Radiolabeled antibiotics are relatively well-studied in the context of imaging diabetic foot infection (DFI). Previously studied radiolabeled antibiotics target crucial features required for bacterial pathogenicity and survival, including the synthesis of folic acid required for nucleic acid synthesis, cell wall construction, cell membrane structure and function, transcription, and translation [131]. Antibiotic-resistant bacteria pose a major problem for radiolabeled antibiotics. Depending on the uptake mechanism, some bacteria may demonstrate radiotracer accumulation, while others may not. ^99mTc-ciprofloxacin is a well-studied antibiotic imaging agent and, although previous studies originally demonstrated its high specificity, subsequent studies have not been able to reproduce the result [132,133,134,135]. In vivo studies of its ability to differentiate gram-positive from gram-negative bacteria have reached contradictory conclusions. The main drawback to ^99mTc-ciprofloxacin is its accumulation in dead bacteria [132], which leads to false positives, as suggested by a large multicenter clinical trial that reported a sensitivity of 66.7%, a specificity of 85.7%, and an accuracy of 72% for both soft tissue infection and OM [136]. Several radiolabeled antibiotics have been studied to date (Table 1); however, the least well-studied are bacterial cell wall synthesis inhibitors such as vancomycin. The SPECT tracer ⁹⁹ Tc-vancomycin, which was directly labeled, has demonstrated an affinity for S. aureus infectious foci. ^99mTc-vancomycin labeled through ^99mTc-Tc-HYNIC tetrazine click chemistry corroborated this observation with a three-fold uptake increase in a S. aureus infection site compared to controls [137]. Studies using fluorescently labeled vancomycin have also reported promising results [138], although isotopic-labeled vancomycin has shown greater uptake [139]. Although the ^99mTc-vancomycin single-photon emission computed tomography (SPECT) tracer may have a valuable place in the clinical management of DFI, the development of a vancomycin PET tracer may provide better image resolution.

Amino acid uptake by bacteria has also been targeted for radiotracer development. D-stereoisomers of amino acids are found in bacterial cell walls but are not present in humans, making these amino acid isoforms promising pathogen-specific imaging targets. One analog of D-methionine, D-[methyl-¹¹C]methionine, has been used to differentiate sterile inflammation from E. coli and S. aureus infections in a murine model, reporting an SUV_max 8–10 times higher (SUV_max = 0.8% ID/cc) in E.coli and S. aureus compared to [¹¹C]L-methionine (SUV_max = 0.1% ID/cc) [146]. An in vitro study demonstrated the broad sensitivity of D-[methyl ¹¹C] methionine across a panel of clinically relevant DFI pathogens, including E. coli, P. aeruginosa, P. mirabilis, S. aureus, S. epidermidis, and E. faecalis, with the greatest uptake occurring in S. aureus and P. aeruginosa [147]. A recent study compared the uptake of D-5-[¹¹C]glutamine and L-5-[¹¹C]glutamine isomers in a murine dual pathogen myositis model with localized MRSA and E.coli infections. Uptake of L-5-[¹¹C]glutamine was higher in the heat-killed control site when compared to D-5-[¹¹C]glutamine, indicating that the D-stereoisomer had better pathogen specificity. When D-5-[¹¹C]glutamine was compared to ¹⁸F-FDG in a sterile inflammation model, ¹⁸F-FDG uptake was significantly higher, further demonstrating the specificity of D-5-[¹¹C]glutamine for bacterial metabolism [148]. Although ¹¹C-labeled amino acid-based imaging shows specificity, widespread adoption for clinical use may be limited because of the 20-min half-life of ¹¹C.

Microorganisms may also be targeted based on their vitamin uptake. Since biotin is used in the production of fatty acids and is an essential growth factor for S. aureus and other bacteria, it has been explored for infection imaging. Although little data are currently available, an in vitro study using [¹¹¹In]-In-DTPA-biotin reported selective uptake in S. aureus cultures [149]. In addition to targeting biotin, a radiolabeled vitamin B12 derivative, ^99mTc-PAMA has also been used for infection imaging. Since PAMA accumulates in rapidly proliferating cells, the imaging agent has shown high uptake in S. aureus as well as E. coli in vitro and in vivo [150]. Microorganisms commonly use a compound called PABA to synthesize folic acid. A study using ³H-PABA reported accumulation in MRSA as well as other therapy resistant microorganisms. Uptake was 100× greater in the microorganisms than in mammalian cells. When labeled with ¹⁸F, uptake was seen in S. aureus [151].

Other targets for bacterial imaging include nucleoside analogs, peptidoglycan targeting aptamers, siderophores, and radiolabeled antibodies. The nucleoside analog ¹²⁴I-fialuridine (¹²⁴I-FIAU) is a bacterial tyrosine kinase substrate trapped in the cell following phosphorylation. Although it has been previously used to detect prosthetic joint infections [152], a more recent study found that tracer uptake was reduced in the presence of metal artifacts [153]. A clinical study of 16 patients with suspected musculoskeletal infection, who underwent ¹⁸F-FDG and ¹²⁴I-FIAU PET/CT, reported a range of specificity of 51.7–72.7% and a maximum sensitivity of 100%; however, the minimum sensitivity was incalculable. The authors were also unable to come to a definitive conclusion on its diagnostic efficacy because of equivocal clinical findings in a large number of patients [154]. Aptamers are oligonucleotides that have been used to target the peptidoglycan cell wall. One group that used ^99mTc-Antibac1 demonstrated its ability to distinguish between bacterial and fungal infections in vivo and in vitro [155].

Siderophores are small molecules with a high affinity for Fe(III) and are produced primarily by bacteria, but are also produced by virtually all microorganisms—including prokaryotes and eukaryotes [156]. Mammalian macrophages have been reported as sources of siderophores [157]. Organisms use these metal-seeking molecules for extra- or intra-cellular metal transport and storage. Biological processes such as DNA replication, transcription, oxidative stress responses, and respiration all require transition metals. Iron and zinc are among the most abundant in living organisms. Iron is an essential cofactor for respiration and central metabolism. Zinc is the catalytic center for many enzymes and is required for 5–6% of all protein functions [158]. Siderophores are often categorized into the following groups according to the chemical moiety that binds iron: catecholates, hydroxamates, and carboxylates. Mixed type siderophores have two or more of these moieties. Siderophores may actively alter their affinity for iron in order to bind other metals by increasing their expression of metal transporters [158]. Desferricoprogen—a natural siderophore produced by Penicillium chrysogenum and Neurospora crass—reportedly bound to both trivalent (Al(III) and In(III)) or divalent (Cu(II), Ni(II), Zn(II), and Fe(II) metal ions [159]. Transport systems and siderophore production are unregulated during infection and can accumulate in bacteria as well as fungi. An in vitro study which used ⁶⁸Ga-ferrioxamine (⁶⁸Ga-FOXE) showed high uptake in S. aureus; however, it was not corroborated in vivo [160]. In fungi, Aspergillus fumigatus-specific fluorescent siderophore conjugates were used to assess pulmonary aspergillosis in vitro [161]; however, recent in vivo studies have only been performed to assess biodistribution in pulmonary aspergillosis [162]. Further research regarding specificity, sensitivity, and use in additional infection models is needed.

In addition to in vitro labeling, WBCs may also be localized in vivo by using radiolabeled antibodies which target WBC antigens—a technique called immunoscintigraphy. For example, ^99mTc-besilesomab uses the monoclonal antibody besilesomab to bind to the granulocyte-specific antigen-95. Antigen-95 is present on the membranes of granulocytes and their precursors. Within 45 min of initial injection, 10% of the ^99mTc-besilesomab is bound to neutrophils, while 20% is unbound but localizes to the site of infection by nonspecific mechanisms [163]. A recent study of 119 patients with suspected OM compared planar ^99mTc-HMPAO WBC and ^99mTc-besilesomab imaging in the diagnosis of peripheral OM. They found that ^99mTc-HMPAO had a lower sensitivity than ^99mTc-besilesomab (59.0% and 74.8%, respectively); however, ^99mTc-HMPAO had a higher specificity than ^99mTc-besilesomab (79.5% and 71.8%, respectively) [63]. ^99mTc-sulesomab also binds to an antigen found on leukocyte membranes—antigen-90—but only 3–6% of ^99mTc-sulesomab is bound to circulating neutrophils, and 35% is found in bone marrow 24 h post-injection [111]. The results of a study of combat-related infections suggest that ^99mTc-sulesomab may not possess sufficient accuracy (78% accuracy) for the identification of the foci of infections, but may have otherwise comparable diagnostic values (93% PPV, 62% NPV, 72% sensitivity, 88% specificity) [71]. One study of Yersinia enterocolitica infections used immunoPET with ⁶⁴Cu-NODAGA-labeled Yersinia-specific polyclonal antibodies to target the membrane protein YadA and reported colocalization in a dose-dependent manner with bacterial lesions in a murine model [164]. Another study that used the [⁶⁴Cu]NODAGA-IgG₃ monoclonal antibody from mice to examine C. albicans infections demonstrated its accuracy in diagnosing infection in vitro as well as in vivo [165]. A meta-analysis found a sensitivity of 81% and specificity of 86% for anti-granulocyte scintigraphy with monoclonal antibodies [166]. The use of intact antibodies takes up to several days to accumulate near infectious foci. Clinically, this could be a limitation because of the need for rapid identification of the infection site. However, the use of antibody fragments may lead to faster accumulation time. Overall, it appears that radiolabeled antibodies do not have the ability to replace existing methods in a clinical setting because of the relatively low diagnostic values reported by the studies discussed above.

One recent therapeutic application of US used in DFI is cavitation, in which ultrasonic waves are used to create microbubbles which form from the dissolved gas that accumulates in the wound. The molecular purpose of this treatment is to induce compression and movement of the wound cells and small ions to increase protein synthesis and the permeability of vascular walls and cell membranes [133]. Ultimately, reduced inflammation, angiogenesis, and increased cell proliferation and recruitment are expected at the site of infection [134]. Currently, only three studies have used this technique and two of them demonstrated either complete wound closure or a significant reduction in wound area [135,136,137].

Although these emerging imaging probes are targeting the microorganism rather than the host response, the quantitative measures of probe retention are generally much lower than the clinically available tracers that target the host response. Further, while most of these have not been used in DFI specifically, this is an opportunity for the expansion of molecular imaging approaches to DFI.