Detection of bioprosthetic valve infection by image fusion of ¹⁸fluorodeoxyglucose-positron emission tomography and computed tomography

Abstract

A 63-year old male with prior bioprosthetic mitral valve replacement and coronary artery bypass graft surgery presented with dyspnea. C-reactive protein and white blood cells were elevated and serial blood cultures were negative. Transesophageal echocardiography showed a paravalvular leak and a thickened anterior leaflet of unclear either infective or degenerative origin. For differential diagnosis, cardiac 128-dual source computed tomography (CT) was performed. The CT image showed a thickened anterior leaflet and further revealed that the paravalvular leak was draining into a large wall thickened pseudoaneurysm with dense tissue adjacent suggestive for an abscess. Therefore, ¹⁸fluorodeoxyglucose-positron emission tomography (¹⁸FDG-PET) was appended and fused with the CT images. There was no tracer-uptake surrounding the leak excluding an abscess. However, an increased ¹⁸FDG-tracer uptake at the thickened anterior leaflet indicated active inflammation. During the subsequent cardiac surgery, vegetations were identified on the anterior cusp of the bioprosthetic valve. Intraoperative biopsy was taken and the cell culture was positive for Staphylococcus aureus. The pseudoaneurysm was repaired and the valve was replaced with a bioprosthesis. The patient was discharged uneventfully from hospital on day 12 and antibiotic treatment was continued for 4 weeks. In conclusion, our case indicates that ¹⁸FDG-PET with cardiac CT image fusion may be a useful tool in patients with unclear focus of inflammation and possible bioprosthesis infection.

CASE PRESENTATION

A 63-year old male with prior triple aorto-coronary bypass grafting and mitral valve replacement with a bioprosthesis (St. Jude Biocor™) 3 years ago presented with acute dyspnea. Blood tests revealed an elevated C-reactive protein (CRP) of 18.8 mg/dl and elevated white blood cells, but serial blood cultures were negative. He suffered from type 2 diabetes and had arterial hypertension.

Transesophageal echocardiography (TEE) showed a paravalvular leak and a thickened anterior leaflet (Fig. 1), suggesting either infection or bioprosthetic structural valve degeneration (SVD).

Figure 1:

TEE: the anterior leaflet of the bioprosthetic valve was diffusely thickened.

Cardiac electrocardiographic (ECG)-gated multislice computed tomography (CT) was performed for clarification, which showed non-specific thickening of the anterior leaflet (Fig. 2a). In addition, CT showed that the paravalvular leak drained into a large subannular pseudoaneurysm of 3 cm size (Fig. 2c). The three bypass grafts were patent. The aneurysm was wall thickened with dense tissue surrounding the leak, suggestive for an abscess.

Figure 2:

CT (a, c) and PET/CT image fusion (b, d): thickened anterior leaflet (a) showed increased ¹⁸FDG uptake indicating active inflammation (b). Note high normal right ventricular myocardial tracer uptake; paravalvular leak and pseudoaneurysm with irregular wall thickening (c) and a soft tissue lesion (white arrow). After PET-CT image fusion (d), no ¹⁸FDG-tracer uptake was found surrounding the pseudoaneurysm (white arrow) correlating with lack of active abscess intraoperatively (‘healed leak’).

To distinguish between active abscess and a chronic wall thickened aneurysm, we performed an ¹⁸fluorodeoxyglucose-positron emission tomography (¹⁸FDG-PET). There was no increased tracer uptake surrounding the leak (Fig. 2d, arrow). However, PET revealed an increased ¹⁸FDG uptake at the thickened anterior leaflet indicative of active inflammation (Fig. 2b).

Cardiac surgery was performed and vegetations were detected on the anterior cusp of the bioprosthetic valve. The pseudoaneurysm and the abscess were cleansed and a new bioprosthetic mitral valve was inserted (Carpentier-Edwards Magna, 29 mm). An intraoperative biopsy of the anterior mitral leaflet was taken and the cell culture was positive for Staphylococcus aureus.

The patient was transferred to the post-operative intensive care unit, and echocardiography on Day 7 showed normal prosthesis function. The patient was discharged from hospital on Day 12. A 6-month follow-up was uneventful and antibiotic treatment was continued for 4 weeks.

DISCUSSION

The differentiation of chronic structural bioprosthetic valve degeneration (SVD) and infection is a major challenge in the absence of a definite clinical diagnosis of infective endocarditis (IE). In our patient, despite a contradictory clinical presentation (CRP elevation and negative serial blood cultures), PET/CT image fusion could accurately detect active inflammation within a bioprosthetic valve by showing high ¹⁸FDG tracer uptake, and was useful to establish diagnosis and to indicate surgery. Furthermore, PET/CT image fusion was able to correctly identify a large paravalvular aneurysm (‘healed abscess cave’) without active inflammation, which was missed by echo.

Prosthetic valve endocarditis (PVE) represents about one-fifth of all definite cases of IE. Reoperation is required in approximately every second case. PVE is associated with poor long-term results (31–37% 10-year survival rate) independent of the patients’ treatment [1].

Imaging is the key to diagnose prosthetic valve dysfunction with TEE being the imaging modality of choice. In case of uncertain findings on TEE, alternative techniques may be helpful. Multislice CT has recently shown good results in detecting IE-associated valvular abnormalities [2].

PET/CT is a new modality which is increasingly used for diagnosis and staging of carcinomas and for the detection of inflammatory foci. FDG is accumulating in cells with increased metabolism, e.g. malignant or inflammatory. In contrast to a body PET/CT, which does not provide sufficient resolution for detailed imaging of a bioprosthesis, we performed a dedicated 128-dual source ECG-gated cardiac CT examination and fused the images to a cardiac PET on an external off-line workstation (Advantage™, GE, Healthcare).

In our case, the fusion of cardiac CT with ¹⁸FDG-PET was useful to detect biosprothesic valve infection which enforced the urgency of surgery indication.

Literature about the use of PET in infected valve prosthesis is scarce, and no case of a bioprosthetic mitral valve infection has been published yet.

Rohde et al. [3] reported a rare case of a recurrent infected aortic prosthesis identified on ¹⁸FDG-PET recently. Tegler et al. [4] confirmed IE and localized the lesion at an aortic prosthesis by PET/CT. Vind and Hess [5] reported two cases with negative TEE results for IE. Subsequent whole-body PET/CT identified PVE and ruled out extra cardiac involvement.

The 128-slice dual-source CT provides high spatial resolution of >0.3 mm, thus enables imaging of small lesions; however, the inferior spatial resolution of PET does not allow for identification of tracer uptake in small vegetations of <4 mm size. Similarly, Van Riet et al. [6] showed less effective PET/CT results based on small vegetations below the PET detection rate [4].

Our case implies the need of validation studies for this new imaging modality.

CONCLUSION

¹⁸FDG-PET/cardiac CT image fusion may be a useful tool for characterization of structural bioprosthetic valve abnormalities and for differentiation of active inflammation, chronic SVD and healed abscess caves.

Conflict of interest: none declared.