Author + information
- Received June 1, 2017
- Revision received August 10, 2017
- Accepted August 31, 2017
- Published online October 16, 2017.
- Vittorio Calzolari, MDa,
- Luca De Mattia, MDa,∗ (, )
- Stefano Indiani, Engb,
- Martino Crosato, MDa,
- Alberto Furlanetto, MDc,
- Claudia Licciardello, MEngb,
- Paolo Antonio Maria Squasi, MDa and
- Zoran Olivari, MDa
- aDepartment of Cardiology, Ca’ Foncello Hospital, Treviso, Italy
- bSt. Jude Medical Italia, Agrate Brianza, Italy
- cDepartment of Anatomical Pathology, Ca’ Foncello Hospital, Treviso, Italy
- ↵∗Address for correspondence:
Dr. Luca De Mattia, Arrhythmia and EP Unit, Cardiology Department, Cà Foncello Hospital, Piazzale Ospedale n°1, 31100 Treviso, Italy.
Objectives In an in vitro model, the authors tested the hypotheses that: 1) lesion dimensions correlate with lesion size index (LSI); and 2) LSI could predict lesion dimensions better than power, contact force (CF), and force–time integral (FTI).
Background When performing radiofrequency (RF) catheter ablation for cardiac arrhythmias, reliable predictors of lesion quality are lacking. The LSI is a multiparametric index incorporating time, power, CF, and impedance recorded during ablation.
Methods RF lesions were created on porcine myocardial slabs by using an open-tip irrigated catheter capable of real-time monitoring of catheter–tissue CF. Initially, 3 power settings of 20, 25, and 30 W were used with a fixed CF of 10 g. A fixed power of 20 W was then set with a CF of 20 and 30 g, thereby yielding a total of 5 ablation groups. In each group, LSI values of 5, 6, 7, and 8 were targeted. Sixty RF lesions were created by using 20 ablation protocols (3 lesions for each protocol).
Results Lesion width and depth were not correlated with power or CF, but the results significantly correlated with FTI (p < 0.01) and LSI (p < 0.0001). Four steam pops occurred with power set at 30 W; no pops were noted with 20 or 25 W even when high LSI values were targeted.
Conclusions In this in vitro model, FTI and LSI predicted RF lesion dimensions, whereas power and CF did not. The LSI predictive value was higher than that of FTI. Steam pops occurred only using high ablation power levels, regardless of the targeted LSI.
- ablation complications
- ablation lesion dimensions
- contact force catheter
- lesion size index
- radiofrequency ablation
Radiofrequency (RF) catheter ablation is the therapy of choice for several cardiac arrhythmias. During ablation, proper RF delivery is crucial both to obtain an effective lesion and to avoid excessive heating that can possibly lead to thrombus formation, steam pop, and/or perforation. Parameters such as RF power, catheter tip temperature, and impedance drop are commonly monitored during RF delivery. However, these parameters display limited accuracy in assessing lesion quality, and more reliable indexes are needed. Conversely, the biophysics of RF lesion formation is known to be the result of a complex interplay of many factors, namely RF delivery time, catheter contact force (CF), power delivered, and tissue impedance.
Recently, a novel, externally irrigated–tip catheter has been developed, allowing real-time monitoring of in vivo catheter tissue CF by means of optical fibers (1). This catheter offers the opportunity to obtain novel ablation indexes, such as force–time integral (FTI) (2). FTI was shown to correlate with lesion size in an in vitro model (3) and to predict lesion size and steam pop occurrence in the beating canine heart (4). The lesion size index (LSI) is another multiparametric index that incorporates time, power, CF, and impedance data recorded during RF ablation in a weighted formula; it could therefore more precisely describe the complexity of in vivo ablation biophysics and help to predict the extent of myocardial tissue lesions (5,6).
In the present study, an in vitro model tested the hypotheses that: 1) lesion dimensions correlate with LSI; and 2) LSI could be a better predictor of lesion dimensions than power, CF, and FTI. As a secondary endpoint, we investigated if 1 of these variables could help to predict the occurrence of steam pops.
A 7-F, 3.5-mm open-tip irrigated TactiCath Quartz (St. Jude Medical, St. Paul, Minnesota) ablation catheter was used for the experiment. As previously described (3), this catheter has a force sensor incorporated into its distal part, which consists of a deformable body and 3 optical fibers that use infrared laser light to measure microdeformations caused by forces applied to the tip of the catheter.
FTI is defined as the total area under the CF curve (contact FTI) applied during RF application, as previously described (3). FTI was found to predict the clinical outcomes in patients with paroxysmal atrial fibrillation after pulmonary vein isolation (5). FTI is derived from a simple multiplication of CF by time and does not take into account the important role of power delivery.
Lesion size index
LSI is a novel ablation index that incorporates CF between the ablation tip and target tissue, impedance, power applied, and duration of RF delivery (6). An empirical model of the effect of these parameters was developed from a series of in vivo experiments encompassing >3,000 lesions in both animals and humans, considering the nonlinear behavior of lesion formation due to the heterogeneity of concurring factors. These factors include: 1) the shift from resistive (joule) heating during the early phase of lesion formation to diffusive (conductive) heating during a later phase; 2) the delay between the variation of force and/or current and the change in lesion growth rate due to thermal latency; and 3) the fact that lesions rapidly grow to a certain depth (typically about 3 mm), beyond which the lesion depth increases at a slower rate.
The equation describing the LSI model is as follows:where f0, f1, and f2 are force parameter coefficients, i1 and i2 are electrical current coefficients, k0 is a diffusive heating coefficient, and τ is a characteristic time value. The coefficients are the results of best curve fitting of experimental data acquired during preclinical studies. The equation comprises a resistive heating component (1-k0) that is independent of time; resistive heating occurs in the first 3 to 5 s of RF delivery and accounts for a maximum lesion depth up to 4 mm. The diffusive heating component, k0 [(1-e–t/τ)/(1−e–60/τ)] in the equation, is time dependent, slower (30 to 50 s), reaches a maximum depth of 6 to 9 mm, and reflects heat conduction from the initial resistive lesion (6). However, LSI has never been validated in a prospective study or in in vitro or in vivo studies.
The TactiCath Quartz catheter was mounted in a fixed standard 8-F, 11-cm sheath and manually maneuvered over a ground platform placed within a tank filled with circulating physiological saline solution at room temperature (Figure 1). The circulating bath used a pump to produce nonpulsatile flow directed perpendicularly to the myocardium surface at a rate of 5 l/min.
Fresh porcine heart muscle slabs (thickness of 2 to 4 cm, taken from the mid-myocardial layer of the left ventricle) were submerged in the saline bath. The animal tissue was purchased from a local butcher, and no procedures were performed on live animals.
The tip of the ablation catheter was held parallel to the myocardial tissue surface and maneuvered by means of the control knob to obtain various CF intensities at different stages of the experiment.
CF, temperature, impedance, FTI, and LSI values were continuously monitored during RF delivery by using the EnSite Velocity cardiac mapping system (St. Jude Medical). A calibrated roller pump (CoolPoint, St. Jude Medical) connected to the catheter delivered saline solution at 17 ml/min during RF delivery. An Ampere RF generator (St. Jude Medical) was connected to deliver 550 kHz unmodulated sine-wave RF energy pulses in a temperature-controlled mode (maximum temperature 41°C). Different power settings and CF intensities were used during the experiment, as discussed in the following sections.
To obtain a range of lesion sizes, myocardial lesions were first created at separate sites with constant CF intensity (10 g) and power settings of 20, 25, and 30 W. For each power setting, RF delivery was maintained to achieve a given LSI value (5, 6, 7, and 8). Three separate lesions were obtained for each LSI target value (total of 36 lesions) (Figures 2A and 2B).
In the second phase of the study, additional myocardial lesions were created with a fixed power (20 W) and 2 different force settings (20 and 30 g). For each force value, RF pulse delivery time was titrated to achieve LSI values of 5, 6, 7, and 8. Three separate lesions were obtained for each LSI target value (total of 24 lesions).
After every lesion, the ablation electrode and the electrode–tissue interface were examined for char formation. The occurrence of an audible pop was noted (Figure 3).
The study design excluded lesions with LSI lower than 5, which needed a very short time of RF delivery, thus preventing complete lesion formation. Moreover, lesions with LSI higher than 8 required excessively long RF pulses (>90 s) even with high-power settings (Table 1) and were therefore not considered (steady state for lesion formation is typically reached between 45 and 60 s) (7).
After RF delivery, the myocardium was cross-sectioned at the level of each lesion. The maximum width and depth values of the blanched zone of the lesion (Figure 4) were measured with a dial caliper with a resolution of 0.1 mm by 1 observer; the observer was not blinded to the lesion protocol.
Data are expressed as mean ± SD. Descriptive statistics were calculated for variables of interest.
The significance of the relationship between lesion size (depth and width) and the different levels of power, force, FTI, and LSI was assessed by means of univariate and multivariate linear regression analysis and analysis of variance. Depth or width was used as the dependent variable and LSI, power, and force as effects. Significant differences were further evaluated by using Bonferroni's method for pairwise multiple comparisons. Values of p < 0.05 were considered statistically significant. Statistical analysis was performed by using IBM-SPSS version 23.01 (IBM SPSS Statistics, IBM Corporation, Armonk, New York) and STATISTICA version 12 (StatSoft, Tulsa, Oklahoma) statistical software.
A total of 60 RF applications were delivered (15 lesions for every LSI value, obtained by varying the intensity of CF or power). Mean impedance, as measured at the tip of the ablation catheter while in contact with the myocardium at the beginning of RF delivery, was 65 Ω Ohms and did not significantly change between the 4 CFs used. The width and depth of the obtained lesions are reported in Figures 2A and 2B and Table 2.
With LSI level 5, the mean lesion width was 6.3 ± 0.4 mm, and mean lesion depth was 2.7 ± 0.4 mm. Ablation lesions targeting the lowest LSI value did not result in any popping phenomena. With LSI level 6, mean lesion width was 7.8 ± 0.5 mm, and mean lesion depth was 3.9 ± 0.2 mm. One pop was heard and crater formation was noted (Figure 3) during high-power (30 W, fixed power) ablation. With LSI level 7, mean lesion width was 9.0 ± 0.4 mm, and mean lesion depth was 5.0 ± 0.2 mm. Two pops were heard, with thrombus and crater formation during high-power (30 W) ablation. With LSI level 8, mean lesion width was 9.9 ± 0.4 mm, and mean lesion depth was 6.0 ± 0.2 mm. One pop was heard, with thrombus and crater formation during high-power (30 W) ablation.
The analysis of variance showed a great effect of LSI increment on lesion width (p < 0.0001) and depth (p < 0.0001). The Bonferroni’s post hoc test confirmed that the lesion size (both width and depth) differed for each level of LSI (LSI 5 vs. LSI 6, and LSI 6 vs. LSI 7, and LSI 7 vs. LSI 8, and LSI 6 vs. LSI 7, and LSI 6 vs. LSI 8, and LSI 7 vs. LSI 8 were significantly different, p < 0.001).
In summary, both lesion width and depth increased with increasing LSI target values, regardless of power values. For instance, lesions obtained with LSI level 8 using low CF (10 g) and power (20 W) settings were significantly larger than lesions obtained with higher CF (30 g) and similar power values (20 W) but targeting an LSI level of 5 (median width 9.9 mm vs. 6.3 mm; median depth 6 mm vs. 2.7 mm).
Moreover, the time needed to achieve each LSI value decreased with increasing power or force values, whereas the FTI values increased with increasing LSI values. Instead, FTI correlation with power, and especially with force, was weaker.
Factors affecting ablation lesion size
According to multiple linear regression analysis, ablation lesion size did not seem to correlate with power or CF (Table 3), whereas a significant correlation was seen between both FTI (p < 0.01) and LSI (p < 0.0001) (Figure 5) and lesion width and depth.
Figure 6A displays the same relationships but highlighting 4 intervals of FTI. Figure 6B displays the relationships between the width and depth of lesions obtained by using the 4 different LSI levels. Such correlation resulted stronger for LSI than for FTI: the linear correlation coefficients for lesion dimensions resulted higher for LSI (r = 0.95 for width and r = 0.97 for depth) than for FTI (r = 0.66 for width and r = 0.71 for depth) (Figure 5).
Impedance drop during RF delivery was analyzed but did not exhibit any significant correlation with lesion size.
Factors affecting ablation safety
No popping phenomena were observed with an RF power of 20 or 25 W, irrespective of the targeted LSI level (Table 1). Only when the power was set to 30 W did steam pops occur in 4 cases (1 pop with LSI level 6, 2 pops with LSI level 7, and 1 pop with LSI level 8). In cases in which pop occurred, the impedance values at the beginning and at the end of RF delivery were similar to those recorded in cases without pops, and no sudden impedance drop was noted.
The main finding of the present in vitro study is that LSI is a better predictor of RF lesion dimensions than power, impedance CF, and FTI. In other words, given that the LSI formula takes into account force, power, and impedance, LSI was more predictive than each of its components if considered alone.
Moreover, the plotted diagrams of width/depth and LSI allowed prediction of both the LSI value required to obtain a pre-specified lesion width/depth and prediction of the lesion width/depth after a RF lesion performed targeting a pre-specified LSI value (Figure 7).
Correlation between RF power and lesion width/depth
As mentioned earlier, in the first part of the study, 36 lesions were created by using a fixed force intensity (10 g) and 3 distinct power settings (20, 25, and 30 W [3 lesions for each power setting]) to achieve 4 different pre-specified LSI levels (5, 6, 7, and 8). To limit the number of lesions needed, the dataset obtained at each successive step was used to verify that the lesion size was independent of power output. Indeed, this assumption was supported by the lack of significant correlation between power/lesion width (p = 0.31) (Table 3) and power/lesion depth (p = 0.67). Therefore, additional sets of lesions were created only with a fixed power setting of 20 W and CF values of 20 and 30 g to achieve the 4 different LSI levels (5, 6, 7, and 8).
Comparison between FTI and LSI
The value of FTI in predicting RF lesion dimensions has been previously explored and validated in both in vitro and clinical studies (3,5). However, FTI describes only 2 (time and force) of many factors involved in RF lesion formation, and it ignores the crucial role of power delivery in lesion creation (3). Conversely, LSI not only comprises CF, power, and time but describes other aspects of the biophysics of RF lesion formation (e.g., resistive heating and conductive heating). Therefore, LSI would be a more robust predictor of RF lesion size than FTI, as confirmed by our observations.
Clinical utility of LSI
An audible pop and crater formation was noted only when the delivered power was 30 W; no complication was encountered with lower power values. Therefore, given that the most important determinant of lesion size in this study was LSI, even at low-power settings, it might be suggested that effective lesions could be safely created by using power settings <30 W, once it has been ascertained that appropriate LSI values have been obtained. By doing so, the required lesion width (e.g., to avoid lesion gaps when a linear lesion is required) and depth (e.g., to ensure lesion transmurality during pulmonary vein isolation) can be achieved even with low-power settings, thus enhancing both ablation effectiveness and safety (fewer steam pops).
As discussed earlier, our results suggest that, at least in an in vitro setting, the LSI value needed to obtain a pre-specified lesion width/depth can be predicted, as well as the resulting lesion width/depth after an RF lesion performed with a given LSI value (Figure 7).
If confirmed in in vivo studies, these data might provide some guidance for a safe and effective setting of the power delivered during ablation with an irrigated-tip contact catheter.
The primary limitation of this study is that it was performed on an in vitro model, which differed from a clinical setting in a number of significant ways, including the absence of cardiac motion. Moreover, the porcine mid-myocardial muscle surface is smooth, whereas the human endocardium is trabeculated, which often hampers catheter stability. This in vitro setting was chosen to easily control electrode orientation and CF, and it can be regarded as a proof of concept for future in vivo experiments.
The catheter tip was held parallel to the myocardial surface in this study; no data were collected with the catheter in perpendicular contact. Data from the literature on this issue are controversial: some authors obtained larger lesions by means of parallel contact (8–10), whereas others suggested that catheter orientation does not play a role in determining lesion size (11). The parallel orientation in our study was derived from the observation that in clinical practice, most lesions are created with such orientation. During RF ablation for right atrial flutter (12) and atrial fibrillation ablation (13), the ablation catheter tip is mostly held parallel to the myocardial surface.
Lesions were measured by a nonblinded observer, thus introducing a possible expectancy bias. In similar studies, other authors (1,3) have reported not only lesion width and depth (2,14) but also lesion volume; lesion volume is derived from width and depth measurements according to the formula for an oblate ellipsoid and may prove less accurate than width and depth, which can, by contrast, be directly measured. Moreover, when RF ablation is performed in clinical practice, lesion width and depth are more relevant indexes than lesion volume in terms of achieving success (e.g., lesion transmurality and continuity).
Another point is that the average impedance measured in our study (65 Ω) differs significantly from the typical values recorded during RF ablation in humans (approximately 110 Ω).
Moreover, the presence in our experimental setup of saline bath at room temperature instead of heparinized blood at 37°C may have introduced further discrepancies from a clinical setting. Lastly, in in vivo settings, stable, pre-specified power levels may not be easily achieved because of catheter instability, ablation site anatomy (e.g., pouches), and variable flow velocity of the surrounding blood.
The main scope of our study, however, was to achieve proof of principle of the predictive power of LSI in predicting RF lesion dimensions if compared with other indexes (e.g., FTI, CF) and not to determine absolute LSI values to translate as they are in the clinical practice. Our results therefore need to be confirmed prospectively in in vivo studies.
LSI was found to be highly predictive of RF lesion width and depth in the in vitro model.
At similar LSI values, lesions obtained by using lower power settings were similar in magnitude, but exhibited a superior safety profile (no steam pops), compared with lesions obtained by means of high-power settings. Further studies are warranted to explore the predictive value of parameters provided by CF catheters, including LSI, in in vivo and human settings.
COMPETENCY IN MEDICAL KNOWLEDGE: RF catheter ablation is a well-established therapy for several cardiac arrhythmias. Proper RF delivery is crucial both to obtain an effective tissue lesion and to avoid excessive tissue heating, steam pop formation, and cardiac perforation. Traditionally, parameters such as RF power, catheter tip temperature, and impedance are monitored during RF delivery. However, these parameters display limited accuracy in assessing both the effectiveness and safety of the procedure, and more reliable lesion quality markers are needed.
TRANSLATIONAL OUTLOOK: Our results showed that LSI is the best in vitro predictor of RF lesion width and depth compared with power, CF, RF time, and FTI. If confirmed in in vivo and human studies, LSI would result in a useful guide for delivering a safer and more effective RF ablation therapy.
The authors are grateful to Franco Noventa (Senior Scientist, Department of Molecular Medicine, University of Padua, Padua, Italy) for the statistical analysis.
Drs. Indiani and Licciardello are St. Jude Medical Italia employees. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
All authors attest they are in compliance with human studies committees and animal welfare regulations of the authors’ institutions and Food and Drug Administration guidelines, including patient consent where appropriate. For more information, visit the JACC: Clinical Electrophysiology author instructions page.
- Abbreviations and Acronyms
- contact force
- force–time integral
- lesion size index
- Received June 1, 2017.
- Revision received August 10, 2017.
- Accepted August 31, 2017.
- 2017 American College of Cardiology Foundation
- Yokoyama K.,
- Nakagawa H.,
- Shah D.C.,
- et al.
- Ikeda A.,
- Nakagawa H.,
- Lambert H.,
- et al.
- ↵Lambert H, Olstad SJ, Fremont OB. Prediction of atrial wall electrical reconnection based on contact force measured during RF ablation. United States Patent Application Publication. Pub. No. US 2012/0209260 A1. Publication Date August 16, 2012.
- Haines D.E.
- Nakagawa H.,
- Wittkampf F.H.,
- Yamanashi W.S.,
- et al.
- Knecht S.,
- Reichlin T.,
- Pavlovic N.,
- et al.