First-In-Human Dual-Modality OCT and Near-Infrared Autofluorescence Imaging of Coronary Artery Disease

Abstract

OBJECTIVES

We present the first clinical imaging of human coronary arteries in vivo using a multimodality OCT and near-infrared autofluorescence (NIRAF) intravascular imaging system and catheter.

BACKGROUND

While intravascular OCT is capable of providing microstructural images of coronary atherosclerotic lesions, it is limited in its capability to ascertain compositional/molecular features of plaque, including the definitive presence of a necrotic core. A recent study in cadaver coronary plaque has shown that endogenous NIRAF is elevated in necrotic core lesions. The combination of these two technologies in one device may therefore provide synergistic data to aid in the diagnosis of coronary pathology in vivo.

METHODS

We developed a dual-modality intravascular imaging system and 2.6-F catheter that can simultaneously acquire OCT and NIRAF data from the same location on the artery wall. This technology was utilized to obtain volumetric OCT-NIRAF images from 12 patients with coronary artery disease undergoing PCI. Images were acquired during a brief, non-occlusive 3-4 ml/sec contrast purge at a speed of 100 frames per second and a pullback rate of 20 or 40 mm/sec. OCT-NIRAF data were analyzed to determine the distribution of the NIRAF signal with respect to OCT-delineated plaque morphological features.

RESULTS

High quality intracoronary OCT and NIRAF image data (>50 mm pullback length) were successfully acquired without complication in all patients (17 coronary arteries). The maximum NIRAF signal intensity of each plaque was compared to OCT-defined type, showing a statistically significant difference between plaque types (one-way ANOVA, p<0.0001). Interestingly, coronary arterial NIRAF intensity was elevated only focally in plaques with a high-risk morphologic phenotype (p<0.05), including OCT fibroatheroma, plaque rupture, and fibroatheroma associated with in-stent restenosis.

CONCLUSIONS

This first-in-human OCT-NIRAF study demonstrates that dual-modality microstructural and fluorescence intracoronary imaging can be safely and effectively conducted in human patients. Our findings show that NIRAF is associated with a high-risk morphologic plaque phenotype. The focal distribution of NIRAF in these lesions furthermore suggests that this endogenous imaging biomarker may provide complementary information to that obtained by structural imaging alone.

INTRODUCTION

Intravascular optical coherence tomography (OCT) is a high-resolution imaging technique that is increasingly being utilized in interventional cardiology for the investigation and management of coronary artery disease (CAD) (1,2). OCT enables the visualization of artery wall microstructure, including morphologic features related to coronary events such as lipid-containing regions, macrophage accumulations, thin-cap fibroatheromas (TCFA), erosion and rupture of coronary plaques, and presence of thrombi and calcified nodules (1,2). Owing to its capability to enable a clear view of the detailed arterial morphology and implanted arterial stents, OCT has also been used to assess the response of the artery wall following percutaneous coronary intervention (PCI) (1,2).

Even though OCT provides an unprecedented level of morphologic detail, it does have limitations that constrain its diagnostic capabilities. Key plaque features such as lipid, for example, manifest as low OCT signal. The use of negative contrast features can confound diagnosis, as signal loss may arise from a variety of different sources such as macrophage shadowing, intraluminal debris, and image artifacts (3,4). Furthermore, the capability of OCT to differentiate non-necrotic intra- and extracellular lipid accumulations from necrotic core lipid has never been shown and therefore remains an unresolved question in the field (2,4). This ambiguity is problematic, as many studies have shown that a definitive diagnosis of necrosis is needed to distinguish the most advanced, progression-prone lesions (5). In addition, microstructure alone does not provide a complete understanding of CAD as the underlying mechanisms of coronary plaque development that lead to disruption and acute thrombosis are multifactorial, involving a complex interaction between structural, compositional and biomechanical characteristics, and cellular and molecular processes in the vessel wall (5).

Fluorescence molecular imaging has been proposed to complement OCT for studying plaque pathobiological mechanisms (6,7). Intravascular near-infrared fluorescence (NIRF) using targeted molecular agents has been shown to elucidate inflammatory activity and fibrin accumulation in mice (8) and rabbit arteries (7,9), but these agents are not yet approved for human use. Detection of fluorescence from naturally occurring molecules, also known as autofluorescence, is closer to clinical application because it can be detected without the administration of exogenous agents. Autofluorescence excited in the UV and the visible portions of the electromagnetic spectrum has been studied in human plaques ex vivo where the signal relates to elastin, collagen, and NADH (10,11). Recently, red-excited (633 nm) near-infrared autofluorescence (NIRAF), with emission detected between 700-900 nm, has been shown in cadaver coronary arteries to be specifically elevated in advanced, necrotic core containing lesions (12), including thin capped fibroatheroma (TCFA), the most common type of plaque implicated in acute coronary syndromes and acute myocardial infarction.

Based on the potential of OCT and NIRAF combination to improve our detection of necrotic core lesions and specifically TCFA, we developed a human-use OCT-NIRAF system and catheter. Here we describe a first-in-human safety and feasibility study of this multimodality intravascular imaging technology in patients and report our findings regarding the spatial distribution of the NIRAF signal with respect to co-localized OCT images of tissue microstructure in vivo.

METHODS

Patients

Patients undergoing PCI at Massachusetts General Hospital (Boston, MA) were enrolled between July 2014 and January 2015. All patients provided informed consent and the study was approved by the Massachusetts General Hospital (Partners Healthcare) institutional review board (IRB).

Dual-modality OCT-NIRAF imaging system

We created a dedicated, dual-modality OCT-NIRAF system that implements state-of-the-art OCT (1250-1370 nm) and NIRAF, excited at 633 nm and detected between 675 and 950 nm (Supplementary Figure 1). This system acquires synchronized OCT and NIRAF data at a rate of 100 frames/second. The OCT-NIRAF coronary imaging procedure is identical to that of current intravascular OCT, where volumetric (3D) OCT-NIRAF data are obtained by rotating and translating the driveshaft at a constant speed producing a helical scan.

Clinical intracoronary OCT-NIRAF imaging

As in routine clinical intravascular OCT imaging (1,2), the dual-modality catheter was advanced over a 0.014-inch guidewire and through a 6-F guide catheter placed into a coronary ostium. The OCT-NIRAF catheter was advanced distal to a lesion of interest and images were acquired during a manual contrast injection at a rate of approximately 3-4 ml/sec for approximately 3 seconds. The driveshaft was retracted at a pullback speed of either 20 or 40 mm/sec. OCT-NIRAF imaging of the vessel undergoing PCI was performed in all cases and additional major coronary vessels were imaged as time permitted.

Data Processing

NIRAF emission intensity data were processed by first subtracting the image background. We then automatically calibrated NIRAF emission intensities based on the distance between the catheter and the artery wall as determined by OCT, so that the fluorescence signal could be quantitatively compared between patients (6,13). Quantitative NIRAF data were normalized between 0 and 1 using the minimum and maximum NIRAF values acquired in the study. NIRAF data was displayed using a linear color look up table (white=high NIRAF, to dark blue=low NIRAF). OCT and NIRAF data were fused with the calibrated, normalized, and color-mapped NIRAF data, presented as an annulus around the gray scale OCT image, in a similar manner to that employed for NIRS-IVUS (14). En face two-dimensional (2D) NIRAF maps were also generated, similar in appearance to a NIRS chemogram, with the catheter's scan angle on the vertical axis and the pullback distance on the horizontal axis. OCT images were displayed using a logarithmic gray-scale look up table (15).

Three-dimensional (3D) reconstructions of OCT-NIRAF data were obtained after frame-to-frame semi-automatic registration to correct for artifacts generated by cardiac motion and mechanical scanning catheter-based acquisition. Plaques were manually segmented and different colors were assigned as follows: artery wall – gray, calcified plaques – white, and lipid-rich plaque – yellow. In addition, stent struts were automatically segmented (16,17) and rendered in red. Processed OCT-NIRAF images were imported into volume rendering software, (OsiriX 6.5.2,The OsiriX Foundation, Switzerland), and 3D data was visualized as volumes (perspective volume rendering) with semi-transparent opacity tables (15). The NIRAF signal was rendered over the artery wall with semitransparent opacity levels, tuned for optimal visualization.

Data analysis

In order to correlate the NIRAF emission intensities with the different plaque features, lesions were manually segmented using standard OCT image interpretation criteria (1,2). Tissue type was categorized as: normal vessel wall, fibrotic, fibrocalcific, thick-cap fibroatheroma (ThCFA) if cap thickness >65 μm, thin-cap fibroatheroma (TCFA) if cap thickness <=65 μm, and plaque rupture. The plaques were selected by an expert OCT image reader (GJU) blinded to NIRAF data to avoid bias. Each plaque included multiple adjacent OCT-NIRAF frames. The maximum NIRAF signal intensity was calculated using all A-scan lines within the plaque. Plaques with a maximum NIRAF intensity >0.4 were arbitrarily classified as having a high NIRAF signal, moderate NIRAF plaques had a maximum signal between 0.2 and 0.4, low NIRAF lesions had a maximum signal between 0.05 and 0.2, and plaques negative for NIRAF were those with a maximum NIRAF signal <0.05. Macrophage accumulations within fibroatheroma caps were quantified using the normalized standard deviation (NSD) parameter (18). NSD values >7 (median value of the NSD range) were considered to be elevated in this study. This analysis was performed on each OCT fibroatheroma frames showing moderate or high NIRAF, and on an equally numbered, randomly selected set of atheroma frames showing low or absent NIRAF. Frames showing OCT image artifacts (e.g., NURD – non-uniform rotational distortion, seamline artifact or blood in the lumen) where excluded from all analyses. NIRAF signal reproducibility among repeated pullbacks from the same coronary segment was assessed by first registering two datasets using anatomical landmarks and known pullback speeds. A one-dimensional pullback plot was then generated for each NIRAF dataset by taking the maximum NIRAF signal for each frame for each pullback position. Reproducibility was quantified using Pearson's correlation coefficient computed from paired, one-dimensional NIRAF pullback datasets.

In order to quantify our observations that NIRAF is only elevated focally in OCT-delineated atherosclerotic plaques, we measured the fibroatheroma arc distribution in degrees as determined by OCT using the lumen's centroid as the origin and repeated the same procedure for NIRAF. These measurements were made for all the plaques imaged in this study that exhibited moderate or high NIRAF emission intensity. The same procedure was used to compare NIRAF and macrophage density (NSD) angular distributions.

Statistics

Statistical analysis reported in this study was performed using Matlab version 8.4.0 (MathWorks®, Natick, MA) and Matlab Statistics Toolbox version 9.1. One-way ANOVA and pairwise Kruskal-Wallis analysis (with Mann-Whitney U tests and without adjustment for multiple comparisons) were used to compare maximum NIRAF between OCT-defined plaque type. No corrections for multiple observations within subjects or vessels were applied. Continuous data are expressed as mean ± standard deviation, or median (interquartile range) when appropriate, and p values <0.05 are considered statistically significant.

RESULTS

At total of 12 patients were imaged using the OCT-NIRAF catheter. Table 1 depicts baseline characteristics of the enrolled patients. A total of 17 major coronary arteries were imaged simultaneously using co-localized OCT and NIRAF, encompassing 33 OCT-NIRAF pullbacks. Good quality OCT-NIRAF datasets (average length: 52±10 mm) were obtained for each patient. The mean number of pullbacks per patient was 2.75±1.23 and the average amount of contrast administered to each patient was 44±26 ml, with a mean of 14±2 ml per OCT-NIRAF pullback. In a substudy of 4 repeated pullbacks, NIRAF reproducibility was excellent with an average Pearson's correlation coefficient of 0.925±0.015 (Supplementary Figure 2). There were no patient complications related to the OCT-NIRAF imaging procedure.

Table 1. Patient baseline characteristics on admission.

All patients presented with stable angina pectoris (SAP), either Class II or Class III. Categorical variables are expressed as numbers (%) and continuous variables are reported as mean ± standard deviation or medians (interquartile ranges) for body-mass index. CCS – Canadian Classification System for stable angina pectoris.

Baseline characteristics	N =12
Mean age (years)	63.3 ± 7.6
Male gender, n (%)	11 (91.2)
Hypertension, n (%) *	11 (91.2)
Family history of premature heart disease, n (%)	3 (25.0)
Prior coronary artery disease	8 (66.6)
Prior myocardial infarction, n (%)	2 (16.6)
Prior PCI, n (%)	5 (41.7)
Smoking status
Never smoked, n (%)	6 (50.0)
Ex-smoker, n (%)	5 (41.7)
Current smoker, n (%)	1 (8.3)
Hyperlipidemia, n (%)	11 (91.2)
Hyperlipidemia currently taking medication, n (%)	9 (75.0)
LDL mg/dL, mean ± SD	99 ± 33
HDL mg/dL, mean ± SD	43 ± 11
Triglycerides mg/dL, mean ± SD	123 ± 63
Body-mass index (kg/m2), n, median, Q1-Q3	28.0 (24.8-30.4)
Diabetes mellitus, n (%)	4 (33.3)
Medication
Statins, n (%)	9 (75.0)
Beta-blocker, n (%)	8 (66.6)
ACEi or A2 receptor blockers, n (%)	6 (50.0)
Nitrates, n (%)	8 (66.6)
Diuretics, n (%)	4 (33.3)
Aspirin or clopidogrel, n (%)	11 (91.2)
Anticoagulant, n (%)	1 (8.3)
Angina pectoris class
CCS I, n (%)	0 (0.0)
CCS II, n (%)	8 (66.6)
CCS III, n (%)	4 (33.3)
CCS IV, n (%)	0 (0.0)
Angiography indication
Chest pain	5 (41.7)
Positive functional study	7 (58.3)
Stented Vessel (PCI)
LAD	4
LCx	3
RCA	5

^*hypertension is defined by systolic blood pressure ≥ 140 mmHg and/or diastolic blood pressure ≥ 90 mmHg.

Description of representative clinical cases

Figure 1 shows an example of an OCT-NIRAF pullback from the right coronary artery (RCA) of a 67-year-old male patient who presented with in-stent restenosis of the mid-RCA. This pullback was acquired adjacent to the stented region located in the distal RCA, along a nonstenotic segment (Figure 1A). OCT imaging revealed a long segment of normal coronary wall and intimal hyperplasia with concomitant minimal NIRAF signal throughout (Figures 1B, D). A small calcific lesion located in the middle of the pullback was NIRAF negative (Figure 1C).

Fig. 1. Coronary segment negative for NIRAF.

(A) Angiography of RCA showing non-significant coronary disease over the OCT-NIRAF pullback segment (ps). (B) 2D NIRAF map demonstrating negligible NIRAF signal. (C) Cross sectional OCT-NIRAF image showing normal coronary wall and a calcification (2 o'clock) with no NIRAF signal detected. (D) 3D cutaway rendering of the OCT-NIRAF pullback. Scale bar in (C) equal to 1 mm; scale bar in (B) equal to 5 mm. *, guide-wire shadowing artifact; ●, side-branch.

Strong NIRAF was only seen in regions of plaque with high-risk structural features (e.g., lipid-containing plaques, thin-fibrous caps, rupture with thrombus) as determined by OCT. Supplementary Figure 4 shows an example of an OCT-NIRAF pullback acquired from the left-anterior descending (LAD) coronary artery of a diabetic 59-year-old male patient, presenting with left-circumflex (LCx) in-stent restenosis. The angiogram demonstrated diffuse atherosclerotic disease in the LAD. OCT imaging revealed multiple lesions in the LAD, including several fibrocalcific and lipid-rich plaques (Supplementary Figures 4C-F, G). A single, focal spot of elevated NIRAF signal was observed in the 2D NIRAF map (Supplementary Figure 4B). The OCT image at this site in the mid LAD indicated the presence of a TCFA with an intact cap. Interestingly, the NIRAF signal was only elevated focally in this OCT-TCFA; in the cross-section displayed in Supplementary Figure 4F, high NIRAF was located at 9 o'clock, whereas the thin cap extended over a much larger arc, from 7 to 2 o'clock. All other plaques in this artery were OCT-delineated as fibrotic (Supplementary Figure 4C), fibrocalcific (Supplementary Figure 4E), or ThCFA (Supplementary Figure 4D), and exhibited negligible or low NIRAF signal.

High NIRAF was also found in arterial sites that contained OCT-evidence of plaque disruption/erosion and overlying thrombus. Figure 2 shows an example of a TCFA rupture located in the LAD of a 66 year-old diabetic patient, presenting with significant stenosis of the proximal LCx (treated artery) and moderate stenosis of the proximal LAD. OCT imaging of the ostial LAD revealed a TCFA with cap rupture (Figures 2C-H), including a platelet-rich thrombus overlying the site of cap disruption (Figures 2D, G). NIRAF was elevated focally within the OCT-TCFA at the rupture site (Figures 2E, H) and at an adjacent location that contained a cholesterol crystal (Figures 2C, F). In this pullback, OCT revealed several other lipid-containing plaques (Figure 2I) that were all negative for NIRAF.

Fig. 2. OCT-NIRAF imaging of TCFA rupture.

(A) Coronary angiogram of LAD and (B) 2D NIRAF map showing a focal region of elevated NIRAF in the ostial LAD. (C, D, E) OCT-NIRAF cross-sections from sites in (B) with elevated NIRAF, revealing subclinical OCT-TCFA fibrous cap rupture. (F) Magnification of a cholesterol crystal below the cap, co-localized with high NIRAF, and (G, H) magnified views of the rupture site. In (G), the rupture site (arrowhead) is covered by a small white luminal thrombus (arrow) and the arrow in (H) points to the site of the thin-cap rupture, demonstrating co-localized and very high focal NIRAF signal. (I) 3D cutaway rendering showing that the highest NIRAF spot appears focally within a large lipid pool (arrow), and the remaining portion of the vessel shows diffuse disease that was negative for NIRAF. Scale bars on OCT images and magnifications are equal to 1 mm and 0.5 mm, respectively; scale bar in (B) is equal to 5 mm. ps, pullback segment; L, lipid; R, rupture site; T, thrombus.

The OCT-NIRAF catheter was also used to image early in-stent restenosis (86 days after BMS implantation) in the left-circumflex coronary artery (LCx) of the diabetic patient described in section 3.2. Because of severe restenosis (70%), OCT-NIRAF was conducted after balloon pre-dilatation. In the proximal and middle region of the stented segment, OCT imaging revealed that a portion of the stent was deployed overlying a large fibroatheroma (Figure 3). An elevated NIRAF signal was observed in this area (Figure 3B, C). The remaining in-stent restenotic tissue had an OCT appearance of neointimal hyperplasia and exhibited minimal NIRAF signal (Figure 3E).

Fig. 3. OCT-NIRAF imaging of in-stent restenosis.

(A) Angiography of LCx and (B) 2D NIRAF map. (C) Cross-sectional OCT-NIRAF image, showing a focal site with elevated NIRAF co-localized with stent struts overlying an OCT-delineated fibroatheroma. The arrow in (D) identifies OCT uncovered stent struts, a marker of incomplete stent healing. (E) Cross-sectional image from the distal portion of the stent that is negative for NIRAF, showing in-stent restenosis and non-attenuating OCT tissue suggesting intimal hyperplasia and presence of fibrotic tissue. (F) 3D cutaway rendering illustrating that the highest NIRAF signal is co-localized with mid and proximal stent segment and with the tissue with high OCT signal attenuation (fibroatheroma). Scale bars on OCT images are equal to 1 mm; scale bar on (B) is equal to 5 mm. ps, pullback segment; L, lipid.

OCT-NIRAF quantitative analysis

Patient imaging findings revealed NIRAF emission intensity patterns that were distinct from structural or compositional patterns seen with other intravascular coronary imaging modalities (e.g. OCT, IVUS, NIRS) in vivo. From the 17 coronary arteries imaged in this study, a high focal NIRAF signal was found in 5 arteries (29%), a low to moderate NIRAF signal was found in 4 arteries (24%) while all the other arteries (n=8, 47%) were NIRAF negative. Following analysis of each pullback, a total of 79 distinct plaques were identified as described in the methods section. The maximum NIRAF signal intensity of each plaque was compared to OCT-defined type (Figure 4). There was a statistically significant difference of maximum NIRAF signal between plaque types (one-way ANOVA, p<0.0001). OCT-delineated TCFA and plaque rupture cases demonstrated much higher maximum NIRAF signal than all other plaque (p<0.05). All other groups were statistically different from each other (p<0.05), with the exception of fibrocalcific plaques and ThCFA (p=0.65). While there were only two plaque ruptures identified in this cohort, the maximum NIRAF signal for rupture sites was non-significantly higher than that of unruptured TCFA (p=0.07) (Figure 4).

Fig. 4. Analysis of maximum NIRAF plaque intensities for different plaque types.

Values from normal vessel wall and fibrotic plaques were significantly different from each other (p<0.05), demonstrating a lower signal intensity than that of fibrocalcific plaques (p<0.05). Fibrocalcific plaques and ThCFA showed moderate maximum NIRAF signal intensities that were not significantly different from each other (p=0.65). A higher maximum NIRAF signal was detected from sites of plaque rupture (p<0.05) and TCFA (p<0.05). Whisker lengths are defined as +/−2.7σ, and points are drawn as outliers if outside the range [q1 – w(q3 – q1)] and [q3 + w(q3 – q1)], where q1 and q3 are the 25th and 75th percentiles, respectively.

By comparing OCT and NIRAF lipid arcs, we observed that NIRAF contiguously spanned 23.2%±13% (mean ± standard deviation) of the total arc of lipid in OCT fibroatheroma. These findings indicate that in all areas where NIRAF was elevated, it only peaked at specific discrete loci.

OCT-macrophage accumulation analysis (92 OCT-fibroatheroma frames analyzed) demonstrated that NIRAF is focally associated with OCT measurement of macrophage-rich inflammation, as estimated by the NSD parameter (18). When NIRAF was high, it was always found in areas of elevated NSD (Supplementary Figure 3) but not all areas of high NSD exhibited high NIRAF. Furthermore, NIRAF was only elevated discretely in regions of elevated NSD as NIRAF covered 39.2%±12.5% of the elevated NSD arc.

DISCUSSION

In this paper, we present a first-in-human investigation of multimodality intracoronary OCT and near-infrared autofluorescence (NIRAF) imaging in vivo. Our results show that OCT-NIRAF uncovers a unique autofluorescence signature of human coronary atherosclerosis in vivo. The intracoronary NIRAF-OCT procedure was safe and was utilized in patients similarly to conventional intravascular OCT. Importantly, co-registration and distance correction of OCT and NIRAF data was inherently automatic and facilitated image interpretation. Data acquisition was reliable and NIRAF was found to be reproducible amongst multiple pullbacks.

By investigating the spatial relationship between NIRAF and arterial morphological features in vivo, we found that elevated NIRAF was associated with morphologic and/or mechanistic features of high plaque risk. NIRAF was negative or low in plaques with a low-risk microstructural phenotype as determined by OCT (intimal hyperplasia, fibrous plaque, and fibrocalcific plaque). In contrast, NIRAF was high focally in certain OCT-delineated fibroatheromas and highest in OCT-delineated TCFA, regions of cap disruption, and areas of fibroatheroma associated with in-stent restenosis. Results from this study also associate elevated NIRAF with indicators of inflammation. NIRAF was found to be only elevated in plaque regions that showed OCT evidence of macrophage accumulations. In combination with our previous ex vivo study (12), these findings support our original rationale for combining OCT and NIRAF to better delineate and characterize atherosclerotic lesions that are at risk of progression. Although the correlation shown here between NIRAF and plaque inflammation has been established using an objective method for OCT image quantification (18,19) these results should be considered as hypothesis generating. Plaque components other than macrophages give rise to an appearance of punctate high OCT signal regions and OCT has not been demonstrated to distinguish between active and inactive macrophages and other macrophage subtypes (20).

The focal spatial distribution of the NIRAF signal was a major unexpected finding of this study that supports the potential additive nature of this imaging biomarker. Elevated NIRAF signal occurred only at discrete locations and for example did not subtend the entire fibroatheroma or the entire arc of thin fibrous caps. In addition, NIRAF was only high focally in regions with OCT evidence of high macrophage accumulations, however the converse was not true in that there were many elevated NSD areas that were NIRAF negative. The observation that NIRAF is only focally elevated in plaques with macrophage accumulations is consistent with concept of multiple macrophage phenotypes in atherosclerotic lesions.

Additional studies are merited to uncover the specific molecular/chemical mechanisms that produce NIRAF in atherosclerotic plaque to fully understand the biological and clinical relevance of this signal. Based on the findings of this study and information gleaned from the literature (21,22), various hypotheses can be made about the potential sources of NIRAF in atherosclerotic plaques and its potential relationship to coronary wall inflammation. NIRAF in atherosclerotic plaque may arise from the modification of lipids and lipoproteins by oxidative stress (21,22). Studies have shown that ceroid, a protein-lipid oxidation byproduct found in atherosclerotic plaque, has a yellow fluorescence spectrum (22) and it is possible that the tail of ceroid's fluorescence emission may extend into the NIR. Oxidative stress caused by inflammatory activity may also cross-link surrounding proteins, creating dimers such as dityrosine that are fluorescent in the NIR (21). Other reports in cancerous tissues (23) have observed tissue autofluorescence, suggesting that it could arise from endogenous porphyrins. Extrapolating to atherosclerosis, it is possible that intraplaque hemorrhage that contributes to lesion development and destabilization may give rise to porphyrins that could in part explain the NIRAF signal observed here.

While the focal appearance of the NIRAF signal could be due high biological/spatial specificity of the molecular/chemical entity generating the fluorescence, it also could in part be influenced by the spatial locations of the fluorophores and NIRAF light propagation given the surrounding tissue's optical properties. To better characterize the autofluorescence signal detected in vivo, OCT-NIRAF instrumentation can be augmented by detecting the emitted light using a spectrometer rather than a single integrating detector. Use of a spectrometer will allow for the analysis of the spectral content of the NIRAF emission signal that may help to determine its molecular/chemical origins. Furthermore, spectral detection can potentially be used to correct for variations of NIRAF signal intensity that are due to the location of the fluorophores in the artery wall and the propagation of NIRAF light through plaque, which may have heterogeneous optical properties.

Study limitations

In this study, we imaged a limited number of patients with relatively low risk clinical presentations (Canadian Classification System Class II and Class III angina pectoris). Intravascular OCT-NIRAF imaging studies in larger cohorts that include ACS and STEMI patients will be required in order to confirm and expand our findings. Although our data allows a preliminary assessment of the associations between the NIRAF signal and plaque morphology and microstructure, the small size of this study and our inability to acquire specimens for advanced tissue analysis prevent us from making definitive conclusions about the biological or molecular nature of the NIRAF signal. While these results demonstrate the potential for this NIRAF plaque signature to refine the identification of high-risk plaques in vivo, the applicability of these new lesion characteristics to clinical screening and coronary event prediction remain to be demonstrated as well as the accuracy in assigning an increased risk to plaques with high NIRAF.

CONCLUSION

In our first-in-human experience with intracoronary OCT-NIRAF, we have demonstrated that a unique human coronary autofluorescence signature can be detected in CAD patients in vivo. The multimodal OCT-NIRAF structural and fluorescence intracoronary imaging can be conducted in patients with similar ease and safety as that of conventional, standalone intravascular OCT. Findings show that NIRAF is focally elevated in plaque locations where most high-risk morphologic phenotypes are evident. Future investigations will elucidate the specific molecular nature of the NIRAF signal and its pathobiological and clinical relevance. In addition to NIRAF, this work paves the way for demonstrating intravascular OCT and targeted molecular fluorescence in human patients. Such multimodality technologies that combine microstructural and fluorescence imaging will hopefully further expand our armamentarium of tools for coronary plaque diagnosis, improving our capacity to predict plaque progression and refine patient and lesion-specific risk.

Clinical Perspectives.

Competency in Medical Knowledge:

A first-in-human study with a multimodality OCT and near-infrared autofluorescence (NIRAF) imaging system and 2.6 F coronary catheter has been conducted. Results showed that co-localized and simultaneously acquired OCT and NIRAF pullback datasets could be safely obtained in human coronary arteries in vivo in a few seconds during a brief, non-occlusive contrast injection. Elevated NIRAF signal was found to be associated with a high-risk morphological phenotype, as determined by OCT.

Translational Outlook:

Our results suggest a potential role for intravascular OCT-NIRAF in improving our capability to detect high-risk plaques. Additional studies are required to confirm these initial findings and to determine the molecular/chemical sources of NIRAF in human coronary atherosclerosis.

ABBREVIATION LIST

DCF

double-clad fiber

NIRF

near-infrared fluorescence

NIRAF

near-infrared autofluorescence

NIRS

near-infrared spectroscopy

NSD

normalized standard deviation

NURD

non uniform rotational distortion

OCT

optical coherence tomography

PCI

percutaneous coronary intervention

TCFA

thin-cap fibroatheromas

ThCFA

thick-cap fibroatheroma