Molecular Imaging of Infective Endocarditis with 6′′-[¹⁸F]Fluoromaltotriose Positron Emission Tomography/Computed Tomography

Infective endocarditis (IE) is a microbial infection of the inner/endocardial tissue of the heart, most commonly the heart valves, and is generally fatal if untreated¹. Staphylococci are the most common etiologic agents of IE¹. The current approach for diagnosing IE includes the modified Duke criteria¹, which are based on clinical examination, blood cultures, and echocardiographic findings. However, blood cultures and echocardiography can be inconclusive (particularly in patients with prosthetic valves and implantable cardiac devices), leading to a high proportion of unconfirmed cases of suspected IE¹. Computed tomography (CT) and magnetic resonance imaging are also limited in diagnosing IE because they reveal anatomic changes only when the infectious process is advanced and tissue damage has occurred^{1, 2}. Moreover, leukocyte scintigraphy and 2-deoxy-2-[¹⁸F]Fluoro-D-glucose ([¹⁸F]FDG) positron emission tomography/computed tomography (PET/CT) visualize sites of inflammation as well as bacterial infection^{1, 2}, and consequently have limited specificity.

To address these challenges, we developed a PET-based strategy for directly imaging bacteria using the novel probe 6′′-[¹⁸F]Fluoromaltotriose. This probe targets the maltodextrin transporter, a carrier protein that is exclusive to bacteria and not expressed in mammalian cells (Figure 1A, left panel). 6′′-[¹⁸F]Fluoromaltotriose has been previously shown to distinguish bacterial infection from sterile inflammation³. Here, we investigate the diagnostic utility and therapeutic monitoring of 6′′-[¹⁸F]Fluoromaltotriose PET/CT imaging in a preclinical Staphylococcus aureus-induced endocarditis model.

Figure 1. PET/CT imaging with 6′′-[¹⁸F]Fluoromaltotriose detects infective endocarditis (IE) with high sensitivity and specificity.

A) Aortic valve IE detection workflow using 6′′-[¹⁸F]Fluoromaltotriose PET/CT. Mouse model of induced Staphylococcus aureus IE was developed as described in main text. B) i) Bioluminescence imaging (BLI) and ii) representative, non-gated [¹⁸F]Fluoromaltotriose PET/CT images showing bacterial signal in an IE mouse vs. control mouse. L – liver; K – kidney; B – bladder. iii) There was a 2.3-fold higher in vivo tracer signal on PET imaging in IE mice (n=6) vs. controls (n=6) (P<0.001). This was corroborated by ex vivo biodistribution results, which showed a 2.8-fold higher tracer uptake in the IE mice hearts (n=3) vs. controls (n=3) (P<0.001). Data presented as mean ± standard deviation. iv) Significant positive correlation existed between the BLI and PET SUV measurements (R²=0.941, P=0.0014; dashed lines indicate 95% confidence intervals for regression line). C) Representative, thresholded images of 6′′-[¹⁸F]Fluoromaltotriose PET/CT acquired at 60-min post-tracer injection in an IE mouse. No cardiac or respiratory gating was performed. The end of the contrast-filled catheter, placed under echographic guidance, was visualized by CT to confirm the aortic valve’s volume-of-interest on fused PET/CT images for quantitative analysis. There was no indication that the radiotracer adhered to the catheter. D) Histology: i) Hematoxylin and eosin (H&E) sections of control aortic valve showing leaflets composed of scant connective tissue overlain by a smooth endothelial lining; ii) H&E and iii) Gram stain of infected aortic valve showing mats of fibrin (light blue asterisks) with numerous embedded Gram-positive bacterial colonies (red asterisks). The cocci bacteria stained dark purple with the Gram dye. Adjacent to the aortic wall are dense aggregates of neutrophils (dark blue asterisks). Boxed areas are regions of higher magnification. AoV – aortic valve; AoL – aortic lumen; LVLu – left-ventricular lumen. E) Signal from 6′′-[¹⁸F]Fluoromaltotriose PET/CT can be used to monitor treatment response in an IE mouse model as demonstrated by in vivo serial tracer uptake in the same mice (n=3) at baseline before bacterial infection, at 1-day post-infection (but before antibiotic treatment), and at 3-days and 20-days after the start of vancomycin treatment. Antibiotics were administered right after PET/CT scanning at the 1-day post-infection time point. Mean PET/CT tracer uptake in the IE mice (n=3) was normalized to the mean aortic valve uptake in the non-IE control mice (n=3) at each of the time points. Error bars denote standard error of the mean.

Aortic valve IE was induced by placing a 32-gauge catheter within the aortic root via the right carotid artery in CD-1 mice to cause valve injury and then injecting bioluminescent Staphylococcus aureus (strain Xen36; 10⁶ CFU) into the tail vein 24-hours after catheter placement⁴. Control mice had catheters similarly placed without intravenous injection of bacteria. One-hour dynamic PET/CT imaging of 6′′-[¹⁸F]Fluoromaltotriose (~200 μCi) was performed 24-hours after bacterial inoculation and quantitatively analyzed (IE mice: n=6; controls: n=6). In vivo bioluminescence imaging (BLI) was acquired before each PET/CT scan. Lastly, animals were euthanized after imaging and organs were harvested for microbial assessment and histology. An overall concept schematic diagram is shown in Figure 1A (right panel). To assess the ability to monitor treatment response with this tracer, a different group of mice with IE (n=3) were imaged before and after administration of vancomycin (110mg/kg twice per day subcutaneously), and compared to kanamycin-treated (n=3; 800mg/kg/day intramuscularly) and non-IE controls (n=3). All animal procedures followed institutional guidelines.

PET/CT showed a 2.3-fold higher tracer uptake in the aortic valves of IE mice when compared to control mice (maximum standardized uptake value [SUVmax]: 2.13±0.45 vs. 0.92±0.12; P=0.0009) (Figures 1B and 1C). Range of SUVmax in the aortic valve region for IE mice (1.40–2.70) showed no overlap with control mice (0.80–1.10), which confers 100% sensitivity and 100% specificity for this small pilot cohort. BLI and PET measurements were strongly correlated (R²=0.941, P=0.0014). On H&E staining, dense colonies of bacterial cocci embedded in fibrin were evident in the aortic valve leaflets of the IE mice (Figure 1D) and the cocci stained dark purple with the Gram dye. In the treatment arm of the study, 6′′-[¹⁸F]Fluoromaltotriose PET/CT successfully revealed therapy responses. After an initial 50% rise in the aortic valve uptake ratio 1-day post-infection, there was a significant drop back to baseline control values after 20-days of vancomycin treatment (P<0.01) (Figure 1E). Mice with IE that were given kanamycin (i.e., ineffective treatment due to bacterial resistance) did not even survive to the 3-day post-therapy PET imaging time point, highlighting the lethal nature of the disease if not treated properly.

Although other radiotracers have targeted the maltodextrin transporter such as 6-[¹⁸F]Fluoromaltose and [¹⁸F]Maltohexaose, both of these tracers suffer from suboptimal pharmacokinetics (PK) and have poor signal-to-noise (S/N) ratios, particularly in the thoracic region, thus limiting their applications in lung and cardiac infections³. Antibiotics and antimicrobial peptides, which kill or disable bacteria at very low concentrations, have also been radiolabeled, but their effectiveness as radiotracers is limited due to lack of signal amplification^{2, 5}. Moreover, radiolabeled prothrombin analogs like ⁶⁴Cu-DTPA-ProT have the potential for false-positive signaling by non-infectious processes and are, therefore, not entirely specific for bacterial pathogens⁵. It is in these areas (better PK, better S/N, higher specificity, amplification of signal, imageability in thorax) that 6′′-[¹⁸F]Fluoromaltotriose may provide advantages. Limitations of this work include the small sample size in the therapy monitoring arm of the study and the IE animal model which requires administering the bacteria systemically after valvular injury.

To our knowledge, this is the first time that a fluorine-18 PET tracer has been used to specifically image bacterial infection of the heart valves with high sensitivity and specificity in an animal model. Responses to antibiotic therapy could also be assessed using 6′′-[¹⁸F]Fluoromaltotriose PET/CT. This tracer promises to offer significant diagnostic and treatment monitoring capability, with the potential to change the clinical management of patients with bacterial endocarditis. Plans are currently underway to translate 6′′-[¹⁸F]Fluoromaltotriose into the clinic for imaging patients with IE and cardiovascular device infections.