Open Access
Add to BibSonomy
Abstract
Epicardial ablation is necessary for the treatment of ventricular tachycardias refractory to endocardial ablation due to arrhythmic substrates involving the epicardium. The human epicardium is composed of adipose tissue and coronary vasculature embedded on the surface and within the myocardium, which can complicate electroanatomical mapping, electrogram interpretation and ablation delivery. We propose using near-infrared spectroscopy (NIRS) to decipher adipose tissue from myocardial tissue within human hearts ex vivo. Histological measurement of epicardial adipose thickness direct correlated (R = 0.884) with the adipose contrast index. These results demonstrate the potential of NIRS integrated catheters for mapping the spatial distribution of epicardial substrates and could aid in improving guidance during epicardial ablation interventions.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Epicardial ablation is a treatment approach for the management of ventricular arrhythmias such as ventricular tachycardia (VT) [1], in which candidate targets for ablation of slowed conduction are post-infarct, scar border zones which constitute the substrate for tachycardia-inducing re-entrant circuits located on the epicardium [2,3]. Previous reports estimate that about a third of VT presentations fail treatment by endocardial ablation, due in part to origins extending mid-wall or epicardially, which render them inaccessible from the endocardium alone [4,5]. While epicardial ablation procedure is recommended in such instances, differences between endo- and epicardial structures, such as lipid distribution and vascularity, can complicate ablation delivery. The presence of epicardial fat has been shown to limit radiofrequency (RF) energy penetration and can lead to ambiguities in electrograms that mimic scar or lesions [6–8]. Moreover, due to the exposure of epicardial vessels, angiograms are routinely performed during catheter navigation to avoid sequelae caused by injury to coronary circulation. A technique that enables tissue discernment at the catheter tip may help improve the safety of therapeutic interventions based on spectral characteristics and preventing unintentional injury of coronary vessels [9].
Conventional MR and CT imaging techniques have been previously employed for quantification of epicardial substrates [10–13]. However, a catheter-based technique may be better suited for real-time, precise evaluation of the tissue prior to ablation. In previous work, optical coherence tomography (OCT) has enabled imaging of epicardial fat in histological detail [14,15]. However, prior studies, indicating the influence of overlying fat on ablation lesions, report significant damping of RF energy penetration in fat layers with thickness >3mm, which exceeds the imaging range offered by OCT in cardiac tissue [16].
Near-infrared spectroscopy (NIRS) is an optical sensing technique influenced by intrinsic signatures of tissue ultrastructure and molecular absorbers. Previously, our group has shown lesion depth tracking up to 4mm using a NIRS-integrated RFA catheter [17]. NIRS can also assess tissue-catheter contact in the presence of blood with high accuracy and discriminate ablated from unablated tissue [17,18]. Furthermore, we have shown unique NIRS spectral signatures for cardiac structural substrates [19]. Importantly, gaps can be readily identified and maps of lesion thickness can be calculated. Marked differences in spectral shape were apparent throughout the entire spectral range and became more prominent with increased lesion depth. Lesion estimation models based on optical indices had high R2 values for lesions created in the right ventricle (R2=0.969) and the LA (R2=0.912) [18]. Also, NIRS has been widely utilized in a variety of medical applications, from standard measurements of pulse oximetry to, more recently, its application toward intravascular discrimination of lipid-core plaques within coronary arteries [20,21]
We hypothesize that NIRS could aid in mapping structures over the epicardial surface for which we developed a custom NIRS-integrated ablation catheter. The optical configuration is designed in such a way to bias measurement influence toward attenuation by molecular absorbers. A parameter, the adipose contrast index (ACI), is introduced to estimate the epicardial fat thickness from NIRS measurements. Three-dimensional mappings of NIRS-extracted ACI and radiofrequency ablation lesions are demonstrated within human hearts ex vivo.
2. Methods
2.1 Experimental protocol
Human donor hearts (n=12) were acquired from the National Disease Research Interchange (NDRI, Philadelphia, PA). NDRI received consent for all donors. All specimens were de-identified and considered not human subjects research, according to the Columbia University’s Institutional Review Board under 45 CFR 46. The inclusion criteria for the NDRI protocol incorporates both healthy hearts and diagnosis of end stage heart failure, cardiomyopathy, coronary heart disease, atrial fibrillation, and myocardial infarction. Table 1 shows a summary of donor medical histories for the hearts used in this study. Experiments were conducted within 24-48 hrs following donor’s death. Ventricular halves were dissected and submerged under temperature-maintained (37°C) phosphate-buffered saline using a circulating water-bath system (SL-12, JULABO, Allentown, PA). A commercial irrigated RFA catheter (Thermocool, Biosense Webster, USA), a generator (Stockert 70, Biosense Webster, USA) and an irrigation system (CoolFlow, Biosense Webster, USA) were utilized to visualize lesion sets over the epicardial surface. To create lesions of different sizes, we varied ablation duration between 10-60s, while power and flow rate settings were fixed at 30W and 5mL/min, respectively. The impact of force was not considered. Following lesion delivery, hearts were 3D-scanned (Einscan-SE, Shining3D Technology Inc., San Francisco, CA) to obtain a point cloud of the ventricular surface. After scanning, the specimen was situated on a platform for subsequent epicardial NIRS sampling with camera tracking of sampled sites. An average of 68 optical spectral measurements across the epicardial surface was made per heart. Within one heart, 202 optical spectral measurements were obtained to evaluate the impact of increased spatial sampling.
Table 1. Donor Medical Historiesa (n=12)
2.2 Optical catheter designs
Figure 1(a) shows a schematic illustration of the NIRS-integrated RFA catheter. The system has been previously described in [17] and is briefly described below. A custom aluminum tip electrode was fabricated to accommodate NIRS illumination and collection optical fibers and was integrated with a commercial RFA catheter. The final diameter of prototype catheter was 12 Fr. The distance between the illumination and collection fibers was adjusted to bias measurement influence toward tissue absorption and minimize the influence of scattering. A broadband lamp (HL-2000HP, Ocean Optics Inc, Dunedin, FL) was used for tissue illumination and the diffusely backscattered collected light from the tissue was recorded by a spectrometer (600-1000nm) (C9405CB, Hamamatsu, Bridgewater, NJ).
Fig. 1. Near-infrared spectroscopy (NIRS) integrated catheters for mapping. A: optical fibers integrated into the catheter tip allow for tissue illumination and sampling of tissue backscattered light. The source-detector separation (SDS) is a design parameter, which influence the amount of collected light and its sensitivity to absorbing molecules. B: shows NIR reflectance spectra normalized at 960 nm for sites absent of and richly coated with epicardial fat (EF) for a given sample. Also plotted is the absorption spectrum of lipid obtained from literature [22]. Black arrow indicates a local minima centered at 930 nm reflectance corresponding to a strong lipid absorption peak (red arrow).
2.3 NIRS model-based processing
NIR spectral measurements were calibrated into relative reflectance (${R_{Rel}}$) in a similar manner that has been reported previously [17,18]. This process included dark subtraction, system-response correction, and normalization from a measurement taken on a phantom of known optical properties. Figure 1(b) shows representative the absorption spectrum of fat along with NIRS measurement for a lipid-rich and lipid-scarce area. The local minima in the ${R_{Rel}}$ spectra centered at 930nm reflects a relative decrease in collected light due to strong attenuation of the signal by lipid. Based on changes in spectral morphology, we introduced a new parameter, the adipose contrast index (ACI), observed to accentuate lipid regions. The parameter is calculated based on spectral characteristics over the lipid absorption region (910 - 950nm), as described in Eq. (1):
2.4 Point-cloud co-registration
Following RFA and prior to optical NIRS mapping, ventricular halves were 3D scanned to generate a point cloud (a set of data points with corresponding to 3-D spatial coordinates) of the epicardial surface topology (Fig. 2). After scanning, optical measurements were taken across the epicardial surface with simultaneous camera tracking of sampled locations. A set of affine transformations was applied to optimize the orientation of the point cloud, such that the 2D projection matched the camera image. This permitted a linear transfer relationship between camera image pixels’ coordinates and those of the 2D projection of the point cloud. NIRS-derived parameters for each measured site was then interpolated (bi-harmonic) over a structured grid enclosed by the convex hull of the measurement sites and mapped onto the 3D surface. This process functioned as a surrogate for clinical mapping systems, which track the position electrical measurement and create a shell using measured and interpolated values. All processing was performed in MATLAB (The Mathworks Inc., Natick, MA)
Fig. 2. Experimental workflow for near-infrared spectroscopic epicardial mapping. Donor hearts are ablated, then 3D scanned to obtain a point cloud of the epicardial surface geometry. Optical measurements taken over the surface are tracked using a camera. Using this position information, optical parameters are then co-registered with the 3-D scanned topology to generate NIRS epicardial maps. This workflow was used as a surrogate to a clinical mapping system to localize catheter position with respect to the cardiac surface.
2.5 Histopathology
Following NIRS sampling, hearts were fixed in formalin, then cut obliquely across the optically sampled region. These slices were photographed for macroscopic assessment and paraffin-embedded for subsequent histological processing. Five micrometer thick serial sections were cut and stained with hematoxylin and eosin (H&E) and Masson’s trichrome. Stained sample slides were digitized under 20X magnification and analyzed using Aperio ImageScope (Leica Biosystems, Buffalo Grove, IL, USA). Slide images were reviewed by a board-certified pathologist and local epicardial fat thickness was measured for correlational purposes (Fig. 3).
Fig. 3. Histological measurement of epicardial fat layer thickness. A: gross pathology cross-section showing epicardial fat (EF) and adjacent myocardium (M) layers. B: shows corresponding trichrome histology and measurement of EF thickness. C: shows a magnification of the box in B showing the appearance of fat in microscopic detail.
2.6 Statistical analysis
Correspondence between ACI values extracted over measured and interpolated sites and histologically-derived fat layer thickness was quantified using the Pearson’s correlation coefficient. Significance was marked by p-values less than 0.05. Prism 8 software (Graphpad, San Diego, CA) was used for all statistical analyses.
3. Results
3.1 Three-dimensional ACI maps for lipid distribution and ablation sites
Figures 4 and 5 shows three-dimensional renderings of ACI and $LO{I_1}$ maps. Regions rich in epicardial fat had an orange-yellow appearance on photographed gross pathology and 3D scanned mesh renderings. Furthermore, the spatial deposition of fat was well-represented in ACI maps. Regions without epicardial fat was well-defined in contradistinction to ACI maps (ACI values < 0.4). ACI values over visibly exposed muscle demonstrated low-to-moderate values (0.1-0.4) which may be due to both a thin fat layer or diffusely integrated lipid infiltration within the myocardial wall. For lesion tracking, maps with $LO{I_1}$ values > 1.5 demonstrated the detection of RFA treatment sites as confirmed by trichrome histology (purple hue regions). While the location of lesion sets was identified within spatial maps of the LOI1 parameter, individual lesions were difficult to discriminate. This may perhaps be indicative of a minimum spatial sampling requirement, which is discussed more in detail in the following section. Qualitatively, finer details within both $LO{I_1}$ and ACI distributions were better appreciated within maps rendered from the more densely sampled heart (Fig. 5).
Fig. 4. Proof-of-concept maps of adipose and lesion deposition contrast derived from near-infrared spectroscopy (NIRS) measurements. A: shows a 3D mesh obtained from 3D scanning of the ventricles. B: shows the sampled region of optical measurements tracked by a separate camera. C: shows the spatial distribution of the adipose contrast index (ACI) rendered onto the 3D scanned point cloud topology. D: shows the spatial distribution of the lesion contrast parameter, the lesion optical index (LOI1). E: shows a composite overlay of both adipose and the thresholded lesion contrast maps. A cross-section of the tissue was taken along a region demonstrating a lesion and transition from a thick to thin epicardial fat coating. NIRS maps showed good agreement with gross pathology and histological correlates of treatment and fat thickness.
Fig. 5. High-density mapping of the epicardial surface with 202 measurements. A: shows a 3D mesh obtained from 3D scanning of the ventricles. B: shows the spatial distribution of the adipose contrast index (ACI) rendered onto the 3D scanned point cloud topology. Greater detail in fat distribution could be seen with the increased sampling. C: shows the spatial distribution of the lesion optical index (LOI1). D: shows a composite overlay of both adipose and the thresholded lesion contrast maps.
3.2 Comparison between ACI vs. fat layer thickness
ACI data from derived maps were sampled along linear segments between identifiable landmarks for all hearts similar to that which is shown in Fig. 4(e). These values we’re compared to the corresponding histologically-determined fat layer thickness. A subset of five hearts were used for this study with 31 points sampled from each heart (total of n=155 comparison points). Fat tissue was identified by the regions of honeycomb-like appearance within histological images (Fig. 2(c)). Comparisons for extracted epicardial fat layer thicknesses and ACI values over the same regions are shown in Fig. 6. ACI values were commensurate with fat layer thickness over the range of values studied (Pearson’s, R = 0.884, p<0.0001).
4. Discussion
In this work, we developed a NIRS-integrated catheter and mapping algorithm and demonstrated in preliminary experiments the feasibility of tracking epicardially-distributed adipose, and acute lesions. Electroanatomical mapping of the epicardial substrate is routinely performed to identify ablation targets and evaluate successful lesion delivery. Visceral fat layers of sufficient thickness present as low voltage areas, which can also be misconstrued as post-infarct scar or necrotic lesions, each of which require different responses when encountered [6–8]. Furthermore, fat interposition has been shown to limit radiofrequency energy penetration and thus lesion delivery [8,16]. Prior studies show that layers of fat >3mm in thickness require greater power and irrigation settings to produce comparable lesion sizes compared to thinner fat layer impositions [16]. The capability to differentiate tissue types with ACI and $LO{I_1}$ maps could help resolve ambiguities in electrogram measurement and better inform ablation strategies. Additionally, prior studies have posited that presence of epicardial fat plays a significant role in the pathogenesis of arrhythmias [12,13,23–25]. The proposed device and technique could help to further assess the role of fat and its spatial distribution on the arrhythmias.
Fig. 6. Relationship between adipose contrast index (ACI) and histologically-measured epicardial fat thickness. Line segments across ACI maps were sampled according to corresponding tissue dissected and preserved regions. Histological evaluation of fat thickness correlated linearly to ACI values (Pearson’s, R=0.884, p<0.0001, n=155, 5 hearts, 31 points per heart) up to 10 mm. Solid red and dash black lines represent the mean and 95% confidence interval, respectively.
Previously, quantification of fat volume over the heart surface has been demonstrated using magnetic resonance (MR) and computed tomography (CT)–based imaging techniques [12,13]. While accurate assessment can be performed, these techniques impose the burden of additional time and cost to the procedure and are potentially contraindicated in patients with implantable devices such as pacemakers. They also do not provide real-time feedback for electrophysiologists. ACI values presented in this work were calculated in <0.23ms on average, making the method suitable for real-time, ad-hoc assessment of lipid and lesion extent [17]. Furthermore, distributions can be acquired alongside the initial electrical mapping phase, adding little to no additional time to the procedure.
In prior work, we have demonstrated lesion tracking up to 4mm [17]. Though we have observed good correlation between NIRS-derived contrast parameters and lesion sizes in this range, it is likely that this is beyond the mean sampling depth of collected photons by our system. Instead, we believe this correlation is likely due to an extant relationship between superficial contrast signatures and the extent of ablation treatment. Similarly, in the case of the ACI parameter presented in this work, we believe the correspondence may be due to an analogous relationship between superficial lipid contrast signatures and epicardial lipid volume. Furthermore, another important factor is that lesion and adipose tissue types have different optical properties and spatial distributions, which can also affect the degree to which superficial measurements correlate to deeper.
In some cases, substantial $LO{I_1}$ values were observed outside of RFA regions. A likely explanation for this could be due to the fact that the dominant contrast mechanism for lesion detection in the $LO{I_1}$ parameter is the transition of myoglobin into its oxidated (metmyoglobin) and denatured (hemichrome) states [17]. These globin derivatives are present in both ablated and naturally decomposing post-mortem muscles tissue, thus the influence of latter is expected to be minimal in the clinical setting. During epicardial ablation procedure, near-infrared spectra can be impaired by presence of blood. In previous studies, we have shown that our optical measurements can discern direct catheter-tissue contact in measurements made with blood [17,18]. This condition may differ within blood-perused tissues encountered in vivo. In such case, LOIs may need to be modified to account for different physiological state. Another limitation is that measurement co-registration was performed using 3D scanning and camera tracking, which functioned as a stand-in for a mapping system. Further studies may be needed to verify the renderings of ACI and $LO{I_1}$ maps produced using MRI, or mapping systems such as EnSite (Abbot, Illinois, USA) or Carto (Biosense Webster, Diamond Bar, CA). Additionally, experiments were conducted in ex vivo human donor hearts. The comparisons between ACI values of lipid thick regions and electrogram voltage presentations could not be made. In this work, though $LO{I_1}$ maps were able to identify the general area of lesion delivery, we did not observe a clear differentiation between ablation sites due to spatial sampling limitations. Further studies are needed to assess the role of sampling density on the presentation of lesions within $LO{I_1}$ maps. Indeed, such experiments could be informative on designing optimal mapping catheters incorporating NIRS for lesion gap identification. Prior studies using fine spatial sampling of NIRS measurements demonstrated a gap detection of <1mm [17]. Moreover, radiofrequency ablation treatment near coronary vessels can risk vessel trauma, which can lead to downstream sequela [9]. Future work will be aimed at incorporating NIRS-detection of high-risk vasculature, either by direct assessment of hemoglobin content or assessment of spectral dynamics driven by arterial perfusion. This could help to reduce the need for repeated angiograms when positioning catheters to avoid vessel injury. Additional work is also needed to explore the feasibility of optical substrate mapping using catheters with smaller tip diameters and source-detector separations.
5. Conclusion
In summary, we present a catheter and treatment approach to obtain NIRS-derived parameter maps for the epicardial surface. The technique is shown to be capable of mapping epicardial fat distributions and acute lesion delivery in human ventricular samples ex vivo. An adipose contrast parameter is introduced, which is shown to correlate with histological fat layer thickness. Direct NIRS assessment of tissue types at the catheter tip could serve as an adjunct improvement in ablation targeting and lesion set validation, alongside current practices.
Funding
National Institute of Health (1DP2HL127776-01, 1R01HL149369-01, 4DP2HL127776-02, K08HL122526); Directorate for Engineering (1454365).
Acknowledgments
The authors would like to thank James McLean and Theresa Lye for the helpful discussion regarding the mapping approach.
Disclosures
The authors declare no conflicts of interest.
References
1. E. Sosa, M. Scanavacca, A. D’Avila, and F. Pilleggi, “A New Technique to Perform Epicardial Mapping in the Electrophysiology Laboratory,” J. Cardiovasc. Electrophysiol. 7(6), 531–536 (1996). [CrossRef]
2. K. Soejima, W. G. Stevenson, J. L. Sapp, A. P. Selwyn, G. Couper, and L. M. Epstein, “Endocardial and epicardial radiofrequency ablation of ventricular tachycardia associated with dilated cardiomyopathy: The importance of low-voltage scars,” J. Am. Coll. Cardiol. 43(10), 1834–1842 (2004). [CrossRef]
3. S. R. Dukkipati, A. D’Avila, K. Soejima, R. Bala, K. Inada, S. Singh, W. G. Stevenson, F. E. Marchlinski, and V. Y. Reddy, “Long-term outcomes of combined epicardial and endocardial ablation of monomorphic ventricular tachycardia related to hypertrophic cardiomyopathy,” Circ.: Arrhythmia Electrophysiol. 4(2), 185–194 (2011). [CrossRef]
4. W. G. Stevenson, D. J. Wilber, A. Natale, W. M. Jackman, F. E. Marchlinski, T. Talbert, M. D. Gonzalez, S. J. Worley, E. G. Daoud, C. Hwang, C. Schuger, T. E. Bump, M. Jazayeri, G. F. Tomassoni, H. A. Kopelman, K. Soejima, and H. Nakagawa, “Irrigated radiofrequency catheter ablation guided by electroanatomic mapping for recurrent ventricular tachycardia after myocardial infarction the multicenter thermocool ventricular tachycardia ablation trial,” Circulation 118(25), 2773–2782 (2008). [CrossRef]
5. F. Sacher, U. B. Tedrow, M. E. Field, J. M. Raymond, B. A. Koplan, L. M. Epstein, and W. G. Stevenson, “Ventricular tachycardia ablation: evolution of patients and procedures over 8 years,” Circ.: Arrhythmia Electrophysiol. 1(3), 153–161 (2008). [CrossRef]
6. S. Dixit, N. Narula, D. J. Callans, and F. E. Marchlinski, “Electroanatomic Mapping of Human Heart: Epicardial Fat Can Mimic Scar,” J. Cardiovasc. Electrophysiol. 14(10), 1128 (2003). [CrossRef]
7. B. Desjardins, F. Morady, and F. Bogun, “Effect of epicardial fat on electroanatomical mapping and epicardial catheter ablation,” J. Am. Coll. Cardiol. 56(16), 1320–1327 (2010). [CrossRef]
8. S. Huang and J. Miller, Catheter Ablation of Cardiac Arrhythmias, 4th ed. (elsevier, 2019).
9. J. F. Viles-Gonzalez, R. De Castro Miranda, M. Scanavacca, E. Sosa, and A. D’Avila, “Acute and chronic effects of epicardial radiofrequency applications delivered on epicardial coronary arteries,” Circ.: Arrhythmia Electrophysiol. 4(4), 526–531 (2011). [CrossRef]
10. L. D’Errico, F. Salituri, M. Ciardetti, R. Favilla, A. Mazzarisi, G. Coppini, C. Bartolozzi, and P. Marraccini, “Quantitative analysis of epicardial fat volume: Effects of scanning protocol and reproducibility of measurements in non-contrast cardiac CT vs. coronary CT angiography,” Quant. Imaging Med. Surg. 7(3), 326–335 (2017). [CrossRef]
11. G. Milanese, M. Silva, L. Bruno, M. Goldoni, G. Benedetti, E. Rossi, C. Ferrari, L. La Grutta, E. Maffei, P. Toia, E. Forte, R. C. Bonadonna, N. Sverzellati, and F. Cademartiri, “Quantification of epicardial fat with cardiac CT angiography and association with cardiovascular risk factors in symptomatic patients: From the ALTER-BIO (alternative cardiovascular bio-imaging markers) registry,” Diagnostic Interv. Radiol. 25(1), 35–41 (2019). [CrossRef]
12. O. Batal, P. Schoenhagen, M. Shao, A. E. Ayyad, D. R. Van Wagoner, S. S. Halliburton, P. J. Tchou, and M. K. Chung, “Left Atrial epicardial adiposity and Atrial fibrillation,” Circ.: Arrhythmia Electrophysiol. 3(3), 230–236 (2010). [CrossRef]
13. S. Nakamori, M. Nezafat, L. H. Ngo, W. J. Manning, and R. Nezafat, “Left Atrial Epicardial Fat Volume Is Associated With Atrial Fibrillation: A Prospective Cardiovascular Magnetic Resonance 3D Dixon Study,” J. Am. Heart Assoc. 7(6), e008232 (2018). [CrossRef]
14. C. P. Fleming, K. J. Quan, and A. M. Rollins, “Toward guidance of epicardial cardiac radiofrequency ablation therapy using optical coherence tomography,” J. Biomed. Opt. 15(4), 041510 (2010). [CrossRef]
15. C. P. Fleming, J. Eckert, E. F. Halpern, J. A. Gardecki, and G. J. Tearney, “Depth resolved detection of lipid using spectroscopic optical coherence tomography,” Biomed. Opt. Express 4(8), 1269 (2013). [CrossRef]
16. A. D’Avila, C. Houghtaling, P. Gutierrez, O. Vragovic, J. N. Ruskin, M. E. Josephson, and V. Y. Reddy, “Catheter ablation of ventricular epicardial tissue: A comparison of standard and cooled-tip radiofrequency energy,” Circulation 109(19), 2363–2369 (2004). [CrossRef]
17. R. P. Singh-Moon, X. Yao, V. Iyer, C. Marboe, W. Whang, and C. P. Hendon, “Real-time optical spectroscopic monitoring of nonirrigated lesion progression within atrial and ventricular tissues,” J. Biophotonics 12(4), e201800144 (2019). [CrossRef]
18. R. P. Singh-Moon, C. C. Marboe, and C. P. Hendon, “Near-infrared spectroscopy integrated catheter for characterization of myocardial tissues: preliminary demonstrations to radiofrequency ablation therapy for atrial fibrillation,” Biomed. Opt. Express 6(7), 2494 (2015). [CrossRef]
19. S. Y. Park, R. P. Singh-Moon, E. Y. Wan, and C. P. Hendon, “Towards real-time multispectral endoscopic imaging for cardiac lesion quality assessment,” Biomed. Opt. Express 10(6), 2829 (2019). [CrossRef]
20. R. D. Madder, M. Khan, M. Husaini, M. Chi, S. Dionne, S. Vanoosterhout, A. Borgman, J. S. Collins, and M. Jacoby, “Combined near-infrared spectroscopy and intravascular ultrasound imaging of pre-existing coronary artery stents: Can near-infrared spectroscopy reliably detect neoatherosclerosis?” Circ. Cardiovasc. Imaging 9(1), e003576 (2016). [CrossRef]
21. A. M. Fard, P. Vacas-Jacques, E. Hamidi, H. Wang, R. W. Carruth, J. A. Gardecki, and G. J. Tearney, “Optical coherence tomography – near infrared spectroscopy system and catheter for intravascular imaging,” Opt. Express 21(25), 30849–30858 (2013). [CrossRef]
22. S. L. Jacques, “Optical properties of biological tissues: A review,” Phys. Med. Biol. 58(11), R37–R61 (2013). [CrossRef]
23. T. De Coster, P. Claus, G. Seemann, R. Willems, K. R. Sipido, and A. V. Panfilov, “Myocyte remodeling due to fibro-fatty infiltrations influences arrhythmogenicity,” Front. Physiol. 9, 1381 (2018). [CrossRef]
24. T. De Coster, P. Claus, I. V. Kazbanov, P. Haemers, R. Willems, K. R. Sipido, and A. V. Panfilov, “Arrhythmogenicity of fibro-fatty infiltrations,” Sci. Rep. 8(1), 2050 (2018). [CrossRef]
25. R. Samanta, J. Pouliopoulos, A. Thiagalingam, and P. Kovoor, “Role of adipose tissue in the pathogenesis of cardiac arrhythmias,” Hear. Rhythm 13(1), 311–320 (2016). [CrossRef]
