Intracoronary near-infrared spectroscopy: an overview of the technology, histologic validation, and clinical applications

Intracoronary near-infrared spectroscopy (NIRS) imaging, which is now clinically available in a combined NIRS and intravascular ultrasound catheter, is a novel catheter-based imaging modality capable of identifying lipid core plaque within the coronary arteries of living patients. The present manuscript provides an overview of intracoronary NIRS imaging with a focus on several concepts essential to individuals seeking to better understand this novel imaging modality. One of the major assets of NIRS is that it has been rigorously validated against the gold standard of histopathology and has been shown to accurately identify histologically-proven fibroatheroma. Clinical studies of NIRS have demonstrated its ability to accurately identify large lipid core plaques at culprit lesions across the spectrum of acute coronary syndromes. NIRS has also been shown to detect lesions at increased risk of causing peri-procedural myocardial infarction during PCI. With regards to predicting future risk, NIRS is seemingly capable of identifying vulnerable patients at increased risk of experiencing subsequent patient-level cardiovascular events. In addition to these clinical applications of NIRS, there are several large prospective observational studies underway to determine if NIRS imaging will be able to identify vulnerable plaques at increased risk of triggering site-specific future coronary events. These studies, once completed, are anticipated to provide valuable data regarding the ability of NIRS imaging to identify plaque vulnerability.


INTRODUCTION
Near-infrared spectroscopy (NIRS) is a widely used technique in analytical chemistry to identify organic substances. NIRS can be simplified into three basic components: (1) a light source; (2) a light detector; and (3) a computer capable of deciphering exiting light into clinically relevant information 1 . NIRS utilizes the attenuation, scatter, and absorption of light emitted at wavelengths of 700-1000 nm to detect a unique optical signature produced by any given unknown molecule 2 . NIRS was first applied in vivo in 1977 by Jobsis et al who devised a method to monitor oxygenation by detecting the particular optical signature of oxygenated and de-oxygenated hemoglobin via NIRS 3 . Since that time, NIRS has been described in a multitude of other in vivo applications, including the recent application of NIRS to detect lipid core plaque (LCP) in the coronary arteries of living patients. Although intracoronary NIRS imaging has been recently reviewed in detail elsewhere [4][5][6] , we provide an overview of intracoronary NIRS imaging with a focus on three aspects essential to individuals seeking to better understand this novel imaging modality: (1) NIRS technology and its interpretation; (2) studies validating NIRS findings against the gold standard of histopathology; and (3) clinical applications of NIRS imaging in contemporary practice.

Intracoronary NIRS imaging device
The intracoronary NIRS imaging catheter (TVC Insight Catheter, Infraredx, Burlington, Massachusetts) has applied the same principles of NIRS commonly used in analytic chemistry to identify lipid-rich coronary plaques in vivo. The current NIRS imaging system enacts a dual-modality catheter that houses both NIRS and intravascular ultrasound (IVUS) technologies to provide the user with information on both the composition and structure of plaques within the coronary arteries [7][8][9] . Using a 6 French or larger guide catheter, a combined NIRS-IVUS catheter is manually advanced into a coronary artery over a 0.014 inch guidewire. By pressing a start button on the device, the operator initiates an automated pullback and the catheter scans the target artery at a rate of 5 mm per second 7 . During the automated catheter pullback, more than 30,000 NIRS measurements are obtained within each 100 mm segment of vessel that is imaged 7 . Upon completion of the automated pullback, a NIRS ''chemogram'' is generated and displayed on the imaging console.

The NIRS chemogram, block chemogram, and lipid burdens NIRS chemogram
Following automated pullback of the NIRS catheter within the target artery, acquired spectra are analyzed to determine the probability of LCP presence at each site within the vessel 4-6 . The probability of LCP presence is displayed visually as a chemogram and a block chemogram, which are automatically generated within a few seconds of completing the NIRS pullback in the artery. The chemogram is oriented such that longitudinal position in the long axis of the vessel is depicted on the x-axis and circumferential position from 0-360 degrees around the inside of the vessel is depicted on the y-axis. For locations in the artery in which NIRS spectra indicate a low probability (i.e. less than 0.6) of LCP presence, the corresponding pixels are depicted as red. For locations in the artery in which NIRS spectra indicate a probability of LCP exceeding 0.6, the corresponding pixels are depicted as yellow. An example of a NIRS chemogram and a description of its orientation are provided in Figure 1.

NIRS block chemogram
The contemporary NIRS imaging system also provides a block chemogram that can be found beneath the chemogram on the imaging display [4][5][6] . When the imaging software creates the block chemogram, the imaged vessel is divided longitudinally into contiguous non-overlapping 2mm segments termed blocks. Each 2mm block is assigned one of four colors based upon the 90th percentile probability of all NIRS measurements within that 2mm segment of vessel: • Red indicates the probability of LCP is <0.57 • Orange indicates the probability of LCP is 0.57-0.83 • Tan indicates the probability of LCP is 0.84-0.97 • Yellow indicates the probability of LCP is 0.98 or greater An example of a block chemogram is presented in Figure 1.

Lipid core burden index
In addition to providing information regarding the probability of LCP presence at any location within an imaged artery, NIRS is also capable of providing a semi-quantitative estimate of the amount of LCP present within any selected region of interest. This estimate of lipid quantity is termed the lipid core burden index (LCBI) and is calculated as the number of yellow pixels divided by the total valid pixels (red plus yellow) in any region of interest multiplied by 1000 [4][5][6] . The LCBI is reported on a 0 to 1000 scale. In clinical studies of NIRS imaging, lipid burden has frequently been reported as the maximum LCBI in any 4mm section of the artery (maxLCBI 4mm ) [10][11][12][13][14] . Because a 4mm section of vessel is rather narrow, the maxLCBI 4mm metric can be considered a surrogate of the circumferential extent of a lipid core. Accordingly, a plaque having a maxLCBI 4mm of 700 indicates that the lipid core occupies approximately 70% of the circumference of the vessel at that site.

NIRS VALIDATION AGAINST HISTOPATHOLOGY
Histologic evaluation of coronary autopsy specimens represents the gold standard for LCP detection. One of the major strengths of NIRS imaging is that NIRS has been validated against histopathology for the detection of LCP in several studies. In the first study to validate intracoronary NIRS against histology, Gardner et al applied a histologic definition of LCP as a fibroatheroma containing lipid core ≥ 200 µm thick, having a ≥60 • circumferential distribution, and having an overlying fibrous cap thickness <450 microns, a definition mandated by the FDA 15 . In this study, the NIRS block chemogram differentiated segments with and without histologically-proven LCP with an area under the receiver-operator characteristic (ROC) curve of 0.80. This study also demonstrated that a tan or yellow block on the NIRS block chemogram identified a histologically-proven LCP with a specificity of 90%. The high specificity of a tan or yellow block on the NIRS block chemogram for the presence of a histologically-proven fibroatheroma has since been confirmed 16 . NIRS positive lesions have also been demonstrated at autopsy to have larger necrotic cores, more inflammatory cells, and thinner fibrous caps compared to NIRS negative lesions 17 . Two autopsy studies have demonstrated the incremental benefit of combined NIRS-IVUS imaging compared to the use of either NIRS or IVUS alone for the identification of histologically-proven fibroatheroma 16,18 . Compared to superficial attenuation by IVUS, Kang et al demonstrated that a tan or yellow block on the NIRS block chemogram was associated with a similarly high specificity, but a greater sensitivity for fibroatheroma detection 16 . Importantly, the combination of the NIRS and IVUS findings increased the sensitivity of fibroatheroma detection and improved the positive predictive value compared to ether imaging modality alone 16 . Puri et al demonstrated that plaque burden by IVUS and LCBI by NIRS differentiated coronary segments with and without histologically-proven fibroatheroma with similar accuracy 18 . However when plaque burden and LCBI were used together, these combined NIRS-IVUS findings resulted in significant improvement in the accuracy of fibroatheroma detection and were associated with a net reclassification index of 43%. Examples of combined NIRS-IVUS images are provided in Figure 2.

CLINICAL APPLICATION OF NIRS IMAGING
In addition to the use of NIRS imaging in the autopsy studies described above, there are several lines of clinical research that have reported NIRS findings in living patients. In general these NIRS studies have focused on the identification of culprit lesions in patients with acute coronary syndromes (ACS), the detection of lesions at greater risk for periprocedural myocardial infarction during percutaneous coronary intervention (PCI), and the identification of patients at greater risk for future patient-level cardiovascular events. The following sections will describe each of these lines of research in more detail.

NIRS findings at culprit lesions in patients with ACS STEMI
Post-mortem studies of patients suffering fatal myocardial infarction have shown that most culprit lesions are ruptured thin-capped fibroatheromas with overlying thrombus [19][20][21] .
Considering that an essential component of such thin-capped fibroatheromas at autopsy is a large necrotic lipid core [19][20][21]   Thrombolysis in Myocardial Infarction (TIMI) grade 3 flow was established but prior to stenting 11 . NIRS identified a large, nearly circumferential LCP at the STEMI culprit site in the majority of these patients and STEMI culprit segments were found to have maxLCBI 4mm that was 5.8-fold higher than non-culprit segments of the same vessel. Furthermore, the culprit segments had a maxLCBI 4mm that was 87-fold higher than segments from an autopsy control group that were known to be free of LCP by histology.
In this study, a threshold maxLCBI 4mm >400 identified STEMI culprit sites with specificity of 98% and a sensitivity of 85%. The significance of this finding is that the LCP generating the maxLCBI 4mm >400 signal was likely present in the artery well before the onset of the STEMI. Hence, the maxLCBI 4mm >400 signal may be a marker of a vulnerable plaque at increased risk of triggering a future myocardial infarction. A larger, multicenter study confirming these initial NIRS observations in STEMI patients is expected to be published in 2016. An example of typical NIRS findings in a patient with STEMI is provided in Figure 3.

Non-STEMI and unstable angina
NIRS evaluations of culprit lesions in patients with non-STEMI and unstable angina have also been performed 12 . Similar to the findings previously reported in STEMI, patients with non-STEMI and unstable angina frequently have a large LCP detected by NIRS at the culprit site that is characterized by a maxLCBI 4mm ≥ 400 12 . Accordingly, a threshold maxLCBI 4mm ≥ 400 accurately differentiated culprit from nonculprit sites in non-STEMI and unstable angina while retaining the high specificity previously demonstrated in STEMI patients. Interestingly, culprit lesions in non-STEMI more often had a maxLCBI 4mm ≥ 400 rather than a moderate (maxLCBI 4mm 200-399) or small (maxLCBI 4mm <200) lipid burden. In contrast, culprit lesions in unstable angina had lipid burdens that were more evenly distributed among lipid cores that were small, moderate or large. This observation suggests that the clinical presentation of a plaque rupture event may depend not only on the presence of a LCP, but also on the overall burden of lipid at the culprit site 12 . Additional studies are needed to determine how the lipid burden impacts the clinical presentation of a plaque rupture event.

Sudden cardiac death
A small series of patients presenting with sudden cardiac death who were successfully resuscitated and subsequently underwent NIRS-IVUS imaging has been published 22 . In this small series of patients, a maxLCBI 4mm ≥ 400 was detected by NIRS at each culprit site thought to be responsible for triggering ischemia and the resulting ventricular arrhythmia that lead to cardiac arrest. When considering this observation and those previously demonstrated in STEMI, non-STEMI, and unstable angina, NIRS has now been shown to identify large LCP at culprit sites across the entire spectrum of ACS clinical presentations. A larger study is needed to further evaluate the association between NIRS findings and sudden cardiac death.

NIRS to predict peri-procedural myocardial infarction
Peri-procedural myocardial infarction remains commonplace, complicating 3-15% of all PCI procedures 13 . One of the pathophysiologic mechanisms thought to account for some instances of peri-procedural myocardial infarction is embolization of lipid-rich material that is mechanically released from a target lesion during angioplasty or stent placement 23,24 . Hence, the identification of lipid at the target lesion prior to PCI may identify those lesions at greater risk of causing PCI-related complications. Although this concept has been demonstrated previously with IVUS [25][26][27] , optical coherence tomography [28][29][30] and non-invasively with computed tomographic angiography 31 , NIRS may offer an advantage over these techniques owing to its automated ability to identify LCP with high chemical specificity.
In an analysis of cases from the COLOR registry, Goldstein et al studied target lesions undergoing NIRS imaging prior to PCI and demonstrated that lesions having a maxLCBI 4mm ≥ 500 had a 50% rate of peri-procedural myocardial infarction (13). This is in stark contrast to the observation that only 4.2% of target lesions having a maxLCBI 4mm <500 were associated with a peri-procedural infarct. In this study target lesions having a maxLCBI 4mm ≥ 500 prior to PCI had a 12-fold increased risk of developing a peri-procedural myocardial infarction during PCI 13 .
In the CANARY trial, Stone et al confirmed that large LCPs identified by NIRS imaging prior to PCI were significantly associated with a higher risk of peri-procedural myocardial infarction 14 . In this study lesions having a maxLCBI 4mm ≥ 600 were associated with peri-procedural myocardial necrosis in 24.7% of cases. However, the CANARY trial, which randomized target lesions having a maxLCBI 4mm ≥ 600 to undergo PCI either with or without a distal protection filter, failed to show benefit of distal protection. This finding was perhaps unexpected considering evidence from a prior small study that presented convincing evidence of debris capture using distal protection during PCI on large LCP 32 and also highlights the uncertainty regarding methods to reduce the risk of peri-procedural complications during PCI of large LCP. It is unknown if other methods shown to reduce the focal lipid burden, such as intensive statin therapy 33 or mechanical aspiration of lipid content from the target lesion 34 , if performed prior to PCI will reduce the risk of PCI-related complications. Additional study in this area is clearly needed.

NIRS to identify the vulnerable patient
Whereas the identification of large LCP at heightened risk of triggering peri-procedural myocardial infarction represents one potential clinical use of NIRS imaging, another proposed application of NIRS is to identify patients at increased risk of future patient-level major adverse cardiovascular events. In the ATHEROREMO-NIRS study, investigators performed NIRS imaging within a non-culprit vessel after performing PCI elsewhere in the coronary tree. Patients having an LCBI in the non-culprit vessel above the median value of 43 for the population had a 4-fold increased risk of patient-level cardiovascular events during 1-year of follow up compared to those with an LCBI below the median 35 . Although prior studies have demonstrated a clear association of serum lipid content and cardiovascular risk [36][37][38] , the observations of the ATHEROREMO-NIRS study suggests that the burden of lipid deposited within coronary tissues, as now detectable by NIRS imaging, also associates with cardiovascular risk. Additional studies examining the ability of NIRS imaging to identify patient-level vulnerability to cardiovascular events are forthcoming 39 .

NIRS to identify the vulnerable plaque
Given the association of large LCP identified by NIRS with culprit lesions across the spectrum of ACS [10][11][12]22 and the propensity of such large LRP to trigger myocardial infarction at the time of PCI 13,14 , it has been proposed that NIRS imaging may be a means to identify vulnerable plaques at increased risk of triggering future site-specific coronary events 4 . Whereas prior studies have demonstrated the ability of IVUS to detect plaques at increased risk of site-specific future events [40][41][42][43] , it remains unproven if NIRS findings of a large LRP are in fact at increased risk of triggering future events. There are currently at least two large multicenter prospective observational studies testing the hypothesis that NIRS is capable of identifying vulnerable plaques. These studies include the Lipid Rich Plaque study being performed in the United States and Europe and the PROSPECT II study currently underway in Scandinavia 4 . These studies, once completed, are anticipated to provide valuable data regarding the ability of NIRS imaging to identify plaque vulnerability.