Slow Intramural Heating With Diffused Laser Light
A Unique Method for Deep Myocardial Coagulation
Background—Catheter ablation of postinfarction ventricular tachycardia (VT) may be limited by insufficient myocardial coagulation or excessive endocardial or epicardial damage. We propose that volumetric heating restricted to intramural sites may improve the outcome and safety of this procedure, especially if delivered at rates that enhance heat conduction and forestall adverse tissue changes.
Methods and Results—A novel optical fiber with a diffusing tip for direct intramural, volumetric laser heating was tested via thoracotomy and percutaneously in normal dogs. Low-power (2.0- to 4.5-W) diode laser light (805 nm) diffused within tissue induced large lesions but no visible surface damage, mural thrombi, or transmural perforation. Mean lesion depth approximated tip length (10 mm). Mean lesion widths in the thoracotomy and percutaneous groups were 5.8±0.5 to 9.1±0.84 mm and 5.2±0.85 to 7.9±1.1 mm, respectively, depending on the light dose. Mean volumes in the percutaneous group were 1006±245 to 2471±934 mm. ST-segment depression, appearing in unfiltered bipolar electrograms recorded from the guiding catheter, was specific for lesion induction. All dogs survived the protocol, which included a 1-hour observation period. In cross section, lesions were elliptical to spherical and characterized by extensive contraction-band necrosis abruptly bordering viable tissue. No platelets or fibrin adhered to the endocardium.
Conclusions—Slow, volumetric, and direct intramyocardial heating induces large, deep lesions without hazardous tissue damage. Such heating might cure postinfarction VT more successfully and safely than present techniques. Further testing and development of this method seem warranted.
Postinfarction ventricular tachycardia (VT) caused by midmyocardial or subepicardial reentrant circuits can be difficult to treat with catheter ablation, which can supplement but has not supplanted the implantable cardioverter-defibrillator.1 2 3 4 5 6 A deep VT substrate may escape coagulation if intervening zones of high temperatures caused by resistive surface heating char or vaporize tissue.7 8 These events both limit heat conduction and increase the risk of endocardial damage, thrombus formation,7 9 10 11 12 13 14 and pericardial tamponade.15
Deep, large-volume lesions could be safely generated by less intense but strictly intramural heating that allows maximal heat conduction but avoids the endocardium and epicardium.16 For this, photons are an ideal source because when scattered, they provide a slowed rate of volumetric heating. Endocardial and epicardial laser irradiation has been used for postinfarction VT,17 18 19 20 but heat-induced surface changes can limit its effectiveness,21 as can blood interposed between the optical fiber and the endocardium. Perhaps the major limitation to laser technology has been a concern over its size, complexity, and expense.22 Each has been greatly reduced by the advent of the diode generator.
A penetrating optical fiber that itself diffuses photons would generate the most effective intramural lesions.23 24 25 Diffusing-tipped fibers “distribute the light over a large area without making use of the scattering properties of the biological tissue itself.”24 They directly heat greater volumes than do standard, nonmodified tips; reduce power density; slow the rate of heating; and reduce the likelihood of char, vaporization, and their attendant complications.21 25 Using such a fiber with a diode generator, we hypothesized that slow intramyocardial heating with diffused laser light could induce deep, large coagulative lesions without significant endocardial or epicardial damage. The studies presented here demonstrate the dosimetry and pathology of lesions induced with this new technique, which may safely improve cure rates, extend catheter ablation to more patients with postinfarct VT, and reduce the need for device implantation or medical therapy.
Optical Fiber Design
Silica fibers 400 and 200 μm in diameter, clad with a hard polymer (total diameter, 600 and 400 μm), were used for the open-chest and percutaneous experiments, respectively (Figure 1⇓). The distal end of each fiber had been modified by attachment of a 5-mm “diffusing element” and a 4-mm sharpened (bare fiber) tip. Titanium dioxide particles uniformly embedded in the diffusing element scatter photons when struck by laser light (Figure 2⇓), increasing the volume and slowing the rate of heating. A dielectric mirror between the diffusing element and sharpened tip reflects photons internally and prevents the direct heating of tissue ahead of the fiber. In pilot studies, large lesions least likely to damage either myocardial surface were induced once the distal 1 cm (including the sharpened tip and diffusing element) had penetrated the tissue at 90°.
A diode laser (Diomed Ltd) operating at 805 nm was used in all experiments. Light transmission through the optical fibers was measured at baseline and before each exposure with an integrating sphere to ensure stable power delivery. Fibers were discarded if output was diminished by charring or fracture of the fiber tip.
Approval was obtained from the Institutional Review Board and Animal Care and Use Committee of the University of Texas Medical Branch at Galveston.
Normal mongrel dogs weighing 15 to 25 kg were intubated after receiving butorphenol 0.1 mg/kg, acepromazine maleate 0.05 mg/kg, and atropine 0.02 mg/kg. Anesthesia was induced with intravenous thiopental sodium 20 mg/kg and maintained with 1% to 2% halothane. Skin electrodes were placed for rhythm monitoring. A left thoracotomy exposed the heart, which was stabilized in a pericardial cradle. The optical fiber was housed in an end-hole catheter, through which saline was infused at room temperature by a Harvard pump (10 mL/min) for epicardial cooling. Before irradiation, the distal fiber was placed 1 cm into the tissue, 90° to the epicardium. Animals were monitored for 1 hour after the last lesion. The hearts were then arrested with intravenous potassium chloride before removal.
A radiopaque, end-hole, deflectable, quadripolar catheter allowed percutaneous guidance of the optical fiber. The distal electrode was 5 mm long; others were 3 mm, with an interelectrode distance of 7 mm. Experiments were performed in a catheterization laboratory on 11 animals anesthetized and monitored as above. In 6, femoral artery pressure was monitored. The fiber was retracted within the guiding catheter as it was introduced via the left carotid artery and placed in firm contact with the endocardium. The cardiac silhouette and a knowledge of canine anatomy in the various radiographic views helped in positioning the distal catheter at angles presumed to be perpendicular to the endocardium. The distal 1 cm of the fiber was then deployed. Cardiac motion did not interfere with this action, and once intramural, the fiber was secure. Unfiltered (1- to 5000-Hz) bipolar electrograms were recorded from the distal and proximal electrode pairs before, during, and after each laser application. After this, the fiber was again retracted into the guiding catheter, and both the catheter and fiber were removed.
Pilot lesions were made to determine “acceptable” doses, ie, doses ranging in 30-second and 0.5-W increments from the lowest that first caused visible coagulation to the highest that did not extensively char or vaporize the lesion. This range was 3 to 4.5 W and 30 to 120 seconds in the open-chest model. In the percutaneous experiments, 2, 2.5, and 3 W were delivered over 60 seconds, and 2.5 W was delivered over 90 seconds. Because 2.5 W over 90 seconds and 3.0 W over 60 seconds induced lesions of similar size, the former was not used for dosimetry analysis.
Lesion Measurement and Pathology
Lesions were bisected in a plane perpendicular to the epicardial or, in the percutaneous experiments, endocardial surface. The maximum lesion width (short axis) and length (long axis) were measured with a micrometer before tissue fixation in 10% neutral-buffered formalin. A “maximum depth” was determined for percutaneously induced lesions, because distal margin depth often exceeded lesion length. Lesion volume was calculated as (4/3)(π)(long axis)(short axis)2, the volume of an ellipse rotated about its long axis (or prolate spheroid).26 Fixed lesions were sectioned parallel to the original bisection plane and processed for standard histopathological analysis.
In open-chest experiments, a 2-way ANOVA determined the effects of exposure time (30, 60, 90, and 120 seconds) and power (3.0, 3.5, 4.0, and 4.5 W), and their interaction, on the continuous outcome variables of lesion depth, width, and volume. Data are analyzed with the caveat that the highest and lowest doses were not used: 3.0 and 3.5 W×30 seconds produced no visible lesions, and 4.5 W×120 seconds charred tissue. Differences among levels of time and power and of their interaction were evaluated with Bonferroni-adjusted t tests, using an experiment-wise error rate of 0.05. For example, for width, which demonstrated a time×power interaction, the α level of significance was 0.0006 for pairwise t tests (or 0.05/78, where 78 was the number of possible comparisons). The GLM procedure in the statistical software SAS with option LSMEANS/PDIFF was used for all analyses.27 Means are given with their ±SD. All t tests mentioned in the Results section are Bonferroni-adjusted.
In percutaneous experiments, lesion dimensions at power levels of 2.0, 2.5, and 3.0 W delivered over 60 seconds were evaluated with 1-way ANOVA followed by Bonferroni-adjusted t tests. The sensitivity and specificity that electrogram changes had for lesion induction were determined at each fluoroscopic position, using gross coagulation at the matching site found postmortem as the “gold-standard,” comparative test.28 Only well-defined ST-segment depression and T-wave peaking, as illustrated in Figure 4⇓, were analyzed.
Nine animals underwent thoracotomy and 88 laser applications (5 to 15 lesions per heart; mean, 10±1 per heart). Of these, 80 could be measured accurately. Fiber insertion caused occasional premature ventricular contractions and brief runs of nonsustained VT. Ventricular fibrillation (VF) occurred in 2 animals during the 10th and in another during the 4th lesion. After defibrillation, this third animal underwent 8 additional lesions without incident. All 9 dogs survived the protocol, including the 1-hour observation period.
Combining all 13 doses (Table 1⇓), mean lesion width was 7.24±1.5 mm (range, 4.8 to 10.8 mm). Mean depth (10.4±1.6 mm; range, 6.3 to 14.7 mm) approximated tip length. Mean lesion volume was 2460.9±1307 mm3 (range, 837.8 to 6316 mm3).
Examining the effects of power and time on width showed a significant power×time interaction (P=0.02). Thus, to understand how width relates to exposure time, one must consider the power level, and vice versa. This interaction permits lesion widths to be compared as in Table 1⇑, yielding a general pattern of increasing width with increasing time and power. The 30-second level was an exception, producing narrow lesions across all powers. The entire range of widths could be induced by varying either time (30 to 120 seconds) against 4 W or power (3.0 to 4.5 W) against 60 seconds. Grouped as such, widths increased linearly with time but nonlinearly with increasing power (Figure 3⇓).
Most photons diffuse laterally23 ; however, with longer exposures, lesion depth increased significantly (P<0.01). No power×time interaction (P=0.48) affected depth. The t tests showed that levels of 90 seconds (10.77±1.23 mm) and 120 seconds (10.73±1.8 mm) differed from 30 seconds (8.96±1.24 mm), but there were no significant pairwise differences between power levels.
Likewise, there was no power×time interaction on volume (P=0.10), but significant effects of power and time were present (both P<0.0001). For power, t tests showed significant pairwise differences comparing levels of 3.5 W (2548±1162 mm3), 4.0 W (2366±1104 mm3), and 4.5 W (3193±1778 mm3) with 3.0 W (1606±445 mm3), but the higher levels did not differ from each other. For time, levels of 60 seconds (2361±1419 mm3), 90 seconds (2722±1209 mm3), and 120 seconds (2805±1347 mm3) differed from 30 seconds (1256±479 mm3) but not from each other.
Eleven dogs underwent 48 laser applications (2 to 7 per heart). Twenty-eight lesions were found (≤5 per heart). Higher and lower doses were identified equally well, suggesting that lesions were missing because of unsuccessful fiber tip penetration. Successful deployments caused ventricular ectopy but no ST-T changes in the endocardial electrogram. Brief, well-tolerated, nonsustained VT occurred rarely during heating. A brief idioventricular rhythm was seen in 1 animal 5 minutes after catheter removal. Sustained VT or VF did not occur, and arterial pressure remained stable in the 6 animals monitored.
With the 400-μm fiber, a smaller and lower-power range induced “acceptable” lesions similar to the open-chest experiments. As before, width increased nonlinearly with power (Table 2⇓), ranging from 4 to 10 mm (6.3±1.4 mm); the mean of maximum depths (10.2±1.8 mm; range, 7.3 to 13.3 mm) approximated tip length and generally exceeded lesion length (8.3±1.2 mm; range, 6 to 11.4 mm). Occasional lesions were centered within the LV wall (Figure 6⇓).
One-way ANOVA followed by Bonferroni-adjusted t tests detected significant pairwise differences when comparing widths for 2.5 and 3.0 W with 2.0 W and when comparing 2.5 with 3.0 W. Increasing power caused a trend but no significant change in lesion length or maximum depth. Volumes at 3 W differed significantly from those at 2.5 and 2.0 W. The greatest widths and volumes were generated by 3 W over 60 seconds.
ST and T waves did not change with endocardial penetration. Epicardial penetration did not occur; thus, possible ST-T changes accompanying this event are unknown. However, heat-induced ST depression with T-wave peaking (by 0.5 to 2.0 and 0.5 to 3.0 mV, respectively) were specific for gross intramyocardial lesion induction (Figure 4⇓ and Table 3⇓), but T-wave peaking alone was not. Marked T-wave peaking and ST depression were unusual when proximal lesion margins were deep. RR and QTc intervals did not change nor did R-waves regularly decrease with coagulation.
The epicardium and endocardium were remarkably free of gross tissue damage in both open-chest and percutaneous experiments; in the latter, lesion sites were not apparent on inspection of the endocardium, and transmural perforation, pericardial tamponade, and epicardial coronary artery damage did not occur.
In cross section, lesions ranged from elliptical to spherical and were well outlined grossly by a line of congested blood vessels (Figures 5⇓ and 6⇓). Microscopically, coagulated tissue abruptly bordered viable tissue (Figure 7A⇓). Necrotic tissue was characterized by extensive, homogeneous contraction bands. No evidence of vascular thrombi or hemorrhage was seen (Figure 7B⇓). Focal subendocardial cell necrosis was noted within a 1- to 2-mm radius from the insertion point, but this was not associated with platelet or fibrin adherence to the endocardial surface (Figure 7C⇓).
This unique optical fiber diffuses light for slow, intramural heating and large, deep, well-circumscribed lesions that do not disrupt either endocardium or epicardium. It therefore appears well suited for the treatment of postinfarction VT.
Thermal conduction through tissue is directly proportional to the temperature gradient across the medium and to the time allowed for conduction to occur29 ; however, a steep gradient that chars, boils, or increases tissue reflectance will reduce heat transfer7 21 and, if it involves the endocardium, may promote mural thrombi.10 11 12 13 14 Surface cooling increases lesion size but transfers high temperatures into the midmyocardium,15 30 where tissue vaporization and explosion may cause pericardial tamponade.15 By heating tissue more slowly, this fiber both reduces risk and generates deep, large-volume lesions.
Once deployed, the entire diffusing element is intramyocardial, and the opposing tissue surface is not directly irradiated. The endocardium is not disrupted but remains free of adherent platelets and fibrin. Well-defined margins reduce the arrhythmogenic potential of the lesions.31 32 Most photons diffuse laterally,23 and within certain dose ranges, lesion width changes linearly or nonlinearly with increases in time and power, respectively. Mean depths approximated the tip length inserted (here 10 mm); thus, a tip shorter than the target tissue is thick should not damage the epicardium. However, the wide range of depths suggests that tip length is not the sole determinant of this parameter. In the open-chest experiments, longer exposure times (90 and 120 seconds) increased lesion depth significantly compared with 30 seconds. In the percutaneous experiments, the highest powers caused a trend toward deeper lesions. Thus, coagulation beyond the fiber tip will be less likely when lower doses are used. Of course, extended tip deployment will increase depth (Figure 6⇑). This would target very deep tachycardia circuits but spare those near the endocardium and increase the chance of epicardial damage. Clearly, refined control over penetration depth is a desirable future objective.
Our data are consistent with established models of tissue heating.16 Lesion size is determined by the extent of light diffusion and heat transport. Steady-state temperature occurs after photons distribute, when heat deposition and dissipation rates are equal. Heating must coagulate but not overwhelm rates of dissipation and thereby cause charring or boiling. Here, 4 W was the highest power that could be delivered over the longest exposure time (120 seconds), and this dose generated the widest lesions (Table 1⇑). Higher powers (≥4.5 W) delivered over 120 seconds seriously carbonized the tissue. Widths gradually approached a maximum with increasing exposure time (Figure 3⇑, top). As power increased over a fixed time, lesions widened more abruptly (Figure 3⇑, bottom), implying rapid heat deposition and an increased risk of carbonization. Finally, lesion length increased significantly only with the longest exposure times, because this is not the primary direction of heat conduction.
Coagulation could be identified in only 58% of all lesions attempted percutaneously. We suspect that the optical fiber had not penetrated the tissue, a problem that should be solved with device improvement. Alternatively, adequate temperatures may not have developed after penetration, because heat delivery did not overcome intramural heat loss.16 In this event, T-wave peaking in the endocardial electrogram may have indicated nonlethal injury, with secondary tissue hyperkalemia.33 34
The most specific repolarization change indicating visible midmyocardial coagulation was ST-segment depression (Figure 4⇑, Table 3⇑). However, electrogram changes were relatively insensitive to the development of intramural coagulation and are not predictive of optical fiber penetration. Changes that might accompany epicardial penetration are not known, because this did not occur. Nevertheless, electrogram changes may have implications that are not explored here and may become more helpful in predicting or understanding subsurface events once detection becomes more sensitive. Intramyocardial heating caused VF, but its infrequency prevents correlation with lesion number, size, or delivered energy, and its long-term implications are unknown. All episodes were in the open-chest experiments (occurring with lesion 10 in 2 animals and during lesion 4 in a third). Other, nonsustained episodes of ventricular ectopy did not require intervention and had no apparent adverse sequelae.
In summary, this diffusing-tip optical fiber is unique in its ability to induce large intramural lesions without directly heating the endocardium. Volumetric heating that is modified to forestall charring and vaporization reduces the likelihood of transmural perforation and allows thermal energy to conduct to its theoretical maximum. Tip penetration sustains tissue contact; verification of adequate deep tissue coagulation may be confirmed by ST-segment depression recorded from the guiding catheter. Because only low powers are necessary to induce these lesions, diode laser generators capable of VT ablation should remain small and thus feasible for use in electrophysiology laboratories. The technique is limited by the requirement that the guiding catheter remain nearly perpendicular to the endocardium and, at the present time, by difficulty with successful tissue penetration. However, its potential for effective VT therapy is incentive for improvements that could make it a very useful addition to our clinical armamentarium.
This project was supported by grants from the American Heart Association, Texas Affiliate (94G-658), the University of Texas Medical Branch Keating Endowment for Cardiovascular Research (2028-93), and the National Institutes of Health (1-R43-HL-54397-01). The optical fibers used in this study were provided by Rare Earth Medical, Inc, West Yarmouth, Mass.
- Received June 26, 1998.
- Revision received November 23, 1998.
- Accepted November 23, 1998.
- Copyright © 1999 by American Heart Association
Scheinman MM. Patterns of catheter ablation practice in the United States: results of the 1992 NASPE survey. Pacing Clin Electrophysiol. 1992;17:873–875.
Morady F, Harvey M, Kalbfleisch SJ, el-Atassi R, Calkins H, Langberg JJ. Radiofrequency catheter ablation of ventricular tachycardia in patients with ischemic heart disease. Circulation. 1993;87:363–372.
Kim YH, Sosa-Suarez G, Trouton TG, O’Nunain SS, Osswald S, McGovern BA, Ruskin JN, Garan H. Treatment of ventricular tachycardia by transcatheter radiofrequency ablation in patients with ischemic heart disease. Circulation. 1994;89:1094–1102.
Blanchard SM, Walcott GP, Wharton M, Ideker RE. Why is catheter ablation less successful than surgery for treating ventricular tachycardia that results from coronary artery disease? Pacing Clin Electrophysiol. 1994;17(pt I):2315–2335.
Kaltenbrunner W, Cardinal R, Dubuc M, Shenasa M, Nadeau R, Tremblay G, Vermeulen M, Savard P. Epicardial and endocardial mapping of ventricular tachycardia in patients with myocardial infarction: is the origin of the tachycardia always subendocardially localized? Circulation. 1991;84:1058–1071.
Kramer JB, Saffitz JE, Witkowski FX, Corr PB. Intramural reentry as a mechanism of ventricular tachycardia during evolving canine myocardial infarction. Circ Res. 1985;56:736–754.
Haines DE, Verow AF. Observations on electrode-tissue interface temperature and effect on electrical impedance during radiofrequency ablation of ventricular myocardium. Circulation. 1990;82:1034–1038.
Haines DE. The biophysics of radiofrequency catheter ablation in the heart: the importance of temperature monitoring. Pacing Clin Electrophysiol. 1993;16(pt II):586–591.
Thakur RK, Klein GJ, Yee R, Zardini M, Li HG, Morillo CA. Embolic complications following radiofrequency catheter ablation in the left heart. Circulation. 1993;88(suppl I):I-296. Abstract.
Thakur RK, Klein GJ, Yee R. Embolic complications in Multicenter European Radiofrequency Survey (MERFS). Eur Heart J. 1994;15:1290–1291. Letter.
Nakagawa H, Imai S, Arruda M, Schleinkofer D, Freihoff F, Shah N, Beckman K, McClelland J, Gonzalez M, Otomo K, Lazzara R, Jackman W. Prevention of pericardial tamponade during radiofrequency ablation of accessory pathways using a saline irrigated electrode. Pacing Clin Electrophysiol. 1997;20:1068. Abstract.
Pearce J, Thomsen S. Rate process analysis of thermal damage. In: Welch AJ, van Gemert MJC, eds. Optical-Thermal Responses of Laser-Irradiated Tissue. New York, NY: Plenum Press; 1995:561–606.
Littman L, Svenson RH, Gallagher JJ, Selle JG, Zimmern SH, Fedor JM, Colavita PG. Functional role of the epicardium in postinfarction ventricular tachycardia: observations derived from computerized epicardial activation mapping, entrainment, and epicardial laser photoablation. Circulation. 1991;83:1577–1591.
Svenson RH, Gallagher JJ, Selle JG, Zimmern SH, Fedor JM, Robicsek F. Neodymium:YAG laser photocoagulation: a successful new map-guided technique for the intraoperative ablation of ventricular tachycardia. Circulation. 1987;76:1319–1328.
Nolsoe C, Torp-Pedersen S, Olldag E, Holm HH. Bare fiber low power Nd:YAG laser interstitial hyperthermia: comparison between diffuser tip and non-modified tip: an in vitro study. Lasers Med Sci. 1992;7:1–7.
Verdaasdonk RM, Borst C. Optics of fibers and fiber probes. In: Welch AJ, Gemert MJC, eds. Optical-Thermal Response of Laser-Irradiated Tissue. New York, NY: Plenum Press; 1995:619–666.
Selby SM. Standard Mathematical Tables. Cleveland, Ohio: The Chemical Rubber Co; 1971:18.
SAS/STAT© User’s Guide.Volumes 1 and 2, Version 6, ed 4. Cary, NC: SAS Institute Inc; 1992.
Sox HC, AB Marshal, MC Higgins, KI Morton. Medical Decision Making. Stoneham, MA: Butterworth-Heinemann; 1988:107–111.
Shitzer A, Eberhart RC. Review of elementary heat transfer. In: Shitzer A, Eberhart RC, eds. Heat Transfer in Medicine and Biology: Analysis and Applications. New York, NY: Plenum Press; 1985:411–416.
Svaasand LO, Boerslid T, Oeveraasen M. Thermal and optical properties of living tissue: application to laser-induced hyperthermia. Lasers Surg Med. 1995;5:589–602.
Levine JH, Spear JF, Weisman HF, Kadish AH, Prood C, Siu CO, Moore EN. The cellular electrophysiologic changes induced by high-energy electrical ablation in canine myocardium. Circulation. 1986;73:818–829.
Lee BI, Gottdiener JS, Fletcher RD, Rodriguez ER, Ferrans VJ. Transcatheter ablation: comparison between laser photoablation and electrode shock ablation in the dog. Circulation. 1985;71:579–586.
Janse MJ, Kleber AG. Electrophysiological changes and ventricular arrhythmias in the early phase of regional myocardial ischemia. Circ Res. 1981;49:1069–1081.
Nath S, Lynch C, Whayne JG, Haines DE. Cellular electrophysiological effects of hyperthermia on isolated guinea pig papillary muscle: implications for catheter ablation. Circulation. 1993;88(pt I):1826–1831.