Imaging the Infected Heart

BACTERIAL ENDOCARDITIS Imaging the Infected Heart Abass Alavi,1* Babak Saboury,1 Sandip Basu2 This Focus discusses the merits of modern imaging techniques for the management of patients with suspected or proven infection and also addresses the challenges of detecting infective endocarditis early. THE EVOLUTION OF IMAGING The current generation of structural imaging techniques, including magnetic resonance imaging (MRI) and computed tomography (CT), provides anatomical scans with exquisite detail and high spatial resolution. However, many diseases start at the molecular and cellular levels, which may never translate to gross structural abnormalities. These technologies have proven to be insensitive for early detection of several diseases, including cancer, when therapeutic intervention would be desirable. In addition, because of the low sensitivity of structural imaging methods, the effects of systemic therapy cannot be adequately assessed, which is pivotal to clinical decisionmaking. In medicine, it is not uncommon to encounter suboptimal or no response to treatment, particularly for infectious diseases. As such, undue delays in using alternate therapies may result in further progression of the disease as well as undesirable side effects from the initial treatment. There is thus a dire need for imaging approaches that detect disease at the molecular and cellular levels during the early stages of pathogenesis. In the 1970s, investigators noticed that the agent 18F-fluorodeoxyglucose (FDG) was able to measure glucose metabolism in vivo quantitatively and in a dynamic manner, thus opening a new era in medical imaging at the molecular level (1). By that time, positron emission tomography (PET) had also emerged as a promising modality for imaging the biodistribution of labeled compounds in the clinic. Ever since, FDG– PET has been the workhorse for imaging glucose metabolism and has played a major role in examining disorders that are associ1 Radiology Department, Division of Nuclear Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA 19104, USA. 2Radiation Medicine Centre (BARC), Tata Memorial Hospital Annexe, Parel, Mumbai, 400012, India. *Corresponding author. E-mail: Abass.Alavi@uphs. upenn.edu ated with altered glycolysis, such as central nervous system disorders, cancer, and inflammation. We discuss in detail in this Focus recent efforts to image infection and inflammation, including recent papers on detecting acute infective endocarditis with advanced imaging methods, such as PET. VISUALIZING INFECTION AND INFLAMMATION In the 1930s, Warburg discovered increased glycolysis in cancer cells in vitro. It was noted thereafter that inflammatory cells also have high glycolytic activity that is similar to that of malignant cells (2–4). In recent years, the range of disorders with aberrant glycolysis that can be assessed by means of FDG-PET has increased and comprises common infections (such as an infected prosthesis, osteomyelitis, or a diabetic foot) and noninfectious inflammatory disorders, such as rheumatoid arthritis, regional ileitis, sarcoidosis, and atherosclerosis (4). FDG as a unique tracer has been used extensively for identification of infected sites in the human body and for monitoring response to treatment (3, 4). Two types of biological structures at the site of infection can be targeted with external imaging. One is microorganisms, such as bacteria; the other is inflammatory cells that home to the infected site. Although either of these could potentially serve as a reliable source for targeted imaging, there are major differences between the two with regard to modern imaging techniques. For example, the volume of microorganisms that reside at these sites is extremely small and provides limited options for detecting the infected areas (because few binding sites are available). Conversely, imaging inflammatory cells has proven to be relatively effective, as shown by a variety of approaches, in particular functional imaging techniques that use radiotracers. Nevertheless, detecting inflammatory cells is nonspecific and provides indirect evidence at best for local or diffuse infection. In other words, simply visualizing the presence of immune cells cannot differentiate between inflammation caused by microorganisms or by noninfectious diseases. Efforts to radiolabel bacteria at the sites of infection have yielded minimal success, and most have not been translated into the clinic. This approach was adopted using single gamma-emitting radionuclides attached to bacteria-targeting compounds, including antibiotics (5). In recent years, efforts have been made to use positronemitting radiotracers instead because of favorable physicochemical characteristics. With the radiolabled tracers, it is clear that positive results, which were reported by this approach, mostly reflected nonspecific leakage of the labeled agents at the sites of infection owing to the presence of a large number of leaky vessels. In other words, similar and positive results would be expected to be noted from inert preparations with no known attraction to the site, such as radiolabeled albumin. Therefore, there is some consensus that compounds that target bacteria and other microorganisms may not be promising enough to be pursued further at this time. There are two possible options for visualizing inflammatory cells at the infected sites. One method is to label a patient’s white cells ex vivo, reinfuse them intravenously, and monitor cell migration to the infected lesions by using conventional scintillation cameras. Unfortunately, there are serious shortcomings to this approach. The procedure is very time-consuming (3 to 5 hours for labeling, 24 hours for imaging), which results in many nonfunctional cells. In addition, the image quality is very poor (nontomographic), and the radiation dose to the sensitive organs is unacceptably high. The other option is to label white cells with positron-emitting compounds, such as FDG, and to image with PET. This method has also experienced minimal success. Because of these limitations to ex vivo labeling and positron-emitting labels, others have explored radiolabeled nanoparticles and FDG for visualizing immune response and inflammation. Further research is needed to determine the viability of this nanoparticlebased technique for routine use in humans. FOCUSING IN ON ENDOCARDITIS In spite of the successes made by modern imaging techniques in detecting infections in many organs, their role in visualizing www.ScienceTranslationalMedicine.org 7 September 2011 Vol 3 Issue 99 99fs3 1 Downloaded from stm.sciencemag.org on September 9, 2011 FOCUS CREDIT: SPRINGER/JOURNAL OF NUCLEAR CARDIOLOGY endocarditis has been limited. The observations made in A Acute endocarditis is classimice with experimental endocally defined as inflammation carditis using both FMT-CT and of the endocardium (inner layPET-CT imaging are intriguer of the heart) and is a major ing and innovative and may clinical problem that progresses provide a powerful means for rapidly. Vegetations, the hallassessing this potentially lifemark lesions of endocarditis, threatening infectious disease are composed of platelets, fiin humans. It is important movB brin, microorganisms, and ing forward that this approach inflammatory cells. Endocardibe tested in larger animals in tis occurs as a result of implanorder to determine the merit of tation of circulating bacteria on optical imaging in detecting lethe cardiac or aortic valves that sions that are distant from the have entered the bloodstream imaging devices. Translation of from the mouth cavity or the this approach would also entail gastrointestinal tracts. Morefurther clarifying the specificity over, preexisting damage to the of this tracer for S. aureus and valve is considered a predisposno other bacteria. Panizzi et al. ing factor for infection. Of all of tested their agent in coagulasethe modern structural imaging negative S. epidermidis–infected techniques, echocardiography mice, which argues for the spec(particularly the transesophaificity of the agent. However, the geal approach) has been the Fig. 1. The role of PET-CT in imaging infection. (A) Transaxial and (B) fact that the S. epidermidis vegmost successful in identifying coronal FDG-PET-CT in a 47-year-old woman with infective endocardi- etations accumulated a small vegetative lesions present in the tis. PET images are on the left; the fused PET-CT images are on the right. amount of tracer (albeit much valves of the heart and the aorta Focal FDG uptake in the heart is noted by arrows in the region of the less than S. aureus) would be in patients with endocarditis. aortic valve. Reproduced with permission from (7). one reason to examine this in Unfortunately, because of the larger vegetation to ensure that small size of the lesions (several the partial-volume effect commillimeters), a large number of lesions go aged endothelium) and fluoresced brightly mon to imaging is not playing a role in the undetected. In addition, structural imag- in ex vivo sections of the mouse aortas (6). minimal visualization of vegetation. ing is nonspecific, cannot differentiate be- A faint response was also noticeable in mice tween active and inactive infection, and is infected with coagulase-negative Staphylo- TRANSLATIONAL CHALLENGES AND unable to assess response to systemic ther- coccus epidermidis, but none at all was seen PROSPECTS apy. Therefore, there is an unmet need for in uninfected control mice. FDG in combination with PET-CT has also methodologies that allow timely detection The goal is to use this noninvasive imag- been used to identify vegetative lesions in the of endocarditis and its complications, such ing method for diagnosis of endocarditis in human heart, as well as at distant sites, owing as embolic lesions at distant locations, and humans. To this end, the authors used their to embolization of detached lesions (7–10). also monitor response to treatment. prothrombin probe with fluorescence mo- FDG binds to both clots and inflammatory A recent technical report by Panizzi et lecular tomography combined with x-ray cells and has therefore been used with great al. (6) describes in vivo detection of en- computed tomography (FMT-CT) to im- interest. Early clinical trial observations usdocarditis caused by coagulase-positive age bacterial growths in vivo in living mice. ing FDG-PET show promise for the routine, Staphylococcus aureus, which is the most They were able to confirm the presence of noninvasive detection of endocarditis (7–10). dangerous and virulent type that might not such vegetations 24 hours after injection As shown in Fig. 1, FDG-PET-CT was used respond readily to treatment. The authors with high specificity and high signal over to detect suspected infective endocarditis in a introduce an interesting method that incor- background when compared with various 47-year-old woman (7). One hour after injecporates both noninvasive fluorescence and control animals. Furthermore, Panizzi and tion, there was substantial uptake of FDG in PET imaging to visualize growing bacte- colleagues were able to visualize in vivo the the infected areas of the aortic valve, which rial vegetations. First, a mouse model of S. response to treatment with the antibiotic was confirmed with separate blood cultures aureus endocarditis was developed that re- vancomycin, which resulted in a decrease in to be coagulase-negative Staphylococci. Other capitulates the bacterial lesions seen in the signal over the course of 48 hours. Lastly, a single-case studies have described the applihuman condition. Because S. aureus is able new prothrombin-based probe was gener- cability of FDG-PET for diagnosing infection to induce blood coagulation via staphyloco- ated for PET-CT imaging, which allowed for in the heart (8–10), especially when echocaragulase, Panizzi et al. created a fluorescent radiological imaging with standard instru- diography presents unclear results, such as in prothrombin-based probe that could be ac- mentation widely available in clinics. This the case of patients with prosthetic vales or tivated in vivo by the enzyme. When inject- radiolabeled prothrombin probe was simi- indwelling pacemakers (7). ed into mice with endocarditis, the probe larly able to confirm the presence of vegetaThe method described by Panizzi et al. deposited at the sites of vegetation (dam- tions in damaged aortas in living mice. allows for concurrent imaging to detect www.ScienceTranslationalMedicine.org 7 September 2011 Vol 3 Issue 99 99fs3 2 Downloaded from stm.sciencemag.org on September 9, 2011 FOCUS vegetative lesions and confirmation of the type of bacteria present by use of the prothrombin dye. This multifaceted approach is promising for early detection and treatment in humans; however, further validation will be needed to assess its utility in routine diagnosis. The major concern is that a very small fraction of the vegetations might be stained superficially with the compounds described, which may not be resolved by existing PET imaging instruments. Therefore, there is a need for novel imaging efforts that overcome some of the deficiencies associated with PET and CT. The major challenges of visualizing relatively small lesions, such as vegetations due to endocarditis in the heart and aortic valves, are twofold: the size of the abnormalities to be detected and the constant rhythmic motion of the heart. Although modern structural imaging modalities, such as CT and ultrasound, have high spatial resolution and can generate scans quickly, they lack high-contrast resolution and therefore suffer from low sensitivity in detecting these lesions. PET-CT has the greatest potential for overcoming these deficiencies and should be pursued as a viable option for managing patients with this serious heart infection. Newly developed PET compounds should label the entire lesion and not bind superficially to the surface. At present, the suboptimal spatial resolution of PET (around several millimeters under ideal conditions) in detecting small lesions will prevent the accurate diagnosis of endocarditis. To overcome further deterioration of image quality owing to cardiac motion, gated imaging based on cardiac cycle by using ECG may improve image resolution and, in turn, the sensitivity of the PET-based approach. Assessing global disease activity in the entire heart may prove to be a possibility in this setting. As described in this Focus, there are several emerging and novel approaches to diagnosing and treating infective endocarditis early (6–10). Large-scale, prospective multicenter trials should be carried out to define the merits of these novel imaging methodologies in the near future. Until more human trials are undertaken, the potential role and benefit of such approaches in the clinic will remain unclear. REFERENCES AND NOTES 1. J. B. Alavi, A. Alavi, J. Chawluk, M. Kushner, J. Powe, W. Hickey, M. Reivich, Positron emission tomography in patients with glioma: A predictor of prognosis. Cancer 62, 1074–1078 (1988). 2. H. Zhuang, M. Pourdehnad, E. S. Lambright, A. J. Yamamoto, M. Lanuti, P. Li, P. D. Mozley, M. D. Rossman, S. M. Albelda, A. Alavi, Dual time point 18F-FDG PET imaging for differentiating malignant from inflammatory processes. J. Nucl. Med. 42, 1412–1417 (2001). 3. S. Basu, A. Alavi, Unparalleled contribution of 18F-FDG PET to medicine over 3 decades. J. Nucl. Med. 49, 17N– 21N, 37N (2008). 4. S. Basu, T. Chryssikos, S. Moghadam-Kia, H. Zhuang, D. A. Torigian, A. Alavi, Positron emission tomography as a diagnostic tool in infection: Present role and future possibilities. Semin. Nucl. Med. 39, 36–51 (2009). 5. K. E. Britton, D. W. Wareham, S. S. Das, K. K. Solanki, H. Amaral, A. Bhatnagar, A. H. Katamihardja, J. Malamitsi, H. M. Moustafa, V. E. Soroa, F. X. Sundram, A. K. Padhy, Imaging bacterial infection with (99m)Tc-ciprofloxacin (Infecton). J. Clin. Pathol. 55, 817–823 (2002). 6. P. Panizzi, M. Nahrendorf, J. L. Figueiredo, J. Panizzi, B. Marinelli, Y. Iwamoto, E. Keliher, A. A. Maddur, P. Waterman, H. K. Kroh, F. Leuschner, E. Aikawa, F. K. Swirski, M. J. Pittet, T. M. Hackeng, P. Fuentes-Prior, O. Schneewind, P. E. Bock, R. Weissleder, In vivo detection of Staphylococcus aureus endocarditis by targeting pathogenspecific prothrombin activation. Nat. Med. (2011). 10.1038/nm.2423 7. S. H. Vind, S. Hess, Possible role of PET/CT in infective endocarditis. J. Nucl. Cardiol. 17, 516–519 (2010). 8. S. Moghadam-Kia, A. Nawaz, B. C. Millar, J. E. Moore, S. E. Wiegers, D. A. Torigian, S. Basu, A. Alavi, Imaging with (18)F-FDG-PET in infective endocarditis: Promising role in difficult diagnosis and treatment monitoring. Hell. J. Nucl. Med. 12, 165–167 (2009). 9. R. F. Yen, Y. C. Chen, Y. W. Wu, M. H. Pan, S. C. Chang, Using 18-fluoro-2-deoxyglucose positron emission tomography in detecting infectious endocarditis/endoarteritis: A preliminary report. Acad. Radiol. 11, 316–321 (2004). 10. J. Van Riet, E. E. Hill, O. Gheysens, S. Dymarkowski, M. C. Herregods, P. Herijgers, W. E. Peetermans, L. Mortelmans, (18)F-FDG PET/CT for early detection of embolism and metastatic infection in patients with infective endocarditis. Eur. J. Nucl. Med. Mol. Imaging 37, 1189–1197 (2010). 10.1126/scitranslmed.3003044 Citation: A. Alavi, B. Saboury, S. Basu, Imaging the infected heart, Sci. Transl. Med. 3, 99fs3 (2011). www.ScienceTranslationalMedicine.org 7 September 2011 Vol 3 Issue 99 99fs3 3 Downloaded from stm.sciencemag.org on September 9, 2011 FOCUS