laboratory and animal investigation The Role of Laser-Induced Fluorescence in Myocardial Tissue Characterization*

Transcription

1 laboratory and animal investigation The Role of Laser-Induced Fluorescence in Myocardial Tissue Characterization* An Experimental In Vitro Study George E. Kochiadakis, MD; Stavros I. Chrysostomakis, MD; Michael D. Kalebubas, MD; George M. Filippidis, MD; Ioannis G. Zacharakis, MD; Theodore G. Papazoglou, MD; and Panos E. Vardas, MD, PhD Objective: The fluorescence of tissue when stimulated by a laser beam is a well-known phenomenon. The resulting emission spectra depend on the biochemical and structural composition of the tissue. In this study, we exaed the spectra of laser-induced fluorescence emitted by myocardial tissue. Methods: We used an argon-ion laser to stimulate the myocardium of 20 intact sheep hearts. For each spectral emission, we calculated the intensity in specific regions in order to characterize the spectra and to reveal intercavitary and intracavitary morphologic differences. Results: The statistical analysis showed significant differences in the emission spectra intensity between atria and ventricles. The intensity was higher in the atria than in the ventricles (p < 0.001). The atrial emission spectra were morphologically different from those of the ventricles. There was no difference in the intensity or morphology of emission spectra within each chamber. All measurements showed good reproducibility after a short period of time. Conclusions: Laser-induced fluorescence of myocardial tissue seems to have the characteristics necessary for tissue recognition. This might prove useful in identifying cardiomyopathies and transplant rejection, as well as for myocardial mapping, assisting electrophysiologists in discovering fibrotic arrhythmogenic foci. (CHEST 2001; 120: ) Key words: ablation; experimental; fluorescence; heart; histology; laser; mapping; transplantation; ventricular arrhythmias Abbreviations: LA left atrium; LV left ventricle; P1 peak one; P2 peak two; P3 peak three; P4 peak four; RA right atrium; RV right ventricle; V1 valley one; V2 valley two; V3 valley three The fluorescence of tissue when stimulated by a laser beam is a well-known phenomenon. The spectrum of the emission depends on the particular characteristics of the tissue under exaation, on its *From the Cardiology Department (Drs. Kochiadakis, Chrysostomakis, Kalebubas, and Vardas), University Hospital of Crete, Heraklion Crete, Greece; and Institute of Electronic Structure and Laser (Drs. Filippidis, Zacharakis, and Papazoglou) Foundation for Research and Technology, Heraklion, Crete, Greece. Manuscript received May 22, 2000; revision accepted January 3, Correspondence to: Panos E. Vardas, MD, PhD, Cardiology Department, University Hospital of Crete, PO Box 1352, Heraklion Crete, Greece biochemical and structural composition. 1 4 In recent years, spectroscopy, the analysis of the emission spectra of tissue, has been used increasingly for tissue identification (type, composition, discriation between healthy and pathologic) with excellent results. 4 9 In this project, we studied the emission spectra of laser-induced fluorescence of myocardial tissue. We exaed the basic characteristics of the spectra (intensity, morphology) recorded from different anatomic sites. We also assessed the stability of the measurements, a fundamental criterion for any technique with research and clinical applications. CHEST / 120 / 1/ JULY,

2 Tissue Samples Materials and Methods For the purposes of this study, we used hearts from 20 young sheep (age 6 months). The samples were collected immediately after the animals were slaughtered, and the hearts were removed together with the pericardial sac. They were transported in isothermal containers wrapped in plastic membrane to avoid dehydration. The hearts were exaed in the laboratories of the Foundation for Research and Technology within, at most, 2 h of the death of the animal. All the samples appeared healthy to macroscopic exaation, and there were no regions of myocardium with scar tissue (infarcts), as expected from the young age of the animals. Experimental Apparatus The experimental apparatus is shown in Figure 1. An argon-ion laser emitting at nm (Stabilite 2016; Spectra-Physics; Mountain View, CA) was used for excitation. The average power irradiating the samples during the experiments was of the order of 10 milliwatts. The output of the laser was reflected at right angles by a dichroic mirror (99.3% at 45 at 442 nm) and was focused on the input of a step index, multimode optical fiber (SPC500N; Fiberguide Industries; Stirling, NJ). Fluorescence emission was collected by the same fiber, passed through the dichroic mirror and was focused at the entrance slit (100 m) of a 0.10-m spectrograph, employing a 450 groove mm -1 holographic grating. Data acquisition and analysis were performed via an optical multichannel analyzer (Cronin GmbH; Eichenau, Germany), using a diode array detector (RL1024S; EG&G Reticon; Santa Clara, CA). The signal from the diode array detector was fed to a computer for analog-to-digital conversion and further processing. Wavelength calibration was performed with a mercury lamp. A high-pass filter (490 nm; Schott CG optical filter; CVI Laser Corp, Albuquerque, NM) was placed in front of the entrance of the spectrograph in order to isolate the fluorescence signal from reflected laser light. The spectrum obtained during the experiment represented the time-integrated fluorescence of the underlying tissue. The optical multichannel analyzer was operated in the free-scanning mode. Measurements and Data Processing During exaation, the temperature of the samples was that of the laboratory, ie, 22 C to 24 C. The atria and the ventricles were opened along their free wall. The preparations were washed with normal saline solution to remove blood, which can absorb tissue radiation and might thus alter the emission spectrum. 10,11 Measurements were performed under low lighting in order to reduce the ambient optical noise. After opening, the samples were placed in a black container on absorbent gauze soaked in normal saline solution. The acquisition strategy was to obtain background corrected fluorescence spectra from each tissue sample. The optical fiber, within a protective guiding catheter, had its free tip placed approximately 1 mm from the point of measurement for as long as was necessary for the recording of the spectrum (3 s). Recordings were made during laser irradiation of the background, as a measure of the fluorescence of the fiber, saline solution, and other external sources of noise. This background spectrum was automatically subtracted from each sample tissue spectrum. Between measurements, the tip was rinsed in saline solution in order to avoid accumulation of tissue on the surface, which might have an effect on the measurements of background that were made in pure saline solution. Four measurements were made for each chamber of every heart. Measurements in the ventricles were made at the apex (left ventricle [LV1]-right ventricle [RV1]), the midpoint of the interventricular septum (LV2-RV2), the midpoint of the anterior Figure 1. Experimental apparatus for measurements of laser-induced fluorescence from myocardial tissue. All optical components, except the pumping laser (argon-ion laser emitting at nm), are located in a diagnostic module. HPF high-pass filter; DM dichroic mirror. 234 Laboratory and Animal Investigations

3 Figure 2. Characteristic examples of fluorescence spectra obtained from the four cardiac cavities. Four peaks and three valleys can be seen in each spectrum. papillary muscle (LV3-RV3), and on the free ventricular wall (LV4-RV4). In the atria, the measurement sites were the free wall (left atrium [LA1]-right atrium [RA1]), the atrial appendage (LA2-RA2 and LA3-RA3), and the interatrial septum (LA4-RA4). For each emission spectrum, we calculated local maximum or imum spectral intensity within selected regions (critical points), which were observed as peaks and valleys in the spectral plots. Statistical Analysis Repeated-measures analysis of variance with one within-four level factor was used to assess whether there were any significant differences among the four sites of each cavity in the emission spectral intensities of the critical points (intracavity comparisons). When no significant intracavity differences were found, the measurements from each critical region for the four different sites were combined into a single sample. We then reemployed a repeated-measures analysis of variance design to test whether the emission spectral intensities in the critical regions within the four cavities were significantly different (intercavity comparisons). In case of significant findings, post hoc tests were used to pinpoint differences. The repeatability of the measurements, for peak one (P1) only, at three different time points (initial, 30, and 24 h) was assessed with simple linear regression; correlation coefficients (r) and slopes (b) close to 1 CHEST / 120 / 1/ JULY,

4 Table 1 95% Confidence Intervals for Intensity at Critical Spectral Points in the RA and RV* Critical Regions RA RV LA P1 246, , , , , , , , , , , , 561 V1 221, , , , , , , , , , , , 524 P2 216, , , , , , , , , , , , 507 V2 158, , , , , , , , , , , , 369 P3 191, , , , , , , , , , , , 429 V3 99, , , , , 96 83, , , , , , , 224 P4 84, , , , , 93 74, , , , , , , 189 *P1 mean intensity in region 547 to 552 nm; V1 mean intensity in region 556 to 561 nm; P2 mean intensity in region 566 to 571 nm; V2 mean intensity in region 580 to 585 nm; P3 mean intensity in region 595 to 600 nm; V3 mean intensity in region 616 to 621 nm; P4 mean intensity in region 666 to 671 nm. are indicators of good repeatability. All p values 5% were the criterion of significance throughout. Results Characteristic examples of fluorescence spectra are shown in Figure 2. All spectra are composed of a fluorescence band with a wide range of peaks and valleys: in the regions of 547 to 552 nm (peak one [P1]), 566 to 571 nm (peak two [P2]), 595 to 600 nm (peak three [P3]), 666 to 671 nm (peak four [P4]), 556 to 561 nm (valley one [V1]), 580 to 585 nm (valley two [V2]), and 616 to 621 nm (valley three [V3]), a high diagnostic potential for tissue characterization is suggested. Table 1 gives 95% confidence intervals of average Figure 3. Mean values and 95% confidence intervals of differences between pairs of cardiac cavities in spectral intensity measured at seven critical spectral points. Differences tend to be greater in the lower frequency bands. P1 difference in spectral intensity at P1 between cavities; V1 difference in spectral intensity at V1 between cavities; P2 difference in spectral intensity at P2 between cavities; V2 difference in spectral intensity at V2 between cavities; P3 difference in spectral intensity at P3 between cavities; V3 difference in spectral intensity at V3 between cavities; P4 difference in spectral intensity at P4 between cavities. 236 Laboratory and Animal Investigations

5 Table 1 Continued LV , , , , , , , , , , , , , , , , , , , , , , , , , , , , 131 spectral intensities in the above-mentioned critical points of the emission spectrum, from all measurement sites. For all these points, the maximum intensity was in the range of P1 at 545 nm and the spectral intensity values showed a decreasing sequence from P1 to P2 to V1 to P3 to V2 to P4 to V3. In each cardiac chamber, there were no significant differences among the four acquisition sites with regard to spectral intensities in any of their critical spectral points (four local peaks and three valleys). Between the chambers, however, there were very significant differences, both in emission spectral intensities and in morphology of spectral plot. More specifically, the average fluorescence emission intensities of all critical points were significantly higher when atria were compared to ventricles (p 0.001) and when left cavities were compared to right (p 0.001). Furthermore, the magnitude of differences was not similar for all critical points. As shown in Figure 3, the magnitude of the differences in spectra, both between atria and ventricles and between left and right chambers, tended to decrease as the spectral wavelength increased. Measurements from all 16 sites (4 sites for each chamber) were used for the comparison among cavities, since there was no intracavity variation. Reproducibility of Measurements There was good reproducibility of the spectral intensity evaluation in all critical points of the emission spectra in all cavities when the first and second measurements were compared (correlation coefficient [r], and slope, [b] very close to 1), but there was no good correlation between the first and third measurements (Table 2). Discussion Optical spectroscopy is a technique that has been used increasingly in recent years for tissue identification. 4 9 It is based on the fact that tissue emits photons when excited by a laser beam. 1 4 More precisely, photon absorption by the tissue causes the excitation of electrons within its molecules and their elevation to a higher energy level. The stimulated electrons remain in this higher-energy state for approximately 10-8 s, after which they decay to the ground state by the emission of photons (fluorescence). The emission spectrum of a tissue is the sum of the emission spectra of the main fluorescent substances: elastin, collagen, and nicotinamide adenine dinucleotide. 1 4,12,13 There are, however, many other substances that may contribute to the creation of the spectrum. In addition, the scattering of the emitted radiation in the tissue is quite high, leading to alteration of the signal. 5,12,14 All the above factors result in the creation of artificial peaks and valleys in the fluorescence spectrum that are characteristic of the particular type of tissue. In the present study, we investigated the characteristics of the emission spectra of myocardial tissue under stimulation by an argon-ion laser, which emits at nm. The wavelength used achieves tissue penetration of the order of hundreds of microns. Even though the radiation within the tissue falls Table 2 Correlation Coefficient (r) and Slope (b): First vs Second and First vs Third Measurements in All Critical Regions of the LA and LV* LA LV RA RV Critical Regions b r b r b r b r b r b r b r b r P P P P V V V CHEST / 120 / 1/ JULY,

6 exponentially (I I o e -kx, where x depth), a substantial number of photons pass through the endothelium and reach the interior sections of the tissue. Our results show that there are significant differences in the emission spectra recorded from different anatomic sites within the heart. The atrial spectra not only have an overall higher intensity, but also a different morphology from those recorded in the ventricles. This morphologic difference is due to the fact that the difference between the mean spectral intensities in the atria and the ventricles is much greater in the short than in the long wavelength bands of the emission spectrum. It is worth noting, however, that within each chamber the emission spectrum is more or less constant and appears to be independent of the site from which it is recorded. These observations are made herein for the first time (to our knowledge), and are clearly crucial if the method is to be used for clinical purposes. Also of crucial importance is the good reproducibility of our results after 30. The worsening reproducibility over a longer period does not invalidate the method. Tissue and biochemical changes occurring over that time might be expected to lead to changes in the emission spectra. At this time, we do not know the reason for the difference in fluorescence between atrial and ventricular tissue. Myocardial fibers are known to be closer to the thin layer of ventricular endocardium, while in the atrial myocardium they lie under a thicker layer of endothelium. 11 Taking into account the low degree of penetration of the laser beam (a few hundred microns), this difference could explain the higher spectral intensity of the atrial fluorescence, but not its different morphology. Further studies are needed to detere the biochemical and histologic structures that are responsible for the peaks and valleys in the emission spectrum, in both the atrial and ventricular myocardium. Another issue that must be addressed is whether the laser beam causes damage to the myocardial tissue. Although it is true that the light energy used is very small (10 to 12 milliwatts) and the time of illuation (approximately 3 s) is short enough to make any tissue damage unlikely, this needs further, more detailed investigation, including in vivo studies. Comparison With Previous Studies A small number of earlier studies 5,8,11,14 also used laser spectroscopy for the identification of myocardial tissue. Nilsson et al 14 recorded the emission spectra from the LV of pig hearts under laser stimulation and compared it with the spectrum of other tissues. Perk et al 5 showed that the characteristics of the emission spectra from fibrotic and/or ischemic myocardial tissue under laser stimulation at 308 nm were different from those of healthy myocardium. The same investigators in another study 11 showed that the emission spectra of the human nodal conduction system were different from that of surrounding tissue. In contrast, Aziz and coworkers, 8 using a laser of similar wavelength in dog hearts, found no difference between the nodal conduction system and surrounding myocardial tissue. In a more recent study, Morgan et al 15 recorded the emission spectra from the LVs of rats, excited by blue light, and showed that the spectrum changed consistently as a heart underwent transplant rejection and could even distinguish the degree of tissue rejection. The above studies were based on the exaation of small histologic preparations and/or were aimed at answering specific questions. In contrast, our study is the only one (to our knowledge) to exae whole and healthy hearts. Thus, we were able to get an overall evaluation of the emission spectra of myocardial tissue, not only exaing the basic characteristics (intensity, morphology, reproducibility of measurements), but also relating them to specific anatomic sites. Limitations of the Study In this study we used sheep hearts. We came to this decision for two reasons: first, the ready availability of intact, young sheep hearts that could be presumed to be healthy; second, because our study protocol includes a second, in vivo phase in which sheep will be used. It remains to be investigated to what extent the constancy of measurements within each cardiac chamber and the reproducibility of the method apply to human hearts. The most important limitation of the method arises from the fact that the depth of penetration of the laser beam is only a few hundred microns. 1 4 As a result, the information obtained by spectroscopy reflects the composition of the endocardium and only a small thickness of myocardial tissue. At this point, it is worth mentioning that the penetration depth could have been greater had we used a laser that emits at a higher wavelength. However, the increase in depth would be at the expense of a loss in diagnostic capability, since there are not many chromophores in the tissue that can absorb the higher wavelength and produce fluorescence. Furthermore, our selection was based on the fact that it enabled us to use a relatively small device that can be employed within the laboratory along with standard techniques. The remaining limitations are due to its being an in vitro study. 238 Laboratory and Animal Investigations

7 Conclusions Future Applications Spectroscopy as a means of identifying tissue has only been used until now for experimental purposes. However, at least with regard to the myocardial tissue that we exaed, it appears to have the necessary properties to be used for clinical applications as well. According to our findings and those of other studies, spectroscopy might prove useful in electrophysiology for the mapping of myocardium and the detection of fibrotic arrhythmogenic foci. The various electrophysiologic tests that are used for this purpose have a sensitivity of 50 to 90%. Spectroscopy, even as a supplementary exaation, could offer significant assistance. Spectroscopy might also prove useful in the detection of pathologic myocardial states (cardiomyopathies, myocarditis, right ventricular dysplasia, intramyocardial fibrosis, etc) or tissue rejection in the case of a transplanted heart, enhancing and in some cases replacing biopsy. It should be noted that spectroscopy appears to have two basic advantages over biopsy: first, the capacity for multiple measurements from different sites at one session, and second, the instant availability of results. In vivo studies will be needed in order to establish whether the above-mentioned clinical applications are indeed feasible. ACKNOWLEDGMENT: The authors acknowledge the contribution of Gregory Chlouverakis, PhD, to the statistical analysis. References 1 Andersson PS, Montan S, Persson T, et al. Fluorescence endoscopy instrumentation for improved tissue characterization. Med Phys 1987; 14: Tsien RY. Fluorescent probes of cell signaling. Ann Rev Neurosci 1989; 12: Papazoglou TG, Arakawa K, Grundfest WS, et al. Laserinduced tissue autofluorescence versus exogenous chemical probe induced fluorescence as an arterial layer detection method: a comparative study. Optic Fiber Med V Spie 1990; 1201: Laifer LI, O Brien KM, Stetz ML, et al. Biochemical basis for the difference between normal and atherosclerotic arterial fluorescence. Circulation 1989; 80: Perk M, Flynn GJ, Smith C, et al. Laser-induced fluorescence emission: I. The spectroscopic identification of fibrotic endocardium and myocardium. Lasers Surg Med 1991; 11: Lam S, Palcic B, McLean D, et al. Detection of early lung cancer using low-dose Photofrin II. Chest 1990; 97: Laufer G, Wollenek G, Hohla K, et al. Excimer laser-induced simultaneous ablation and spectral identification of normal and atherosclerotic arterial tissue layers. Circulation 1988; 78: Aziz D, Caruso A, Aguirre M, et al. Fluorescence response of selected tissues in the canine heart: an attempt to find the conduction system [abstract]. Proc Int Soc Optic Eng 1992; 1642: Schomaker KT, Frisoli JK, Richter JM, et al. Discriation of adenomatous from hyperplastic colonic polyps in vitro by laser-induced fluorescence [abstract]. Lasers Surg Med 1990; 2(suppl):5 10 Leon MB, Lu DY, Prevosti LG, et al. Human arterial surface fluorescence: atherosclerotic plaque identification and effects of laser atheroma ablation. J Am Coll Cardiol 1988; 12: Perk M, Flynn GJ, Gulamhusein S, et al. Laser induced fluorescence identification of sinoatrial and atrioventricular nodal conduction tissue. Pacing Clin Electrophysiol 1993; 16: Papazoglou TG. Malignancies and atherosclerotic plaque diagnosis: is laser induced fluorescence spectroscopy the ultimate solution? J Photochem Photobiol B 1995; 28: Horvath KA, Torchiana DF, Daggett WM, et al. Monitoring myocardial reperfusion injury with NADH fluorometry. Lasers Surg Med 1992; 12: Nilsson AM, Heinrich D, Olajos J, et al. Near infrared diffuse reflection and laser-induced fluorescence spectroscopy for myocardial tissue characterization. Spectrochim Acta A Mol Biomol Spectrosc 1997; 53A: Morgan DC, Wilson JE, MacAulay CE, et al. New method for detection of heart allograft rejection: validation of sensitivity and reliability in a rat heterotopic allograft model. Circulation 1999; 100: CHEST / 120 / 1/ JULY,