A first-generation intravascular NIRF imaging prototype was developed in 2008 to test the feasibility of detecting NIRF activity in living subjects [54]. This early system was constructed from a near-infrared optical sensing fiber that was manually pulled back through the vessel to identify local NIRF signal (Fig. 4.3). Performance characteristics demonstrated NIRF detection in a 40 μm diameter focal spot approximately 2 mm from the catheter imaging source. However, without automated translation and rotation capabilities, this one-dimensional NIRF system lacked comprehensive 360° vessel NIRF detection and therefore could only collect data from a localized sector of the vessel wall. In addition, this stand-alone NIRF imaging system did not provide anatomic information and therefore required the use of a sequential structural intravascular imaging platform (e.g., IVUS, OCT/OFDI) to attempt localizing NIRF signal to endovascular regions of interest.

To overcome the lack of circumferential NIRF detection and inability for precise longitudinal anatomic localization in the one-dimensional NIRF catheter design, a second-generation NIRF imaging system was engineered [46]. This two-dimensional NIRF catheter was coupled to computer-controlled motors for automatic 360° rotation and 0.5 mm/s translation pullback (Fig. 4.4). In fluorescent phantom experiments submerged in an artificial blood-like media, the two-dimensional NIRF system had an angular resolution of 35–42° (i.e., ability to resolve >8 sectors) and longitudinal resolution of 1.0 mm at catheter-to-phantom distances of 3.0 and 2.0 mm, respectively. Although the two-dimensional intravascular NIRF imaging system marked a major step forward toward achieving detailed, comprehensive plaque characterization, it suffered from the same drawback as the one-dimensional NIRF prototype – being a stand-alone modality without accompanying structural imaging information.

In a major advance for intravascular NIRF imaging, NIRF was recently paired with OCT/OFDI in a hybrid catheter system to overcome the limitations of earlier one- and two-dimensional stand-alone NIRF catheter systems [55]. As intravascular NIRF imaging lacks anatomical information, the addition of exactly co-registered OFDI allows precise mapping of the local NIRF signal with plaque microstructure. The design of the NIRF-OFDI system employs specialized fiber optics that can simultaneously transmit NIRF through the outer fiber cladding and OFDI through the central core. The distal end of the imaging fiber is coupled to a hand-drawn ball lens that directs light into the tissues at 90° to the optical fiber axis. At the back end, the optical fiber is coupled to the image-processing interface by a customized rotary junction that allows high-speed image acquisition simultaneously from both NIRF and OFDI channels during catheter pullback (Fig. 4.5). In benchmark testing, the NIRF-OFDI system performance was excellent at approximately 7 and 30 μm axial and transverse resolution, respectively, 25.4 frames per second image acquisition, up to 20 mm per second automated pullback speed, a 4.6 mm imaging field depth in saline, and <1 nM sensitivity for near-infrared fluorochromes. Overall, the NIRF-OFDI imaging system offers distinct advantages for enhanced diagnosis over either imaging platform alone and holds great promise for clinical use.

New catheter designs pairing intravascular NIRF with IVUS are also being developed, given the longstanding experience with IVUS in modern-day cardiac catheterization laboratories, where IVUS has been the clinical standard intravascular imaging modality [56,57]. Furthermore, NIRF-IVUS holds technical advantages over NIRF-OFDI in certain areas, despite OFDI exhibiting higher spatial resolution than IVUS. Compared to OCT/OFDI that requires rapid flushing with saline or, more commonly, iodinated contrast to displace blood from the imaging field, both IVUS and NIRF can image the vessel wall through blood without the need for flushing. Eliminating coronary artery flushing by using a NIRF-IVUS system can be time saving, decreases the small incremental risk of vessel trauma and arrhythmias associated with flushing, and lessens the overall contrast load that can result in acute kidney injury. Furthermore, IVUS has improved tissue depth penetration over OCT/OFDI and the ability to image coronary ostia that cannot be evaluated with OCT/OFDI due to difficulty effectively flushing blood from these locations. In evaluation of the left main coronary artery, these characteristics give IVUS a distinct competitive advantage over OCT/OFDI [58].

Complications related to atherosclerotic plaques prototypically involve rupture of the overlying protective fibrous cap, exposing thrombogenic elements to the circulating bloodstream [16,59]. Enhanced plaque proteolytic activity, which results in degradation of structurally important collagen and elastin fiber networks, has been mechanistically linked to plaque rupture [19]. In particular, cathepsins, a well-described cysteine protease family, have been implicated in important atherosclerosis pathology [25]. Cathepsins S and K are present at sites of plaque rupture at autopsy in humans [26], and other cathepsin family members such as cathepsin B associate with atherosclerotic plaque complications [60–63]. Therefore, cathepsins represent high-value molecular imaging targets for identifying high-risk atherosclerotic plaques.

In a significant technological achievement, smart tissue-activatable cathepsin protease NIRF molecular imaging agents have been developed [31,64,65], preclinically tested for cathepsin imaging [29,46,54,61], and commercialized (ProSense/VM110, PerkinElmer, Waltham MA). These specialized agents tightly pack 15–20 near-infrared fluorochromes onto an inert protected graft copolymer scaffold that causes the fluorochromes to be “quenched” and thus nonfluorescent at baseline [65,66]. The polymer scaffold backbone is comprised of the amino acid polylysine covalently bound to methoxypoly (ethylene glycol) chains, which in the presence of protease activity is cleaved to liberate individual fluorochromes detectable by NIRF imaging systems (Fig. 4.6). Compared to constitutively reporting agents, these smart bioreporters have major signal-to-noise advantages, since they exhibit minimal background at baseline, emit NIRF only when active tissue proteolysis is present, contain an intrinsic signal amplification strategy that liberates multiple fluorochromes from each nanosensor, and have proven to be nontoxic. To date, NIRF protease-activatable molecular imaging agents have been engineered to report on cathepsin B [65], cathepsin D [67], cathepsin K [31], and MMP 2 and 9 [68]. For further development potential, the oligopeptide, graft copolymer design of these agents form a versatile platform, enabling linkage of additional drugs or molecules.

Using the early prototype one-dimensional intravascular NIRF imaging system, cathepsin protease activity was assessed in the balloon-injured atherosclerotic iliac arteries of hypercholesterolemic rabbits [54]. Prosense VM110 injected 24 h prior to NIRF imaging revealed increased cathepsin protease activity at sites of atheroma formation by manual catheter pullback that was confirmed by ex vivo macroscopic fluorescence reflectance imaging (FRI) and matched histopathology (Fig. 4.7). Significantly, NIRF plaque detection was accomplished through flowing blood without the need for vessel flushing of occlusion, demonstrating the advantageous in vivo light transmission properties of NIRF light wavelengths. Normal, healthy rabbits injected with Prosense VM110 and atherosclerotic rabbits injected with a saline control vehicle exhibited minimal tissue NIRF signal consistent with background autofluorescence, thus demonstrating the specificity of this agent for NIRF protease detection. Since the rabbit iliac artery is of similar caliber to a small coronary artery (2.0–2.5 mm), this study demonstrated the feasibility of NIRF catheter detection systems to imaging human coronary plaques.

Atherosclerotic plaque cathepsin proteolytic activity in the high-cholesterol rabbit model was assessed more comprehensively with the second-generation, two-dimensional NIRF imaging catheter that provides automated pullback with circumferential imaging capabilities [46]. Furthermore, this study extended in vivo NIRF imaging to the larger 3.0–4.0 mm diameter aorta, more consistent with the typical size range of proximal epicardial human coronary arteries. Eight weeks after aortic balloon injury, rabbits were administered i.v. Prosense VM110 or saline control 24 h prior to performing in vivo NIRF imaging. Anatomic plaque information was evaluated by x-ray angiography and grayscale IVUS and then carefully matched to the NIRF data set using side branches and other fiducial markers. Fusion images revealed co-localization of enhanced NIRF cathepsin protease activity with aortic atherosclerotic plaques, demonstrating the ability to identify inflamed plaques in vivo using the second-generation NIRF system (Fig. 4.8). Quantitative analysis of the plaque NIRF signal revealed excellent performance characteristics with significantly greater signal-to-noise (SNR) and target-to-background (TBR) ratios in aortic plaques from rabbits receiving Prosense VM110 compared to saline control (SNR 12.6 vs. 1.3, p = 0.02; TBR 6.3 vs. 1.1, p = 0.02). Ex vivo tissue analysis confirmed the in vivo NIRF findings, including the presence of histological plaque inflammation (i.e., cathepsins and macrophages) in all balloon-injured animals regardless of whether they were administered saline as a control injection with minimal in vivo NIRF signal or Prosense VM110. Notably, immunohistochemistry for plaque cathepsin proteases may be positive in regions without Prosense VM110 NIRF signal, since Prosense VM110 reports only on biologically active cathepsin protease activity rather than total (inactive plus active) protease content.

Building on prior experiments with earlier-generation stand-alone NIRF imaging systems, inflamed aortic atherosclerotic plaques in cholesterol-fed, balloon-injured rabbits were assessed with integrated NIRF-OFDI molecular-structural imaging using Prosense VM110 injected 24 h beforehand [55]. OFDI-defined plaques revealed enhanced NIRF cathepsin protease signal that was absent in normal vessel segments. Within a single plaque, heterogeneous NIRF proteolysis was observed at high spatial resolution and precisely mapped to the OFDI microstructure (Fig. 4.9). At histopathology, NIRF-positive aortic plaques exhibited increased macrophage infiltration and cathepsin B expression. By combining exact anatomically matched biological and structural plaque information as demonstrated in this feasibility study, intravascular NIRF-OFDI is poised to inform upon previously unappreciated and inaccessible high-risk atherosclerotic plaque features in vivo such as cysteine protease activity, with the possibility to perform serial imaging studies that evaluate drug treatment effects on plaque inflammation in living subjects. Furthermore, as OFDI signal standard deviation threshold measurements can detect resident plaque macrophages in macrophage-dense lesions [69], ongoing studies are attempting to understand whether adding NIRF imaging of proteolytic activity can identify a subset of the OFDI macrophage population that represent activated or M1-type macrophages.

For clinical translation of intravascular NIRF imaging technology, targeted NIRF molecular imaging agents for patient administration must be made available. At present, Prosense VM110 is not approved for human use, and therefore, alternative NIRF inflammation-sensing molecular probes have been pursued to speed clinical translation. Indocyanine green (ICG) is a candidate near-infrared fluorochrome already approved by the Food and Drug Administration (FDA) for intravascular blood flow measurements in ophthalmic retinal imaging and for assessment of cardiac output and hepatic function [70,71]. ICG is an amphiphilic fluorescent dye that binds circulating plasma proteins including atherogenic low-density lipoprotein (LDL) [72] and has been reported to localize within inflamed tissues [73].

Given these promising attributes, ICG was selected for further evaluation of its ability to identify inflamed atheroma and plaque lipid accumulation in vivo with one-dimensional stand-alone NIRF imaging [74]. Eight weeks following aortoiliac balloon injury, hyperlipidemic rabbits were injected with an FDA-approved ICG dose (1.5 mg/kg) prior to in vivo NIRF imaging. Automated pullbacks matched with IVUS for plaque topography revealed locally increased ICG signal in atheroma that was durable for up to 45 min after ICG administration (Fig. 4.10). Ex vivo FRI macroscopically corroborated the in vivo NIRF ICG signal. Detailed fluorescence microscopy and histopathology matched to NIRF and IVUS was then performed to localize the tissue source of ICG plaque signal. Within atheroma, ICG co-localized with lipid (Oil Red O neutral triglyceride staining) and macrophages (RAM-11 antibody). Extending this association further to human tissues, in vitro testing revealed ICG binding to acetylated LDL and ingestion by human macrophages and experimentally derived foam cells, and ex vivo incubation of resected carotid endarterectomy specimens with ICG demonstrated ICG uptake in plaque regions with increased lipid and macrophage content. Therefore, although additional mechanisms of ICG binding to lipid populations and macrophage subsets need to be elucidated, ICG appears to be a promising clinically available NIRF molecular imaging agent for inflamed, lipid-rich atherosclerotic plaques that may accelerate human translation of intravascular NIRF imaging studies.