;114:635—40.
Infect Dis 2000;30:633—8. [33] Thuny F, Habib G. When should we operate on patients with
[15] Hansalia S, Biswas M, Dutta R, et al. The value of live/real acute infective endocarditis? Heart 2010;96:892—7.
time three-dimensional transesophageal echocardiography in [34] Kang DH, Kim YJ, Kim SH, et al. Early surgery versus con-
the assessment of valvular vegetations. Echocardiography ventional treatment for infective endocarditis. N Engl J Med
2009;26:1264—73. 2012;366:2466—73.
[16] Liu YW, Tsai WC, Lin CC, et al. Usefulness of real-time [35] Shapira Y, Weisenberg DE, Vaturi M, et al. The impact of
three-dimensional echocardiography for diagnosis of infective intraoperative transesophageal echocardiography in infective
endocarditis. Scand Cardiovasc J 2009;43:318—23. endocarditis. Isr Med Assoc J 2007;9:299—302.
[17] Thuny F, Di Salvo G, Belliard O, et al. Risk of embolism and [36] Vieira ML, Grinberg M, Pomerantzeff PM, et al. Repeated
death in infective endocarditis: prognostic value of echo- echocardiographic examinations of patients with suspected
cardiography: a prospective multicenter study. Circulation infective endocarditis. Heart 2004;90:1020—4.
2005;112:69—75. [37] Piazza N, Lange R, Martucci G, et al. Patient selection for
[18] Vilacosta I, Graupner C, San Roman JA, et al. Risk of transcatheter aortic valve implantation: patient risk pro-
embolization after institution of antibiotic therapy for infec- file and anatomical selection criteria. Arch Cardiovasc Dis
tive endocarditis. J Am Coll Cardiol 2002;39:1489—95. 2012;105:165—73.
[19] Tornos P, Iung B, Permanyer-Miralda G, et al. Infective endo- [38] Fagman E, Perrotta S, Bech-Hanssen O, et al. ECG-gated
carditis in Europe: lessons from the Euro heart survey. Heart computed tomography: a new role for patients with sus-
2005;91:571—5. pected aortic prosthetic valve endocarditis. Eur Radiol
[20] Kiefer T, Park L, Tribouilloy C, et al. Association between valvu- 2012;22:2407—14.
lar surgery and mortality among patients with infective endo- [39] Feuchtner GM, Stolzmann P, Dichtl W, et al. Multislice com-
carditis complicated by heart failure. JAMA 2011;306:2239—47. puted tomography in infective endocarditis: comparison with
[21] Vikram HR, Buenconsejo J, Hasbun R, et al. Impact of valve transesophageal echocardiography and intraoperative findings.
surgery on 6-month mortality in adults with complicated, left- J Am Coll Cardiol 2009;53:436—44.
sided native valve endocarditis: a propensity analysis. JAMA [40] Gahide G, Bommart S, Demaria R, et al. Preoperative evalu-
2003;290:3207—14. ation in aortic endocarditis: findings on cardiac CT. AJR Am J
[22] Anguera I, Miro JM, Evangelista A, et al. Periannular Roentgenol 2010;194:574—8.
complications in infective endocarditis involving native aortic [41] Klein I, Iung B, Labreuche J, et al. Cerebral microbleeds are
valves. Am J Cardiol 2006;98:1254—60. frequent in infective endocarditis: a case-control study. Stroke
[23] Anguera I, Miro JM, San Roman JA, et al. Periannular 2009;40:3461—5.
complications in infective endocarditis involving prosthetic [42] Klein I, Iung B, Wolff M, et al. Silent T2* cerebral microbleeds: a
aortic valves. Am J Cardiol 2006;98:1261—8. potential new imaging clue in infective endocarditis. Neurology
[24] Anguera I, Miro JM, Vilacosta I, et al. Aorto-cavitary fistulous 2007;68:2043.
tract formation in infective endocarditis: clinical and echocar- [43] Iung B, Klein I, Mourvillier B, et al. Respective effects of early
diographic features of 76 cases and risk factors for mortality. cerebral and abdominal magnetic resonance imaging on clini-
Eur Heart J 2005;26:288—97. cal decisions in infective endocarditis. Eur Heart J Cardiovasc
[25] Graupner C, Vilacosta I, SanRoman J, et al. Periannu- Imaging 2012;13:703—10.
lar extension of infective endocarditis. J Am Coll Cardiol [44] Peters PJ, Harrison T, Lennox JL. A dangerous dilemma:
2002;39:1204—11. management of infectious intracranial aneurysms complicating
[26] Cooper HA, Thompson EC, Laureno R, et al. Subclinical brain endocarditis. Lancet Infect Dis 2006;6:742—8.
embolization in left-sided infective endocarditis: results from [45] Corr P, Wright M, Handler LC. Endocarditis-related cerebral
the evaluation by MRI of the brains of patients with left-sided aneurysms: radiologic changes with treatment. AJNR Am J Neu-
intracardiac solid masses (EMBOLISM) pilot study. Circulation roradiol 1995;16:745—8.
2009;120:585—91. [46] White PM, Teasdale EM, Wardlaw JM, et al. Intracranial
[27] Duval X, Iung B, Klein I, et al. Effect of early cerebral magnetic aneurysms: CT angiography and MR angiography for detection
resonance imaging on clinical decisions in infective endocardi- prospective blinded comparison in a large patient cohort. Radi-
tis: a prospective study. Ann Intern Med 2010;152:497—504 ology 2001;219:739—49.
[W175]. [47] Alavi A, Saboury B, Basu S. Imaging the infected heart. Sci
[28] Snygg-Martin U, Gustafsson L, Rosengren L, et al. Cere- Transl Med 2011;3:99fs3.
brovascular complications in patients with left-sided infective [48] Merhej V, Cammilleri S, Piquet P, et al. Relevance of the
endocarditis are common: a prospective study using magnetic positron emission tomography in the diagnosis of vascular
resonance imaging and neurochemical brain damage markers. graft infection with Coxiella burnetii. Comp Immunol Microbiol
Clin Infect Dis 2008;47:23—30. Infect Dis 2012;35:45—9.
[29] Thuny F, Avierinos JF, Tribouilloy C, et al. Impact of cerebrovas- [49] Spacek M, Belohlavek O, Votrubova J, et al. Diagnostics
cular complications on mortality and neurologic outcome of ‘‘non-acute’’ vascular prosthesis infection using 18F-FDG
62 F. Thuny et al.

PET/CT: our experience with 96 prostheses. Eur J Nucl Med [57] Cerqueira MD, Jacobson AF, Matsuda M, et al. Indium-111
Mol Imaging 2009;36:850—8. leukocyte scintigraphic detection of mitral valve vegeta-
[50] Van Riet J, Hill EE, Gheysens O, et al. (18)F-FDG PET/CT tions in active bacterial endocarditis. Am J Cardiol 1989;64:
for early detection of embolism and metastatic infection in 1080—1.
patients with infective endocarditis. Eur J Nucl Med Mol Imag- [58] Erba PA, Conti U, Lazzeri E, et al. Added value of 99mTc-
ing 2010;37:1189—97. HMPAO-labeled leukocyte SPECT/CT in the characterization
[51] Vind SH, Hess S. Possible role of PET/CT in infective endocardi- and management of patients with infectious endocarditis. J
tis. J Nucl Cardiol 2010;17:516—9. Nucl Med 2012;53:1235—43.
[52] Yen RF, Chen YC, Wu YW, et al. Using 18-fluoro-2-deoxyglucose [59] Miller SW, Palmer EL, Dinsmore RE, et al. Gallium-67 and mag-
positron emission tomography in detecting infectious endo- netic resonance imaging in aortic root abscess. J Nucl Med
carditis/endoarteritis: a preliminary report. Acad Radiol 1987;28:1616—9.
2004;11:316—21. [60] Morguet AJ, Munz DL, Ivancevic V, et al. Immunoscintigraphy
[53] Bertagna F, Bisleri G, Motta F, et al. Possible role of F18-FDG using technetium-99m-labeled anti-NCA-95 antigranulocyte
PET/CT in the diagnosis of endocarditis: preliminary evidence antibodies as an adjunct to echocardiography in subacute
from a review of the literature. Int J Cardiovasc Imaging infective endocarditis. J Am Coll Cardiol 1994;23:1171—8.
2011;28:1417—25. [61] Pena FJ, Banzo I, Quirce R, et al. Ga-67 SPECT to detect endo-
[54] Bensimhon L, Lavergne T, Hugonnet F, et al. Whole body [(18) carditis after replacement of an aortic valve. Clin Nucl Med
F]fluorodeoxyglucose positron emission tomography imaging 2002;27:401—4.
for the diagnosis of pacemaker or implantable cardioverter [62] Liberatore M, Fiore V, Iurilli AP, et al. The role of preoperative
defibrillator infection: a preliminary prospective study. Clin Tc-99m HMPAO-labeled leukocyte total-body scans in aortic
Microbiol Infect 2011;17:836—44. prosthetic reconstruction. Clin Nucl Med 2001;26:1024—7.
[55] Ploux S, Riviere A, Amraoui S, et al. Positron emission tomo- [63] Rouzet F, Dominguez Hernandez M, Hervatin F, et al.
graphy in patients with suspected pacing system infections Technetium 99m-labeled annexin V scintigraphy of platelet
may play a critical role in difficult cases. Heart Rhythm activation in vegetations of experimental endocarditis. Circu-
2011;8:1478—81. lation 2008;117:781—9.
[56] Sarrazin JF, Philippon F, Tessier M, et al. Usefulness of fluorine- [64] Panizzi P, Nahrendorf M, Figueiredo JL, et al. In vivo
18 positron emission tomography/computed tomography for detection of Staphylococcus aureus endocarditis by targeting
identification of cardiovascular implantable electronic device pathogen-specific prothrombin activation. Nat Med 2011;17:
infections. J Am Coll Cardiol 2012;59:1616—25. 1142—6.