Abstract
The International Atomic Energy Agency organized a technical meeting at its headquarters in Vienna, Austria, in 2022 that included 17 experts representing 12 countries, whose research spanned the development and use of radiolabeled agents for imaging infection. The meeting focused largely on bacterial pathogens. The group discussed and evaluated the advantages and disadvantages of several radiopharmaceuticals, as well as the science driving various imaging approaches. The main objective was to understand why few infection-targeted radiotracers are used in clinical practice despite the urgent need to better characterize bacterial infections. This article summarizes the resulting consensus, at least among the included scientists and countries, on the current status of radiopharmaceutical development for infection imaging. Also included are opinions and recommendations regarding current research standards in this area. This and future International Atomic Energy Agency–sponsored collaborations will advance the goal of providing the medical community with innovative, practical tools for the specific image-based diagnosis of infection.
Infections remain a major threat to human health globally (1). The coronavirus disease 2019 pandemic has highlighted a pressing need to develop and translate innovative technologies to detect and treat infectious disease. Even before the onset of the pandemic, infections ranked third in mortality but first in morbidity among all human diseases in 2017, primarily affecting younger, healthier populations (2). There were an estimated 11 million sepsis-related deaths in 2017, accounting for about 20% of deaths globally, with the highest incidence reported in developing countries (3). By 2050, antimicrobial drug-resistant infections are expected to become the leading cause of death globally and surpass those due to cancer (4). The potential cost of drug-resistant infections has been estimated to be as high as $100 trillion worldwide (5). We have also observed a dramatic rise in hospital-acquired (nosocomial) infections affecting at-risk patients during the pandemic. Enterobacterales pathogens, especially K. pneumoniae, and fungi including Aspergillus spp. are an important cause of secondary pneumonias in hospitalized patients with coronavirus disease 2019 (6).
Current diagnostic approaches to detecting bacterial infections, such as microscopy, microbiology, and molecular techniques (nucleic acid amplification and mass spectrometry), require clinical samples (blood, urine, stool, or cerebrospinal fluid) for culturing and sensitivity testing and infection-relevant assays. However, it is increasingly recognized that many different infectious foci with distinct bacterial burdens, antimicrobial exposures, and local biology can coexist in the same host (7–9). Clinical samples may not accurately represent the local biology at infectious sites and thus are either not sensitive to or not representative of the bacterial infection (10,11). Surgical resection or biopsy is often the last resort for obtaining infected tissues, and because of the associated morbidity, these techniques are generally limited to the most accessible lesions identified at a single time point. Additionally, sampling methods fail to capture the heterogeneity of multiple lesions existing simultaneously in the same patient, as well as the temporal changes occurring over the course of the infection and its treatment.
Available imaging tools used in the clinic for detection of bacterial infections include radiography, ultrasonography, CT, and MRI, but these imaging tools are based on anatomic changes during disease, which are delayed compared with the biochemical events occurring within the affected tissues. Structural abnormalities are also nonspecific and reflect a combination of both the infectious agents and the host inflammatory response (12). Molecular hybrid imaging platforms have proven to efficiently localize pathology and assist in the clinical management of several diseases (13,14). These technologies, such as SPECT or PET, can measure molecular pathways in situ and are often used in combination with anatomic imaging (PET/CT, PET/MRI, or SPECT/CT). The use of radiolabeled leukocytes (white blood cells) is considered the gold standard technique for prosthetic (15), vascular graft (16), and diabetic foot (17) infections but requires skill and equipment for sterile blood manipulation. Radiolabeled antibodies against granulocyte antigens can induce human antimouse antibody production in approximately 4% of patients (18), limiting their use. The glucose analog [18F]FDG has also been used to image infection but lacks specificity for pathogens (19).
Despite these advanced tools, there is no universally accepted approach to the specific detection of bacterial infections. Therefore, there is an urgent, unmet need for the development of radiopharmaceuticals that can demonstrate the presence of living pathogens in vivo. However, because of bacterial diversity and the frequency of polymicrobial infections, developmental efforts have focused on both radiopharmaceuticals for panbacterial imaging and radiopharmaceuticals for type- or species-specific imaging. For potential clinical applications, there are advantages and disadvantages to both concepts. Given the breadth of research interests, in general all approaches have been hindered by a relative lack of funding (compared with oncology, for example), lack of uniformly reported data for imaging agents, misconceptions regarding radiation risks, and hurdles in the clinical translation and dissemination of promising radiopharmaceuticals (12).
In March 2022, the International Atomic Energy Agency (IAEA) organized a technical meeting titled “The Status of Radiolabeled Molecules for Infection and Inflammation Imaging” in Vienna, Austria, to evaluate and address these challenges. This summary should be used as a road map for advancing research in this field, understanding the potential clinical use of radiopharmaceuticals and their role in clinical decision-making, and most importantly motivating funding agencies and industry to support and develop pathogen-specific imaging technologies. Although the focus of this meeting was bacterial infection, the conclusions rendered may be expanded and tailored to nonbacterial pathogens whose detection via nuclear imaging is increasingly reported in the peer-reviewed literature.
CLINICAL MOTIVATIONS
With the advance of medical imaging technologies, there has been a sustained interest in developing new tools to detect and monitor bacterial infections noninvasively—particularly in nuclear medicine. Ideally, nuclear imaging probes should have high sensitivity and specificity for a wide range of pathogens, with enough tissue penetration to reach infected areas despite poor vascular supply while providing quantitative signals proportionate to the bacterial burden. They should also be chemically stable in blood and tissues; safe, with acceptable radiation exposure; and manufacturable at a reasonable expense (20,21). However, this magic bullet is not feasible for all imaging applications since disease location, type of pathogen, presence of comorbidities, chronicity of infection, and therapeutic interaction may influence the diagnostic accuracy of a given technique.
Therefore, analogous to the development of imaging agents in oncology, pathogen-specific imaging will greatly benefit from having multiple complementary agents with applicability to different clinical conditions. Several radiopharmaceuticals should be developed targeting variable pathways, allowing differentiation between infection and sterile inflammation and characterization of individual or classes of pathogens (e.g., gram-positive vs. gram-negative bacteria). A collaborative multidisciplinary environment with expert perspectives is essential, as is sharing information regarding how to best conduct imaging studies, interpret data, and include appropriate controls. A central agency (e.g., the IAEA) with a global focus represents an ideal platform to share this information and conduct multicenter comparisons. To allow replication of experiments at different sites, transparency in experimental methods for both preclinical and clinical studies is required, as well as willingness by the researchers to distribute data and standardize reporting.
The bacterial imaging field, both for SPECT and for PET, has grown significantly from 2000 to 2019, followed by a reduction in publications during the coronavirus disease 2019 pandemic, based on articles found via PubMed from 2001 to 2022 by searching “bacteria AND imaging AND scintigraphy/PET AND [year].” This refocusing of research effort may reflect a new focus on coronavirus disease 2019–related diagnostics, a pandemic-related loss of resources needed to conduct this type of research, or most likely both. Therefore, it is essential to reinvigorate this field, particularly with our new knowledge of infections—their transmission, morbidity, and mortality.
With respect to pathogen-specific imaging, it is important to learn from our previous mistakes. One of the first agents to be clinically translated for infection imaging was [99mTc]ciprofloxacin, commercially known as Infecton (Draxis). Although it was rapidly evaluated in hundreds of infected patients with promising results, these initial hopes were dashed when subsequent clinical studies (22–24) showed poor specificity of [99mTc]ciprofloxacin for bacterial infections. The panel at the IAEA agreed that from the beginning [99mTc]ciprofloxacin was a poorly chosen and validated tracer for numerous reasons, including its limited affinity for bacteria (reflected by fast efflux rates from affected tissues) and binding to both bacteria and mammalian cells (25,26). This was a valuable lesson in the need to thoroughly characterize bacteria-specific imaging agents to confirm their mechanism of action and specificity before investing extensive resources in their clinical translation. In this case, in vitro tests were not satisfactory (27–29) and heterogeneity of clinical studies limited their credibility. These studies had variable imaging indications, divergent gold standards, and poor controls and proved challenging to interpret (23,30–33).
To attract industry investment in new imaging technologies, attention should be paid to their potential profitability at all stages of development. For example, the preservation of intellectual property via patents is essential and often overlooked. Indeed, patented radiotracers can be more easily acquired and produced by pharmaceutical companies, reducing the risk of competing technologies. A lack of intellectual protection will likely discourage industry investment, even if the science itself is promising.
DEVELOPMENT OF PATHOGEN-SPECIFIC IMAGING METHODS
The discussion at the IAEA focused on the development of infection-targeted radiopharmaceuticals, including the basic in vitro and in vivo studies needed to validate their clinical relevance.
Infrastructure Requirements
An important consideration in developing any imaging agents is the required infrastructure, with the safe handling of infectious agents representing a special challenge. For any radiopharmaceutical, relevant technologies include a cyclotron or radionuclide generator, shielded fume hoods (i.e., hot cells), a radioactivity-counting instrument (e.g., a γ-counter), high-performance liquid chromatography, and preclinical scanning equipment (PET or SPECT). For clinical translation, quality control evaluation including identity testing, pyrogen evaluation, radiochemical evaluation (yield, purity), and chemical stability is required. Staff with radiochemistry and regulatory expertise imply potentially high costs, which can be a limitation for the development of new agents. Considerations for handling pathogens and biosafety training are critical for conducting these studies. However, groups focused on radiochemistry are not familiar with regulations surrounding pathogens, as they have traditionally focused on cancer and neurologic disorders. Additional infrastructure challenges are related to infections with drug-resistant pathogens requiring special regimes for patient care and cleaning of hospital spaces (waste management) and hospital beds (isolation areas).
Conception and Planning—Identifying the Target
Infectious diseases are widely heterogeneous illnesses associated with microorganisms that cause disease in humans. Although there is considerable overlap in the pathogenic mechanisms of these microorganisms and the host response to them, the intrinsic characteristics of these pathogens are highly variable. Bacteria, viruses, fungi, and parasites are genetically, biochemically, and metabolically different. Additionally, reference laboratory strains can be significantly different from pathogenic strains. For example, reference laboratory E. coli strains—the uropathogenic CFT073, enterohemorrhagic EDL933, and laboratory strain MG1655—all display a mosaic genome structure that can compose up to 40% of their genes (34). Therefore, it is fundamental to identify the target pathogen and its clinical presentation when planning to develop pathogen-specific imaging agents. The specific characteristics of the agent will vary depending on the target, and the use of clinical strains during the initial studies may be crucial.
Developing new infection-targeted agents depends both on the clinical need and on understanding of the mechanism the technology uses to generate image contrast. Bacteria (prokaryotes) are evolutionary and phylogenetically distinct from eukaryotic cells. These basic differences provide opportunities to leverage fundamental biochemical differences between bacteria and mammalian cells—that is, energetic pathways, nucleic acid use, and cell surface components for the discovery of novel molecules that could be developed into pathogen-specific agents. Although initial efforts to develop pathogen-specific radiopharmaceuticals were based on radiolabeled antibiotics, recent approaches have focused on radiopharmaceuticals that are incorporated by the cell wall or are metabolized by microbe-specific pathways. For example, d-methyl-[11C]methionine and other positron-labeled d-amino acids have targeted bacterial peptidoglycan (Fig. 1) (35–38), whereas 2-deoxy-2-[18F]fluoro-d-sorbitol detects bacteria via the unique metabolism of sorbitol by Enterobacteriaceae (Fig. 2) (39,40). For a more thorough understanding of the target, collaboration with microbiologists and infectious disease physicians is helpful. Generally, attention to the literature and careful screening (including in silico) can identify the targets most relevant to probe design (41). Using an artificial intelligence approach in the selection of potentially pathogen-specific radiopharmaceuticals can make radiopharmaceutical agent development more efficient (42).
Compound Screening and Radiochemistry
Once the bacterial target has been identified, the next step is to obtain lead molecules, which may require a conventional compound screen, structure-based design, modification of molecular probes developed for other imaging techniques, or radiosynthesis of metabolite analogs (43,44). Using an unbiased screening approach is essential to the discovery of candidates for pathogen-specific imaging. Screening of candidate compounds should be performed in whole bacterial cell cultures since working with an isolated target ignores critical determinants of clinical performance, such as cell wall penetration.
Multiple comprehensive reviews have been published on the radiochemistry of pathogen-specific imaging radiopharmaceuticals (43,45,46). However, many publications in this field lack the minimal requirements to allow validation and reproducibility of the described radiochemistry methodology. When a new agent is described, information regarding its radiochemical purity, radiochemical yield, molar activity, stability, and metabolism is necessary to allow other researchers to evaluate the presented approach and potentially reproduce it. Molar activity is particularly important to report and evaluate to address the presence of competing cold materials in a radiopharmaceutical sample. Numerous other considerations are relevant to the chemical specifics of the probe. For example, with radiometal modification of antibodies or related protein formats (i.e., single-domain antibodies), stoichiometry should be evaluated and reported. Both retention of activity and stability of conjugated antibodies should be determined. To rule out in vivo transchelation of 99mTc, a cysteine challenge study may be considered. Basic radiopharmaceutical design criteria are beyond the scope of this discussion, but a summary has been provided in a previous review (47).
In Vitro Testing
Evaluating uptake of radiopharmaceuticals by bacteria in vitro is a critical aspect of their validation. Since certain radionuclides both are costly and have short half-lives (e.g., 11C for PET), the in vitro study of radiopharmaceutical analogs can begin with β-emitting nuclei (14C, 3H) and scintillation counting (41), stable isotope MR spectroscopy (48) (e.g., 13C, 1H, 2H, and 19F), or mass spectroscopy (49). The essential validation of a new radiotracer concept is when it has been successfully labeled and incubated with bacterial cultures to detect specific incorporation (37). These studies are usually conducted with bacterial cultures in the growth or exponential phase, although of course other assessments are possible. After bacterial washing and detection of retained radioactive signals, a gross assessment of tracer retention can be made by the percentage of tracer retained, that is, the percentage uptake. However, these data should be normalized to bacterial count, which is not obtained via an estimate (e.g., an E. coli culture with an optical density of 1 measured at a wavelength of 600 nm represents 8 × 108 organisms) but by serial dilutions and plating to determine the number of colony-forming units to be reported. The most relevant controls include the use of heat-killed organisms and blocking using a nonradiolabeled version of the radiopharmaceutical. These blocking, or competition, studies can be used to explore the effect of molar activity on radiotracer performance. Finally, several pathogenic species, as well as multiple strains of the same species, should be included in these analyses. In addition to clinical isolates, commercially available bacteria should be used to allow reproduction of results by other groups.
Several in vitro studies are infrequently performed or use variable experimental conditions. For example, some investigators perform efflux studies whereby after bacterial radioactivity retention, the cells are washed and the subsequent loss of radioactivity over time is determined (50). There is also variability in the medium used, and some components may compete with exogenous radiopharmaceuticals for bacterial incorporation. At this point, there is no standard medium used although investigators should consider appropriate mimicry of the nutrient makeup of the human body.
In Vivo Validation
Once the in vitro characterization of a probe has been completed, subsequent validations in animal models are frequently performed. Regulatory requirements for the development of animal models of infections vary considerably across different countries and institutions. If excessive, these can be an additional burden to researchers (51). Animal models and relevant controls are well summarized in the consensus report by Signore et al. (52). When choosing an animal model, it is important to thoroughly understand the human infection that is being studied, as well as the strengths and limitations of a given model. The European Association of Nuclear Medicine recently published useful guidelines for choosing the appropriate animal model for preclinical experiments (53). In most cases, the models used should recapitulate human pathologies. When tracer sensitivity to different pathogens is being compared, a dual infection model (e.g., a mouse infected with 2 pathogens) (38) or separate carefully generated cohorts (e.g., in comparing [99mTc]hydrazinonicotinamide polymyxin B accumulation in Pseudomonas aeruginosa and Staphylococcus aureus) may be used (Fig. 3) (54). The volume of distribution, metabolism, excretion, vascular leakage, etc., are also key variables that should be considered before choosing a specific model.
Determining and standardizing the readouts used to quantify signals from pathogen-specific agents in animal models are also key to comparing different agents and reproducing the reported findings. For example, for PET imaging of bacterial infections, the agent should be injected at a time point when the infection has been allowed to incubate for a sufficient time (e.g., 8–24 h) to resemble human pathology when inflammatory response peaks and bacteria are in different metabolic states. Determination of the stability of the agent in blood (or tissues if applicable) at the time of imaging should also be reported. Because of replication, the bacterial burden injected at a given site is much lower than found hours later. Therefore, the bacterial burden at the site evaluated should be determined immediately after imaging has been performed.
The use of animal models beyond rodents in development of imaging research has also been suggested (53,55). The rabbit stands out among nonrodent mammals used in research because of its relatively small size, short gestation period (29–31 d), and potential for timed mating and superovulation (56). Myriad imaging studies have reported the successful use of rabbit infection models (57–59). Given ethical concerns, using nonhuman primates (macaques, baboons, marmosets, and African green monkeys) in biomedical research is usually allowed only in research areas for which no alternative is available (60,61); however, research on nonhuman primates regularly necessitates special facilities and expertise. Although expensive, nonhuman primates are invaluable tools to study complex infection pathogenesis (simian HIV or tuberculosis) (62) and are suitable for preclinical imaging studies (63,64), in particular for tracer biodistribution or radiation dosimetry (65–67), translational research on human respiratory infections, and pharmaceutical drug development (68–70). Beyond these common animals, the study of species-specific diseases might involve pigs, cats, dogs, cattle, horses, fish, and birds (71,72).
Distinguishing infection from inflammation depends on the typical host response via initiation of the innate immune system followed by an adaptive response targeting the pathogen (19,73). Frequently, inflammation may persist despite infection control (74). Inclusion of appropriate controls is an important determinant when developing specific infectious imaging agents. Whenever possible, contralateral limb or skin site controls, as demonstrated by the tissue cage model (75), can be used to test the specificity of the agent and further distinguish between infectious and sterile inflammation. Therefore, a rational preclinical screen should be designed to, first, understand the kinetics of the test molecular imaging agent in sterile (76–78) as well as infectious (39,79) animal inflammation models and, second, understand target organ function to compare the sensitivity of the imaging agent in animal models of infection and sterile inflammation. The imaging agents should be validated with the capability for dual or hybrid imaging platforms such as SPECT/CT, PET/CT, and PET/MRI (80).
An accurate analysis of the preclinical images is fundamental to determine the viability of the candidate agent for pathogen-specific imaging. The most used approach to determining the region (or volume) of interest via images is intrinsically operator-dependent. Therefore, efforts should be made to minimize bias (e.g., using the CT instead of the PET images to determine the region of interest). A frequently used unit to represent imaging results in PET/SPECT is SUVmean, which considers average signals in a region of interest, corrected for the dose-decay–adjusted injected dose and the weight of the animal. Other methods of data quantification are available, and researchers should explain the methodology used for the analysis. Careful dissection and ex vivo analyses of all tissues should be performed via a radiation-detecting instrument (i.e., γ-counter). Accurate identification of infected and noninfected tissues can be used to generate an uptake value, normalized to mass (i.e., percentage injected dose per gram). The tissues can be subsequently homogenized and plated to normalize the data for the number of viable bacteria (colony-forming units); this type of analysis is essential for evaluating the sensitivity of a radiopharmaceutical or comparing the sensitivity of different tracers.
TRANSLATION AND THE FUTURE
The general process of translating new nuclear medicine technologies has been explored in numerous reviews (81,82) and in the context of dedicated workshops, such as that organized by the National Institute of Biomedical Imaging and Bioengineering (National Institutes of Health) (83). The basic approach involves the approval of both government and institutional regulatory bodies, toxicology studies as required, radiochemical optimization, and first-in-humans studies usually initiated for dosimetry evaluation. For any tracer, a major challenge is securing the funding to accomplish this work, given administrative expenses and the high cost of radiopharmaceutical production. Many researchers at the meeting felt these costs diminished the number of patients who could reasonably be scanned using a new tracer—thus limiting the conclusions obtained. A second challenge is the difficulty of proving the utility of infection-targeted radiopharmaceuticals in rigorous, multicenter studies. Even for researchers who wish to share and collaborate, securing the funding required for this effort is difficult. Finally, infection-targeted tracers face particular barriers to widespread clinical adoption, described below.
There Is Currently Limited Engagement of Stakeholders
To be successful, physicians in numerous disciplines need to consider infection imaging essential to clinical practice. The collaboration of radiologists, nuclear medicine physicians, infectious disease doctors, surgeons (especially orthopedic surgeons and neurosurgeons), and other specialists will be essential in driving this field forward. In addition, partnerships with industry, including commercial radiopharmacies, are crucial to rendering these technologies profitable and sustainable.
Current Patient Studies Are Not Sufficiently Convincing
The latest generation of microbe-specific tracers is highly compelling, but few carefully conducted patient studies support their use. Infectious disease is a broad topic, requiring time for researchers to produce relevant data in patients.
Infection-Targeted Nuclear Medicine Tools Do Not Fit into an Existing Clinical Workflow
Access to nuclear medicine tools may be limited for the diagnosis of infection in the acute care setting. Most radiotracers cannot be synthesized on demand even during the regular operating hours of a radiopharmaceutical facility and, as a result, can often be used only when the patient is already undergoing antimicrobial therapy due to the urgency of treatment in acute infections. This is a significant limitation to first-in-humans studies, as imaging results can be confounded by the effects of the therapeutic regimen.
CONCLUSION
Meetings such as that recently sponsored by the IAEA are essential in identifying ways for researchers and physicians to better diagnose and treat bacterial infections. The remarkable progress made over the last decade indicates that the successful application of new molecular imaging tools in the clinic will profoundly impact patient care.
DISCLOSURE
No potential conflict of interest relevant to this article was reported.
Footnotes
↵* Contributed equally to this work.
Published online Sep. 28, 2023.
- © 2023 by the Society of Nuclear Medicine and Molecular Imaging.
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.
- 29.↵
- 30.↵
- 31.
- 32.
- 33.↵
- 34.↵
- 35.↵
- 36.
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.
- 67.↵
- 68.↵
- 69.
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- Received for publication April 27, 2023.
- Revision received August 18, 2023.