- Email: [email protected]
Clinical value of assessment of left atrial late gadolinium enhancement in patients undergoing ablation of atrial fibrillation
International Journal of Cardiology 179 (2015) 351–357
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Clinical value of assessment of left atrial late gadolinium enhancement in patients undergoing ablation of atrial fibrillation Marek Sramko a,⁎, Petr Peichl a, Dan Wichterle a, Jaroslav Tintera b, Jiri Weichet c, Radoslav Maxian a, Silvia Pasnisinova a, Radka Kockova a, Josef Kautzner a a b c
Department of Cardiology, Institute for Clinical and Experimental Medicine (IKEM), Videnska 1958/9, 14021 Prague, Czech Republic Department of Radiology, Institute for Clinical and Experimental Medicine (IKEM), Videnska 1958/9, 14021 Prague, Czech Republic Department of Radiology, Na Homolce Hospital, Roentgenova 2/37, 15030 Prague, Czech Republic
a r t i c l e
i n f o
Article history: Received 12 August 2014 Received in revised form 1 October 2014 Accepted 5 November 2014 Available online 7 November 2014 Keywords: Magnetic resonance imaging Fibrosis Atrial fibrillation Ablation
a b s t r a c t Background: Left atrial (LA) fibrosis begets atrial fibrillation (AF). Cardiovascular magnetic resonance (CMR) using the late gadolinium enhancement (LGE) technique might visualize the LA fibrosis and thus help to choose an appropriate strategy for treatment of AF. In this regard, we investigated whether the extent of preablation LA LGE would predict AF recurrence after ablation in a non-selected patient population. Methods: CMR was performed in 95 patients before radiofrequency ablation of AF. An interpretable scan was available in 73 patients (age, 59 ± 8 years; men, 71%; persistent/paroxysmal AF, 55/45%). The extent of LA LGE was quantified by three established thresholding techniques. In addition, CMR was used to quantify LA volume and reservoir function. The patients were followed for AF recurrence for 1.3 ± 0.8 years. Results: The arrhythmia recurred in 29 (40%) of the patients. The extent of LA LGE did not differ between paroxysmal and persistent AF and it did not predict the AF recurrence. Moreover, the extent of LA LGE did not correlate with LA volume, reservoir function and bipolar voltage. Conclusions: Our data indicate a limited value of a routine assessment of LA LGE before ablation of AF. Further experimental and clinical researches should be done before applying the method to a wide clinical practice. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Radiofrequency catheter ablation has proved an effective therapy of atrial fibrillation (AF) [1]. Yet, efficiency of the treatment depends mainly on selecting suitable candidates. One possible strategy for stratification of the patients for catheter ablation is to quantify the extent of fibrosis in the left atrium (LA). The rationale for this approach is based on the evidence that more extensive LA fibrosis begets AF [2]. Late gadolinium enhancement cardiovascular magnetic resonance imaging (LGE–CMR) appears a promising tool for assessing of the LA fibrosis in vivo [3]. The study by Oakes et al. and several subsequent studies have demonstrated the use of the LGE–CMR for predicting the clinical outcome of AF ablation [4–7]. But the evidence for use of this technique comes only from a few specialized centers. As a result, assessment of the LA fibrosis by LGE–CMR has not yet been widely adopted in clinical practice. Our study aimed to evaluate the clinical value of the LA LGE–CMR in a non-selected patient population undergoing ablation for AF. Because the quantification of LA LGE may be biased by inter-individual ⁎ Corresponding author at: Department of Cardiology, IKEM, Videnska 1958/9, 140 21 Prague, Czech Republic. E-mail address: [email protected] (M. Sramko).
http://dx.doi.org/10.1016/j.ijcard.2014.11.072 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
variability [8], we have quantified the LGE by three different techniques. We expected that patients with more extensive preablation LA LGE would have a higher rate of AF recurrence after ablation. To further evaluate validity of the LA imaging, we investigated the relationship between the LA LGE and other LA characteristics that are pathophysiologically related to the LA fibrosis, such as LA volume, reservoir function and endocardial bipolar voltage. 2. Methods 2.1. Study population The study included 95 consecutive patients undergoing catheter ablation of AF. Only individuals without a previous ablation and without a contraindication for CMR were included. All patients underwent magnetic resonance imaging within a week before the ablation. After ablation, the patients were followed-up during regular visits at 3 and 6 months, and every 6 months thereafter. Each follow-up examination included 24 h Holter ECG monitoring. In addition, patients who continued complaining of palpitations underwent repeated Holter ECG monitoring (up to 7 days) or monitoring with a portable event recorder, as necessary, until the underlying etiology of the palpitations was clarified. Withdrawal of antiarrhythmics during the follow-up was left on the clinical judgment of the treating physician, who was blinded to the CMR results. AF recurrence was considered only after a post-ablation blanking period of 3 months. The recurrence was defined as any detected episode of AF or atrial tachycardia lasting N30 s [4]. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki. All patients gave written consent with the investigation.
352
M. Sramko et al. / International Journal of Cardiology 179 (2015) 351–357
2.2. Cardiovascular magnetic resonance The imaging was performed on a 1.5 Tesla Scanner (Avanto, Siemens Medical Solutions, Erlangen, Germany) using a 12-channel body coil. A stack of contiguous short-axis cine images of the entire LA was acquired to assess the LA function (typical temporal resolution of 40 ms, voxel size of 2.1 × 1.6 × 8 mm). Subsequently, time-resolved 3D angiography of the LA (voxel size of 1.7 × 1.7 × 1.7 mm) was obtained with a bolus injection of 0.2 mmol/kg of gadobutrol. Acquisition of the LGE images started 10 min after administration of the contrast using a segmented 3D FLASH sequence with FatSat. Typical parameters were: TR/TE of 4.8/1.5 ms, TI of 270 ms, 1 inversion pulse per RR, flip angle of 10°, linear k-space filling, FOV of 400 × 400 × 100 mm, voxel size of 1.6 × 1.6 × 3 mm interpolated to 0.78 × 0.78 × 1.7 mm. The images were acquired during free breathing with a navigator (PACE, acceptance window of ±2.5 mm at end-expiration). ECG-triggering was adjusted using a 4-chamber view cine loop to obtain a subset of images during a cardiac phase with the least LA motion. The average acquisition time for the LGE-MRI study was 11 ± 4 min.
2.3. Image analysis The extent of the LA LGE was quantified according to Oakes et al. [4] and by additional two automated techniques. First, the LA wall was manually segmented using Slicer 4.2 (Scientific Computing and Imaging Institute, Salt Lake City, Utah). The LA boundaries were anteriorly at the mitral annular plane and posteriorly at the pulmonary vein ostia. The segmented image stack was processed using a custom program written in Matlab. For each slice the program calculated intensity histogram of the voxels belonging to the LA wall. Voxels with intensity values within the 2nd and the 40th percentiles of the histogram (i.e., dark voxels) were classified as a reference tissue. The examiner then interactively selected a threshold at 3, 4 or 5 standard deviations (SDs) above the mean signal intensity of the reference tissue — voxels brighter than the threshold were classified as abnormally enhanced (Fig. 1). The extent of the LA LGE was calculated as the percentage of the abnormally enhanced voxels of all the LA wall voxels. The examiner's choice of the particular threshold was based on the image contrast-to-noise ratio (Fig. 2). The most frequently selected threshold was at 4 SDs (64% of the patients), followed by 3 SDs (19% of the patients) and 5 SDs (17% of the patients). The second method for quantification of LGE was similar to the method above. However, instead of selecting a threshold by the examiner, a fixed threshold was applied for all patients at 2, 3 and 4 SDs above the mean signal intensity of the reference tissue. Finally, the LGE was quantified by the full-width-at-half-maximum method [8]. LA volume was determined by manual segmentation of the 3D angiogram [9]. LA reservoir function, represented by LA emptying fraction, was assessed on the cine images according to Jarvinen [10]. The images were analyzed by a single CMR specialist (M.S.), who was blinded to the follow-up data. Subjective image quality was assessed at the end of each CMR exam. Scans that contained unidentifiable LA borders or significant artifacts were discarded. To assess intra- and inter-observer reproducibility of the image analysis, 10 scans with an
acceptable image quality were re-evaluated by the primary examiner and another two CMR specialists (J.W. and R.K.). 2.4. Catheter ablation Our method of ablation has been described in detail elsewhere [11]. The procedure was guided by a 3D mapping system (CARTO, Biosense Webster, Diamond Bar, CA) and intra-cardiac echocardiography. Ablation was performed using a 3.5-mm irrigated-tip catheter (NaviStar Thermocool, Biosense Webster, Diamond Bar, CA). The pulmonary veins were isolated by circumferential point-by-point lesions; additional linear lesions were performed in the patients with persistent AF. Radiofrequency energy of up to 35 W was applied at each point for 30–60 s. At the end of the procedure, patients who remained in AF were electrically cardioverted. A high-density LA voltage map (219 ± 49 equally spaced points) was obtained before ablating. The electrograms were filtered at 30–500 Hz. To correct for the influence of heart rhythm on the electrogram amplitudes, the voltage maps were obtained during either spontaneous or pacing-induced AF. Low-voltage points were defined by two cut-offs: ≤0.2 and 0.5 mV [4,12]. The mean bipolar voltage and the percentage of the low-voltage points were compared with the extent of LA delayed enhancement. All operators were blinded to the CMR results. 2.5. Statistical analysis Data are reported as the mean ± standard deviation or frequency and percentage. Reproducibility of the LA wall segmentation was evaluated by Dice's coefficient. Agreement between the observers in the quantification of LA LGE was assessed by Pearson's correlation test. Associations between the investigated variables and AF recurrence were explored using Student's t-test or chi-square test, as appropriate. The patients were divided into groups according to the extent of the LA LGE (Table 2). The groups were compared with regard to AF recurrence by one-way ANOVA test. Mutual relationship between the extent of LA LGE, volume, reservoir function and bipolar voltage was analyzed by Pearson's correlation test. P values b 0.05 were considered statistically significant. All analyses were conducted in R (http://www.R-project.org).
3. Results 3.1. Feasibility of LA LGE–CMR CMR was completed in 91 patients. Of the completed scans, 17 (19%) were corrupted mostly by motion blurring, contrast flow artifacts or diffuse noise (Fig. 3). The artifacts occurred more often in the patients who had AF at the time of the imaging, compared to those who were in sinus rhythm (10 of 30 [33%] vs. 7 of 60 [12%], P = 0.01). In total,
Fig. 1. The figure outlines the principles of the quantification of LA LGE based on the analysis of the image histogram. (A) A native LGE–CMR image of the LA. (B) The identical image with overlaid manual segmentation of the LA wall (blue) and blood pool (green). The areas of LGE (yellow) were identified by the supervised thresholding technique according to Oakes et al. [4]. (C) An image histogram of the LA wall voxels shown on the panel B. The abnormally enhanced voxels are separated from the rest of the LA tissue by arbitrary thresholds. (D) The final thresholded image was reconstructed in 3D using volume surface rendering. The red arrow indicates a corresponding region of abnormal delayed enhancement.
M. Sramko et al. / International Journal of Cardiology 179 (2015) 351–357
353
Fig. 2. The figure demonstrates the patient-specific selection of an arbitrary threshold at 3, 4 or 5 SDs above the mean signal intensity of the reference tissue. (Patient #1) The thresholds at 4 and 5 SDs underestimate the extent of the LGE. A threshold at 3 SDs seems to be the most appropriate. (Patient #2) The extent of LGE seems slightly overestimated at 3 SDs and largely underestimated at 5 SDs. In this patient, the examiner chose the 4 SD threshold. (Patient #3) This image has a low contrast-to-noise ratio. Therefore the thresholds below 5 SDs tend to overestimate the extent of the LGE. In this patient the examiner chose the 5 SD threshold.
an interpretable scan for analysis of LGE was available in 73 (77%) of the study participants (age, 59 ± 8 years; men, 71%; persistent/paroxysmal AF, 55/45%). The baseline characteristics of the patients with an interpretable LGE–CMR scan are presented in Table 1. Of note, besides having AF more often during the imaging, the excluded patients did not differ significantly from those with the interpretable CMR scan. 3.2. Reproducibility of image analysis At the repeated analysis, the primary examiner remained fairly consistent in the manual segmentation of the LA wall (Dice's coefficient, 0.65 ± 0.09) and also in the quantification of the LGE (absolute paired difference, 6 ± 5% points; R = 0.67; P = 0.046). In contrast, there was a poor inter-observer agreement between the three examiners, both in the segmentation of the LA wall (average Dice's coefficient, 0.42 ± 0.09) and also in the quantification of the LA LGE (absolute paired difference, 11 ± 8% points; R = 0.29, P = 0.4). The poor reproducibility of the image analysis resulted mainly from ambiguity of the LA images, especially at the level of the LA roof and the mitral annulus (Fig. 4). 3.3. LA LGE and outcome of catheter ablation AF recurred in 29 (40%) patients during a mean follow-up of 1.3 ± 0.8 years (range, 4 months to 3.5 years). All the AF recurrences were clinically relevant: 12 patients underwent reablation(s), another 14 underwent electric cardioversion(s) and the remaining 3 patients had
symptomatic AF paroxysms which lasted 49 min to N24 h on Holter monitoring. At the time of the AF recurrence, 13 (45%) of the patients were using antiarrhythmic drugs compared to 12 (27%) of those without arrhythmia recurrence (P = 0.1). In univariate analysis, the arrhythmia recurrence was associated with the persistent form AF, increased LA volume, lower LA bipolar voltage, the need for more extensive ablation and persistence of the arrhythmia at the end of the procedure (Table 1). Importantly, the extent of the LA LGE did not differ between the patients with and without AF recurrence (16 ± 8 vs. 18 ± 10%, P = 0.4), irrespective of whether they had paroxysmal AF (17 ± 9 vs. 20 ± 10%, P = 0.4) or persistent AF (15 ± 8 vs. 17 ± 10%, P = 0.5) and irrespective of the method for quantification of the LGE. Table 2 presents the rates of AF recurrence according to the stage of the LGE; a Kaplan–Meier analysis of the arrhythmia-free survival is shown in Fig. 5.
3.4. Extent of LA LGE and LA structural remodeling A high density voltage map along with an interpretable LGE–CMR scan was available in 56 individuals. We found no correlation between the extent of the LA LGE and LA voltages, volume or reservoir function, irrespective of the method for quantification of the LGE. On the other hand, lower LA voltage correlated with increased volume and worse reservoir function (Fig. 6); and the LA voltage negatively correlated with reservoir function (R = −0.51, P b 0.001). There was also a poor visual correlation between the spatial distribution of the LGE and the
354
M. Sramko et al. / International Journal of Cardiology 179 (2015) 351–357
Fig. 3. The study flowchart indicating feasibility of the LA LGE–CMR.
low-voltage areas (whether using the 0.2 or the 0.5 mV cut-off). A representative sample of the visual comparisons is shown in Fig. 7.
4. Discussion 4.1. Major findings
Table 1 Comparison of the patients' characteristics according to recurrence of atrial fibrillation. Variable
Clinical characteristics Male Age (years) Body mass index (kg/m2) Lone AF Persistent AF AF persisting N12 months History of AF (years) CHA2DS2-VASc score Antiarrhythmic drugs at baseline Magnetic resonance imaging LA volume (mL) LA emptying fraction (%) Extent of LA LGE, method by Oakes et al. (%) [4] Extent of LA LGE, FWHM method (%) Extent of LA LGE, 2 SDs threshold (%) Extent of LA LGE, 3 SDs threshold (%) Extent of LA LGE, 4 SDs threshold (%) Radiofrequency ablation Mean LA bipolar voltage (mV) Proportion of low voltage LA points (%) Cumulative radiofrequency energy (W min) AF non-termination by ablation Echocardiography LA parasternal dimension (mm) Left ventricular ejection fraction (%) Tissue Doppler E/E′ Mitral regurgitation (Grades 1–4)
No. of AF recurrence n = 44
AF recurrence n = 29
P value
33 (75%) 59 ± 8 29 ± 4 14 (32%) 20 (45%) 1 (2%) 5±3 1.4 ± 1.2 15 (34%)
19 (66%) 60 ± 7 30 ± 5 6 (21%) 20 (69%) 5 (17%) 4±4 1.7 ± 1.1 14 (48%)
0.38 0.35 0.32 0.28 0.048a 0.033a 0.49 0.25 0.22
89 ± 32 30 ± 16 18 ± 10
106 ± 36 26 ± 18 16 ± 8
0.047a 0.25 0.38
21 46 29 17
21 45 29 17
0.63 0.52 0.55 0.93
± ± ± ±
1 2 3 3
± ± ± ±
2 3 5 5
1.1 ± 0.7 46 ± 25 1161 ± 520
0.8 ± 0.5 60 ± 22 1599 ± 654
0.047a 0.034a 0.004a
16 (36%)
19 (66%)
0.015a
43 ± 6 56 ± 7 7.6 ± 1.9 1.3 ± 0.5
44 ± 6 52 ± 4 7.9 ± 2.5 1.3 ± 0.5
0.71 0.69 0.35 0.97
AF = atrial fibrillation; LA = left atrial; LGE = late gadolinium enhancement. a Signifies P b 0.05.
First, the extent of preablation LA LGE did not predict AF recurrence during post-ablation follow-up, regardless of the used image processing method. On the other hand, other traditionally reported factors, such as the type of AF, increased LA volume and lower LA voltage were significantly associated with the AF recurrence. Second, validity of the LGE– CMR was undermined by the fact that the extent of the LGE did not correlate with other LA characteristics that are related to the LA structural remodeling. At last, in our non-selected patient population the LGE scans were frequently corrupted by unavoidable imaging artifacts. The poor image quality was particularly common in the patients with ongoing AF, in whom every third examination was uninterpretable. 4.2. Previously published data Recent studies conducted at the CARMA center in Utah found a strong association between preablation LA LGE and AF recurrence after ablation [4,5,13]. The extent of the LGE correlated with lower LA voltage [4] and decreased LA strain [14]. The authors also proposed adjusting the ablation technique in each individual patient based solely on the extent of LA LGE [3]. The CARMA group has recently conducted a multicenter trial DECAAF which included 260 patients from 15 centers in 6 countries [7]. The extent of preablation LA LGE independently predicted AF recurrence after ablation [4]. Unfortunately, the findings of the DECAAF trial may be difficult to extrapolate to a wide clinical practice, because all the images were analyzed only by the CARMA group [15]. Using a different approach to the image analysis, a group of authors from London also reported association between preablation LA LGE and clinical outcome after catheter ablation [6]. The normalized LA signal intensity correlated with LA volume and mean LA bipolar voltage. In addition, the authors demonstrated that a suboptimal LGE sequence could not reliably detect even postablation LA scarring [16]. This finding implies that only an optimal quality scan may be used for assessment of the LA LGE.
M. Sramko et al. / International Journal of Cardiology 179 (2015) 351–357
355
Fig. 4. Typical LA LGE images.
Our study is comparable to the study by Oakes et al. [4], in terms of size, clinical endpoints and the approach to quantification of LGE. Based on our clinical experience we chose slightly different parameters of the LGE sequence and a different contrast agent. Scanning of the LA started somewhat earlier after contrast injection, because acquisition of the whole LA image stack took often over 15 min (especially in the individuals with irregular fast heart rate). We believe, however, that these differences have not fundamentally influenced our results. In fact, recently we have demonstrated that the same method of LGE–CMR can reliably detect even discrete changes in serially assessed LGE in the left ventricle during the course of dilated cardiomyopathy [17]. In agreement with our experience, Oakes had to discard 25% of the scans because of poor image quality. Although the average extent of LA LGE did not differ between the two studies, in our patients the LGE did not predict AF recurrence. We can only speculate whether the opposing findings of our study could be explained by the use of a different ablation technique or merely by the inherent subjective variability in the
image processing. On the other hand, our results were consistent even after avoiding the inter-observer variability by employing automated techniques for quantification of LGE [8]. Finally, we must underline, that assessment of preablation LA fibrosis by LGE–CMR has not yet been properly validated by systematic histological studies in human or in an animal model. In fact, assessment of LA LGE may be challenging even for detecting a dense postablation scar in a rigorously controlled experiment [18]. Therefore, while the finding of LGE in the left ventricle corresponds well to a compact fibrotic scar [19], and areas of LGE may reflect a scar after a previous ablation [18], LGE on a preablation scan should not be interpreted as LA fibrosis without any precautions. 4.3. Limitations Our study was powered only to investigate predictors of AF recurrence after a single ablation procedure. Also, including a non-selected patient
Table 2 Recurrence of atrial fibrillation according to the stage of LA LGE. Stage of LA LGE (% cut-off)
Tertiles of the present study sample 1st tertile (b13%) 2nd tertile (13–20%) 3rd tertile (N20%) Staging according to Oakes et al. [4] Minimal enhancement (b15%) Moderate enhancement (15–35%) Extensive enhancement (N35%)
AF recurrence
ANOVA F-statistics
P value
11/24 (46%) 9/24 (38%) 9/25 (36%)
0.28
0.8
14/34 (41%) 14/34 (41%) 1/5 (20%)
0.19
0.8
0.41
0.94
Staging according to the DECAAF trial [7] Stage I (b10%) 7/18 (39%) Stage II (10–19%) 10/24 (42%) Stage III (20–29%) 10/24 (42%) Stage IV (N30%) 2/7 (30%)
AF = atrial fibrillation; LA = left atrial; LGE = late gadolinium enhancement. The cut-off values in the brackets in the left column represent the relative extent of LA LGE.
Fig. 5. A Kaplan–Meier plot showing postablation AF-free survival depending on the preablation extent of LA LGE. The data are censored at 3 years of follow-up.
356
M. Sramko et al. / International Journal of Cardiology 179 (2015) 351–357
Fig. 6. The figure shows the relationship between LA LGE and other variables related to LA remodeling. Each point on the plots represents a mean value of the analyzed variable in one of the N patients with available quality data. R and P represent Pearson's correlation coefficients and the corresponding P-values. In this figure, the extent of LA LGE was quantified according to Oakes et al. [4]. However, similar results were achieved also by the other two thresholding techniques. EA map = electroanatomic map.
population resulted to a high heterogeneity of the sample. On the other hand, such study population better reflects clinical reality. Restoring of sinus rhythm by electrical cardioversion before the CMR exam could have provided a better image quality in some patients, but such indication for electric cardioversion is currently not supported by sound evidence. Although decreased endocardial bipolar voltage is a generally accepted surrogate of fibrotic myocardium, the relationship between decreased LA voltage and fibrosis has been investigated only in postablation LA scarring [18]. Finally, we must underline that the aim of the study was not to evaluate accuracy of the LGE–CMR for quantification of LA fibrosis but to evaluate the use of the method—as it currently stands—for routine clinical practice in a non-selected patient population.
that are related to LA fibrosis. Before applying the method to clinical practice, it needs to be validated by solid histopathological evidence and efforts should be taken to improve the imaging and image processing technique. Conflict of interest All the authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. The authors report no relationships that could be construed as a conflict of interest. Funding sources
5. Conclusions This study found a limited clinical value of routine assessment of preablation LA LGE in patients undergoing ablation of AF. Moreover, the extent of LA LGE did not correlate with other LA characteristics
This study was supported by the Research Grant of the Ministry of Health of the Czech Republic — conceptual development of research organization (“Institute for Clinical and Experimental Medicine — IKEM, IN 00023001”).
M. Sramko et al. / International Journal of Cardiology 179 (2015) 351–357
357
Fig. 7. The figure shows three representative examples of a visual comparison between LA bipolar voltage maps and LGE images. The red-colored areas on the electroanatomic maps represent voltage b0.5 mV. The LGE images were reconstructed using maximum intensity projection rendering. The yellow color highlights abnormal delayed enhancement.
References [1] A.N. Ganesan, N.J. Shipp, A.G. Brooks, et al., Long-term outcomes of catheter ablation of atrial fibrillation: a systematic review and meta-analysis, J. Am. Heart Assoc. 2 (2013) e004549. [2] H. Kottkamp, Human atrial fibrillation substrate: towards a specific fibrotic atrial cardiomyopathy, Eur. Heart J. 34 (2013) 2731–2738. [3] G.R. Vergara, N.F. Marrouche, Tailored management of atrial fibrillation using a LGE– MRI based model: from the clinic to the electrophysiology laboratory, J. Cardiovasc. Electrophysiol. 22 (2011) 481–487. [4] R.S. Oakes, T.J. Badger, E.G. Kholmovski, et al., Detection and quantification of left atrial structural remodeling with delayed-enhancement magnetic resonance imaging in patients with atrial fibrillation, Circulation 119 (2009) 1758–1767. [5] N. Akoum, M. Daccarett, C. McGann, et al., Atrial fibrosis helps select the appropriate patient and strategy in catheter ablation of atrial fibrillation: a DE-MRI guided approach, J. Cardiovasc. Electrophysiol. 22 (2011) 16–22. [6] L.C. Malcolme-Lawes, C. Juli, R. Karim, et al., Automated analysis of atrial late gadolinium enhancement imaging that correlates with endocardial voltage and clinical outcomes: a 2-center study, Heart Rhythm. 10 (2013) 1184–1191. [7] N.F. Marrouche, D. Wilber, G. Hindricks, et al., Association of atrial tissue fibrosis identified by delayed enhancement MRI and atrial fibrillation catheter ablation: the DECAAF study, JAMA 311 (2014) 498–506. [8] R. Karim, R.J. Housden, M. Balasubramaniam, et al., Evaluation of current algorithms for segmentation of scar tissue from late gadolinium enhancement cardiovascular magnetic resonance of the left atrium: an open-access grand challenge, J. Cardiovasc. Magn. Reson. 15 (2014) 105. [9] C. Jahnke, J. Fischer, J.G. Mirelis, et al., Cardiovascular magnetic resonance imaging for accurate sizing of the left atrium: predictability of pulmonary vein isolation success in patients with atrial fibrillation, J. Magn. Reson. Imaging 33 (2011) 455–463. [10] V. Jarvinen, M. Kupari, P. Hekali, V.P. Poutanen, Assessment of left atrial volumes and phasic function using cine magnetic resonance imaging in normal subjects, Am. J. Cardiol. 73 (1994) 1135–1138.
[11] B. Aldhoon, D. Wichterle, P. Peichl, R. Cihak, J. Kautzner, Complications of catheter ablation for atrial fibrillation in a high-volume centre with the use of intracardiac echocardiography, Europace 15 (2013) 24–32. [12] M. Fiala, D. Wichterle, J. Chovancik, et al., Left atrial voltage during atrial fibrillation in paroxysmal and persistent atrial fibrillation patients, Pacing Clin. Electrophysiol. 33 (2010) 541–548. [13] C. Mahnkopf, T.J. Badger, N.S. Burgon, et al., Evaluation of the left atrial substrate in patients with lone atrial fibrillation using delayed-enhanced MRI: implications for disease progression and response to catheter ablation, Heart Rhythm. 7 (2010) 1475–1481. [14] S.S. Kuppahally, N. Akoum, N.S. Burgon, et al., Left atrial strain and strain rate in patients with paroxysmal and persistent atrial fibrillation: relationship to left atrial structural remodeling detected by delayed-enhancement MRI, Circ. Cardiovasc. Imaging 3 (2010) 231–239. [15] S. Vijayakumar, E. Kholmovski, M. Haslam, N. Burgon, N. Marrouche, Dependence of image quality of late gadolinium enhancement MRI of left atrium on number of patients imaged: results of multi-center trial DECAAF, J. Cardiovasc. Magn. Reson. 16 (Suppl. 1) (2014) 146. [16] R.J. Hunter, D.A. Jones, R. Boubertakh, et al., Diagnostic accuracy of cardiac magnetic resonance imaging in the detection and characterization of left atrial catheter ablation lesions: a multicenter experience, J. Cardiovasc. Electrophysiol. 24 (2013) 396–403. [17] M. Kubanek, M. Sramko, J. Maluskova, et al., Novel predictors of left ventricular reverse remodeling in individuals with recent-onset dilated cardiomyopathy, J. Am. Coll. Cardiol. 61 (2013) 54–63. [18] J.L. Harrison, H.K. Jensen, S.A. Peel, et al., Cardiac magnetic resonance and electroanatomical mapping of acute and chronic atrial ablation injury: a histological validation study, Eur. Heart J. 35 (2014) 1486–1495. [19] R.J. Kim, D.S. Fieno, T.B. Parrish, et al., Relationship of MRI delayed contrast enhancement to irreversible injury, infarct age, and contractile function, Circulation 100 (1999) 1992–2002.
Contents lists available at ScienceDirect
International Journal of Cardiology journal homepage: www.elsevier.com/locate/ijcard
Clinical value of assessment of left atrial late gadolinium enhancement in patients undergoing ablation of atrial fibrillation Marek Sramko a,⁎, Petr Peichl a, Dan Wichterle a, Jaroslav Tintera b, Jiri Weichet c, Radoslav Maxian a, Silvia Pasnisinova a, Radka Kockova a, Josef Kautzner a a b c
Department of Cardiology, Institute for Clinical and Experimental Medicine (IKEM), Videnska 1958/9, 14021 Prague, Czech Republic Department of Radiology, Institute for Clinical and Experimental Medicine (IKEM), Videnska 1958/9, 14021 Prague, Czech Republic Department of Radiology, Na Homolce Hospital, Roentgenova 2/37, 15030 Prague, Czech Republic
a r t i c l e
i n f o
Article history: Received 12 August 2014 Received in revised form 1 October 2014 Accepted 5 November 2014 Available online 7 November 2014 Keywords: Magnetic resonance imaging Fibrosis Atrial fibrillation Ablation
a b s t r a c t Background: Left atrial (LA) fibrosis begets atrial fibrillation (AF). Cardiovascular magnetic resonance (CMR) using the late gadolinium enhancement (LGE) technique might visualize the LA fibrosis and thus help to choose an appropriate strategy for treatment of AF. In this regard, we investigated whether the extent of preablation LA LGE would predict AF recurrence after ablation in a non-selected patient population. Methods: CMR was performed in 95 patients before radiofrequency ablation of AF. An interpretable scan was available in 73 patients (age, 59 ± 8 years; men, 71%; persistent/paroxysmal AF, 55/45%). The extent of LA LGE was quantified by three established thresholding techniques. In addition, CMR was used to quantify LA volume and reservoir function. The patients were followed for AF recurrence for 1.3 ± 0.8 years. Results: The arrhythmia recurred in 29 (40%) of the patients. The extent of LA LGE did not differ between paroxysmal and persistent AF and it did not predict the AF recurrence. Moreover, the extent of LA LGE did not correlate with LA volume, reservoir function and bipolar voltage. Conclusions: Our data indicate a limited value of a routine assessment of LA LGE before ablation of AF. Further experimental and clinical researches should be done before applying the method to a wide clinical practice. © 2014 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Radiofrequency catheter ablation has proved an effective therapy of atrial fibrillation (AF) [1]. Yet, efficiency of the treatment depends mainly on selecting suitable candidates. One possible strategy for stratification of the patients for catheter ablation is to quantify the extent of fibrosis in the left atrium (LA). The rationale for this approach is based on the evidence that more extensive LA fibrosis begets AF [2]. Late gadolinium enhancement cardiovascular magnetic resonance imaging (LGE–CMR) appears a promising tool for assessing of the LA fibrosis in vivo [3]. The study by Oakes et al. and several subsequent studies have demonstrated the use of the LGE–CMR for predicting the clinical outcome of AF ablation [4–7]. But the evidence for use of this technique comes only from a few specialized centers. As a result, assessment of the LA fibrosis by LGE–CMR has not yet been widely adopted in clinical practice. Our study aimed to evaluate the clinical value of the LA LGE–CMR in a non-selected patient population undergoing ablation for AF. Because the quantification of LA LGE may be biased by inter-individual ⁎ Corresponding author at: Department of Cardiology, IKEM, Videnska 1958/9, 140 21 Prague, Czech Republic. E-mail address: [email protected] (M. Sramko).
http://dx.doi.org/10.1016/j.ijcard.2014.11.072 0167-5273/© 2014 Elsevier Ireland Ltd. All rights reserved.
variability [8], we have quantified the LGE by three different techniques. We expected that patients with more extensive preablation LA LGE would have a higher rate of AF recurrence after ablation. To further evaluate validity of the LA imaging, we investigated the relationship between the LA LGE and other LA characteristics that are pathophysiologically related to the LA fibrosis, such as LA volume, reservoir function and endocardial bipolar voltage. 2. Methods 2.1. Study population The study included 95 consecutive patients undergoing catheter ablation of AF. Only individuals without a previous ablation and without a contraindication for CMR were included. All patients underwent magnetic resonance imaging within a week before the ablation. After ablation, the patients were followed-up during regular visits at 3 and 6 months, and every 6 months thereafter. Each follow-up examination included 24 h Holter ECG monitoring. In addition, patients who continued complaining of palpitations underwent repeated Holter ECG monitoring (up to 7 days) or monitoring with a portable event recorder, as necessary, until the underlying etiology of the palpitations was clarified. Withdrawal of antiarrhythmics during the follow-up was left on the clinical judgment of the treating physician, who was blinded to the CMR results. AF recurrence was considered only after a post-ablation blanking period of 3 months. The recurrence was defined as any detected episode of AF or atrial tachycardia lasting N30 s [4]. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki. All patients gave written consent with the investigation.
352
M. Sramko et al. / International Journal of Cardiology 179 (2015) 351–357
2.2. Cardiovascular magnetic resonance The imaging was performed on a 1.5 Tesla Scanner (Avanto, Siemens Medical Solutions, Erlangen, Germany) using a 12-channel body coil. A stack of contiguous short-axis cine images of the entire LA was acquired to assess the LA function (typical temporal resolution of 40 ms, voxel size of 2.1 × 1.6 × 8 mm). Subsequently, time-resolved 3D angiography of the LA (voxel size of 1.7 × 1.7 × 1.7 mm) was obtained with a bolus injection of 0.2 mmol/kg of gadobutrol. Acquisition of the LGE images started 10 min after administration of the contrast using a segmented 3D FLASH sequence with FatSat. Typical parameters were: TR/TE of 4.8/1.5 ms, TI of 270 ms, 1 inversion pulse per RR, flip angle of 10°, linear k-space filling, FOV of 400 × 400 × 100 mm, voxel size of 1.6 × 1.6 × 3 mm interpolated to 0.78 × 0.78 × 1.7 mm. The images were acquired during free breathing with a navigator (PACE, acceptance window of ±2.5 mm at end-expiration). ECG-triggering was adjusted using a 4-chamber view cine loop to obtain a subset of images during a cardiac phase with the least LA motion. The average acquisition time for the LGE-MRI study was 11 ± 4 min.
2.3. Image analysis The extent of the LA LGE was quantified according to Oakes et al. [4] and by additional two automated techniques. First, the LA wall was manually segmented using Slicer 4.2 (Scientific Computing and Imaging Institute, Salt Lake City, Utah). The LA boundaries were anteriorly at the mitral annular plane and posteriorly at the pulmonary vein ostia. The segmented image stack was processed using a custom program written in Matlab. For each slice the program calculated intensity histogram of the voxels belonging to the LA wall. Voxels with intensity values within the 2nd and the 40th percentiles of the histogram (i.e., dark voxels) were classified as a reference tissue. The examiner then interactively selected a threshold at 3, 4 or 5 standard deviations (SDs) above the mean signal intensity of the reference tissue — voxels brighter than the threshold were classified as abnormally enhanced (Fig. 1). The extent of the LA LGE was calculated as the percentage of the abnormally enhanced voxels of all the LA wall voxels. The examiner's choice of the particular threshold was based on the image contrast-to-noise ratio (Fig. 2). The most frequently selected threshold was at 4 SDs (64% of the patients), followed by 3 SDs (19% of the patients) and 5 SDs (17% of the patients). The second method for quantification of LGE was similar to the method above. However, instead of selecting a threshold by the examiner, a fixed threshold was applied for all patients at 2, 3 and 4 SDs above the mean signal intensity of the reference tissue. Finally, the LGE was quantified by the full-width-at-half-maximum method [8]. LA volume was determined by manual segmentation of the 3D angiogram [9]. LA reservoir function, represented by LA emptying fraction, was assessed on the cine images according to Jarvinen [10]. The images were analyzed by a single CMR specialist (M.S.), who was blinded to the follow-up data. Subjective image quality was assessed at the end of each CMR exam. Scans that contained unidentifiable LA borders or significant artifacts were discarded. To assess intra- and inter-observer reproducibility of the image analysis, 10 scans with an
acceptable image quality were re-evaluated by the primary examiner and another two CMR specialists (J.W. and R.K.). 2.4. Catheter ablation Our method of ablation has been described in detail elsewhere [11]. The procedure was guided by a 3D mapping system (CARTO, Biosense Webster, Diamond Bar, CA) and intra-cardiac echocardiography. Ablation was performed using a 3.5-mm irrigated-tip catheter (NaviStar Thermocool, Biosense Webster, Diamond Bar, CA). The pulmonary veins were isolated by circumferential point-by-point lesions; additional linear lesions were performed in the patients with persistent AF. Radiofrequency energy of up to 35 W was applied at each point for 30–60 s. At the end of the procedure, patients who remained in AF were electrically cardioverted. A high-density LA voltage map (219 ± 49 equally spaced points) was obtained before ablating. The electrograms were filtered at 30–500 Hz. To correct for the influence of heart rhythm on the electrogram amplitudes, the voltage maps were obtained during either spontaneous or pacing-induced AF. Low-voltage points were defined by two cut-offs: ≤0.2 and 0.5 mV [4,12]. The mean bipolar voltage and the percentage of the low-voltage points were compared with the extent of LA delayed enhancement. All operators were blinded to the CMR results. 2.5. Statistical analysis Data are reported as the mean ± standard deviation or frequency and percentage. Reproducibility of the LA wall segmentation was evaluated by Dice's coefficient. Agreement between the observers in the quantification of LA LGE was assessed by Pearson's correlation test. Associations between the investigated variables and AF recurrence were explored using Student's t-test or chi-square test, as appropriate. The patients were divided into groups according to the extent of the LA LGE (Table 2). The groups were compared with regard to AF recurrence by one-way ANOVA test. Mutual relationship between the extent of LA LGE, volume, reservoir function and bipolar voltage was analyzed by Pearson's correlation test. P values b 0.05 were considered statistically significant. All analyses were conducted in R (http://www.R-project.org).
3. Results 3.1. Feasibility of LA LGE–CMR CMR was completed in 91 patients. Of the completed scans, 17 (19%) were corrupted mostly by motion blurring, contrast flow artifacts or diffuse noise (Fig. 3). The artifacts occurred more often in the patients who had AF at the time of the imaging, compared to those who were in sinus rhythm (10 of 30 [33%] vs. 7 of 60 [12%], P = 0.01). In total,
Fig. 1. The figure outlines the principles of the quantification of LA LGE based on the analysis of the image histogram. (A) A native LGE–CMR image of the LA. (B) The identical image with overlaid manual segmentation of the LA wall (blue) and blood pool (green). The areas of LGE (yellow) were identified by the supervised thresholding technique according to Oakes et al. [4]. (C) An image histogram of the LA wall voxels shown on the panel B. The abnormally enhanced voxels are separated from the rest of the LA tissue by arbitrary thresholds. (D) The final thresholded image was reconstructed in 3D using volume surface rendering. The red arrow indicates a corresponding region of abnormal delayed enhancement.
M. Sramko et al. / International Journal of Cardiology 179 (2015) 351–357
353
Fig. 2. The figure demonstrates the patient-specific selection of an arbitrary threshold at 3, 4 or 5 SDs above the mean signal intensity of the reference tissue. (Patient #1) The thresholds at 4 and 5 SDs underestimate the extent of the LGE. A threshold at 3 SDs seems to be the most appropriate. (Patient #2) The extent of LGE seems slightly overestimated at 3 SDs and largely underestimated at 5 SDs. In this patient, the examiner chose the 4 SD threshold. (Patient #3) This image has a low contrast-to-noise ratio. Therefore the thresholds below 5 SDs tend to overestimate the extent of the LGE. In this patient the examiner chose the 5 SD threshold.
an interpretable scan for analysis of LGE was available in 73 (77%) of the study participants (age, 59 ± 8 years; men, 71%; persistent/paroxysmal AF, 55/45%). The baseline characteristics of the patients with an interpretable LGE–CMR scan are presented in Table 1. Of note, besides having AF more often during the imaging, the excluded patients did not differ significantly from those with the interpretable CMR scan. 3.2. Reproducibility of image analysis At the repeated analysis, the primary examiner remained fairly consistent in the manual segmentation of the LA wall (Dice's coefficient, 0.65 ± 0.09) and also in the quantification of the LGE (absolute paired difference, 6 ± 5% points; R = 0.67; P = 0.046). In contrast, there was a poor inter-observer agreement between the three examiners, both in the segmentation of the LA wall (average Dice's coefficient, 0.42 ± 0.09) and also in the quantification of the LA LGE (absolute paired difference, 11 ± 8% points; R = 0.29, P = 0.4). The poor reproducibility of the image analysis resulted mainly from ambiguity of the LA images, especially at the level of the LA roof and the mitral annulus (Fig. 4). 3.3. LA LGE and outcome of catheter ablation AF recurred in 29 (40%) patients during a mean follow-up of 1.3 ± 0.8 years (range, 4 months to 3.5 years). All the AF recurrences were clinically relevant: 12 patients underwent reablation(s), another 14 underwent electric cardioversion(s) and the remaining 3 patients had
symptomatic AF paroxysms which lasted 49 min to N24 h on Holter monitoring. At the time of the AF recurrence, 13 (45%) of the patients were using antiarrhythmic drugs compared to 12 (27%) of those without arrhythmia recurrence (P = 0.1). In univariate analysis, the arrhythmia recurrence was associated with the persistent form AF, increased LA volume, lower LA bipolar voltage, the need for more extensive ablation and persistence of the arrhythmia at the end of the procedure (Table 1). Importantly, the extent of the LA LGE did not differ between the patients with and without AF recurrence (16 ± 8 vs. 18 ± 10%, P = 0.4), irrespective of whether they had paroxysmal AF (17 ± 9 vs. 20 ± 10%, P = 0.4) or persistent AF (15 ± 8 vs. 17 ± 10%, P = 0.5) and irrespective of the method for quantification of the LGE. Table 2 presents the rates of AF recurrence according to the stage of the LGE; a Kaplan–Meier analysis of the arrhythmia-free survival is shown in Fig. 5.
3.4. Extent of LA LGE and LA structural remodeling A high density voltage map along with an interpretable LGE–CMR scan was available in 56 individuals. We found no correlation between the extent of the LA LGE and LA voltages, volume or reservoir function, irrespective of the method for quantification of the LGE. On the other hand, lower LA voltage correlated with increased volume and worse reservoir function (Fig. 6); and the LA voltage negatively correlated with reservoir function (R = −0.51, P b 0.001). There was also a poor visual correlation between the spatial distribution of the LGE and the
354
M. Sramko et al. / International Journal of Cardiology 179 (2015) 351–357
Fig. 3. The study flowchart indicating feasibility of the LA LGE–CMR.
low-voltage areas (whether using the 0.2 or the 0.5 mV cut-off). A representative sample of the visual comparisons is shown in Fig. 7.
4. Discussion 4.1. Major findings
Table 1 Comparison of the patients' characteristics according to recurrence of atrial fibrillation. Variable
Clinical characteristics Male Age (years) Body mass index (kg/m2) Lone AF Persistent AF AF persisting N12 months History of AF (years) CHA2DS2-VASc score Antiarrhythmic drugs at baseline Magnetic resonance imaging LA volume (mL) LA emptying fraction (%) Extent of LA LGE, method by Oakes et al. (%) [4] Extent of LA LGE, FWHM method (%) Extent of LA LGE, 2 SDs threshold (%) Extent of LA LGE, 3 SDs threshold (%) Extent of LA LGE, 4 SDs threshold (%) Radiofrequency ablation Mean LA bipolar voltage (mV) Proportion of low voltage LA points (%) Cumulative radiofrequency energy (W min) AF non-termination by ablation Echocardiography LA parasternal dimension (mm) Left ventricular ejection fraction (%) Tissue Doppler E/E′ Mitral regurgitation (Grades 1–4)
No. of AF recurrence n = 44
AF recurrence n = 29
P value
33 (75%) 59 ± 8 29 ± 4 14 (32%) 20 (45%) 1 (2%) 5±3 1.4 ± 1.2 15 (34%)
19 (66%) 60 ± 7 30 ± 5 6 (21%) 20 (69%) 5 (17%) 4±4 1.7 ± 1.1 14 (48%)
0.38 0.35 0.32 0.28 0.048a 0.033a 0.49 0.25 0.22
89 ± 32 30 ± 16 18 ± 10
106 ± 36 26 ± 18 16 ± 8
0.047a 0.25 0.38
21 46 29 17
21 45 29 17
0.63 0.52 0.55 0.93
± ± ± ±
1 2 3 3
± ± ± ±
2 3 5 5
1.1 ± 0.7 46 ± 25 1161 ± 520
0.8 ± 0.5 60 ± 22 1599 ± 654
0.047a 0.034a 0.004a
16 (36%)
19 (66%)
0.015a
43 ± 6 56 ± 7 7.6 ± 1.9 1.3 ± 0.5
44 ± 6 52 ± 4 7.9 ± 2.5 1.3 ± 0.5
0.71 0.69 0.35 0.97
AF = atrial fibrillation; LA = left atrial; LGE = late gadolinium enhancement. a Signifies P b 0.05.
First, the extent of preablation LA LGE did not predict AF recurrence during post-ablation follow-up, regardless of the used image processing method. On the other hand, other traditionally reported factors, such as the type of AF, increased LA volume and lower LA voltage were significantly associated with the AF recurrence. Second, validity of the LGE– CMR was undermined by the fact that the extent of the LGE did not correlate with other LA characteristics that are related to the LA structural remodeling. At last, in our non-selected patient population the LGE scans were frequently corrupted by unavoidable imaging artifacts. The poor image quality was particularly common in the patients with ongoing AF, in whom every third examination was uninterpretable. 4.2. Previously published data Recent studies conducted at the CARMA center in Utah found a strong association between preablation LA LGE and AF recurrence after ablation [4,5,13]. The extent of the LGE correlated with lower LA voltage [4] and decreased LA strain [14]. The authors also proposed adjusting the ablation technique in each individual patient based solely on the extent of LA LGE [3]. The CARMA group has recently conducted a multicenter trial DECAAF which included 260 patients from 15 centers in 6 countries [7]. The extent of preablation LA LGE independently predicted AF recurrence after ablation [4]. Unfortunately, the findings of the DECAAF trial may be difficult to extrapolate to a wide clinical practice, because all the images were analyzed only by the CARMA group [15]. Using a different approach to the image analysis, a group of authors from London also reported association between preablation LA LGE and clinical outcome after catheter ablation [6]. The normalized LA signal intensity correlated with LA volume and mean LA bipolar voltage. In addition, the authors demonstrated that a suboptimal LGE sequence could not reliably detect even postablation LA scarring [16]. This finding implies that only an optimal quality scan may be used for assessment of the LA LGE.
M. Sramko et al. / International Journal of Cardiology 179 (2015) 351–357
355
Fig. 4. Typical LA LGE images.
Our study is comparable to the study by Oakes et al. [4], in terms of size, clinical endpoints and the approach to quantification of LGE. Based on our clinical experience we chose slightly different parameters of the LGE sequence and a different contrast agent. Scanning of the LA started somewhat earlier after contrast injection, because acquisition of the whole LA image stack took often over 15 min (especially in the individuals with irregular fast heart rate). We believe, however, that these differences have not fundamentally influenced our results. In fact, recently we have demonstrated that the same method of LGE–CMR can reliably detect even discrete changes in serially assessed LGE in the left ventricle during the course of dilated cardiomyopathy [17]. In agreement with our experience, Oakes had to discard 25% of the scans because of poor image quality. Although the average extent of LA LGE did not differ between the two studies, in our patients the LGE did not predict AF recurrence. We can only speculate whether the opposing findings of our study could be explained by the use of a different ablation technique or merely by the inherent subjective variability in the
image processing. On the other hand, our results were consistent even after avoiding the inter-observer variability by employing automated techniques for quantification of LGE [8]. Finally, we must underline, that assessment of preablation LA fibrosis by LGE–CMR has not yet been properly validated by systematic histological studies in human or in an animal model. In fact, assessment of LA LGE may be challenging even for detecting a dense postablation scar in a rigorously controlled experiment [18]. Therefore, while the finding of LGE in the left ventricle corresponds well to a compact fibrotic scar [19], and areas of LGE may reflect a scar after a previous ablation [18], LGE on a preablation scan should not be interpreted as LA fibrosis without any precautions. 4.3. Limitations Our study was powered only to investigate predictors of AF recurrence after a single ablation procedure. Also, including a non-selected patient
Table 2 Recurrence of atrial fibrillation according to the stage of LA LGE. Stage of LA LGE (% cut-off)
Tertiles of the present study sample 1st tertile (b13%) 2nd tertile (13–20%) 3rd tertile (N20%) Staging according to Oakes et al. [4] Minimal enhancement (b15%) Moderate enhancement (15–35%) Extensive enhancement (N35%)
AF recurrence
ANOVA F-statistics
P value
11/24 (46%) 9/24 (38%) 9/25 (36%)
0.28
0.8
14/34 (41%) 14/34 (41%) 1/5 (20%)
0.19
0.8
0.41
0.94
Staging according to the DECAAF trial [7] Stage I (b10%) 7/18 (39%) Stage II (10–19%) 10/24 (42%) Stage III (20–29%) 10/24 (42%) Stage IV (N30%) 2/7 (30%)
AF = atrial fibrillation; LA = left atrial; LGE = late gadolinium enhancement. The cut-off values in the brackets in the left column represent the relative extent of LA LGE.
Fig. 5. A Kaplan–Meier plot showing postablation AF-free survival depending on the preablation extent of LA LGE. The data are censored at 3 years of follow-up.
356
M. Sramko et al. / International Journal of Cardiology 179 (2015) 351–357
Fig. 6. The figure shows the relationship between LA LGE and other variables related to LA remodeling. Each point on the plots represents a mean value of the analyzed variable in one of the N patients with available quality data. R and P represent Pearson's correlation coefficients and the corresponding P-values. In this figure, the extent of LA LGE was quantified according to Oakes et al. [4]. However, similar results were achieved also by the other two thresholding techniques. EA map = electroanatomic map.
population resulted to a high heterogeneity of the sample. On the other hand, such study population better reflects clinical reality. Restoring of sinus rhythm by electrical cardioversion before the CMR exam could have provided a better image quality in some patients, but such indication for electric cardioversion is currently not supported by sound evidence. Although decreased endocardial bipolar voltage is a generally accepted surrogate of fibrotic myocardium, the relationship between decreased LA voltage and fibrosis has been investigated only in postablation LA scarring [18]. Finally, we must underline that the aim of the study was not to evaluate accuracy of the LGE–CMR for quantification of LA fibrosis but to evaluate the use of the method—as it currently stands—for routine clinical practice in a non-selected patient population.
that are related to LA fibrosis. Before applying the method to clinical practice, it needs to be validated by solid histopathological evidence and efforts should be taken to improve the imaging and image processing technique. Conflict of interest All the authors take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation. The authors report no relationships that could be construed as a conflict of interest. Funding sources
5. Conclusions This study found a limited clinical value of routine assessment of preablation LA LGE in patients undergoing ablation of AF. Moreover, the extent of LA LGE did not correlate with other LA characteristics
This study was supported by the Research Grant of the Ministry of Health of the Czech Republic — conceptual development of research organization (“Institute for Clinical and Experimental Medicine — IKEM, IN 00023001”).
M. Sramko et al. / International Journal of Cardiology 179 (2015) 351–357
357
Fig. 7. The figure shows three representative examples of a visual comparison between LA bipolar voltage maps and LGE images. The red-colored areas on the electroanatomic maps represent voltage b0.5 mV. The LGE images were reconstructed using maximum intensity projection rendering. The yellow color highlights abnormal delayed enhancement.
References [1] A.N. Ganesan, N.J. Shipp, A.G. Brooks, et al., Long-term outcomes of catheter ablation of atrial fibrillation: a systematic review and meta-analysis, J. Am. Heart Assoc. 2 (2013) e004549. [2] H. Kottkamp, Human atrial fibrillation substrate: towards a specific fibrotic atrial cardiomyopathy, Eur. Heart J. 34 (2013) 2731–2738. [3] G.R. Vergara, N.F. Marrouche, Tailored management of atrial fibrillation using a LGE– MRI based model: from the clinic to the electrophysiology laboratory, J. Cardiovasc. Electrophysiol. 22 (2011) 481–487. [4] R.S. Oakes, T.J. Badger, E.G. Kholmovski, et al., Detection and quantification of left atrial structural remodeling with delayed-enhancement magnetic resonance imaging in patients with atrial fibrillation, Circulation 119 (2009) 1758–1767. [5] N. Akoum, M. Daccarett, C. McGann, et al., Atrial fibrosis helps select the appropriate patient and strategy in catheter ablation of atrial fibrillation: a DE-MRI guided approach, J. Cardiovasc. Electrophysiol. 22 (2011) 16–22. [6] L.C. Malcolme-Lawes, C. Juli, R. Karim, et al., Automated analysis of atrial late gadolinium enhancement imaging that correlates with endocardial voltage and clinical outcomes: a 2-center study, Heart Rhythm. 10 (2013) 1184–1191. [7] N.F. Marrouche, D. Wilber, G. Hindricks, et al., Association of atrial tissue fibrosis identified by delayed enhancement MRI and atrial fibrillation catheter ablation: the DECAAF study, JAMA 311 (2014) 498–506. [8] R. Karim, R.J. Housden, M. Balasubramaniam, et al., Evaluation of current algorithms for segmentation of scar tissue from late gadolinium enhancement cardiovascular magnetic resonance of the left atrium: an open-access grand challenge, J. Cardiovasc. Magn. Reson. 15 (2014) 105. [9] C. Jahnke, J. Fischer, J.G. Mirelis, et al., Cardiovascular magnetic resonance imaging for accurate sizing of the left atrium: predictability of pulmonary vein isolation success in patients with atrial fibrillation, J. Magn. Reson. Imaging 33 (2011) 455–463. [10] V. Jarvinen, M. Kupari, P. Hekali, V.P. Poutanen, Assessment of left atrial volumes and phasic function using cine magnetic resonance imaging in normal subjects, Am. J. Cardiol. 73 (1994) 1135–1138.
[11] B. Aldhoon, D. Wichterle, P. Peichl, R. Cihak, J. Kautzner, Complications of catheter ablation for atrial fibrillation in a high-volume centre with the use of intracardiac echocardiography, Europace 15 (2013) 24–32. [12] M. Fiala, D. Wichterle, J. Chovancik, et al., Left atrial voltage during atrial fibrillation in paroxysmal and persistent atrial fibrillation patients, Pacing Clin. Electrophysiol. 33 (2010) 541–548. [13] C. Mahnkopf, T.J. Badger, N.S. Burgon, et al., Evaluation of the left atrial substrate in patients with lone atrial fibrillation using delayed-enhanced MRI: implications for disease progression and response to catheter ablation, Heart Rhythm. 7 (2010) 1475–1481. [14] S.S. Kuppahally, N. Akoum, N.S. Burgon, et al., Left atrial strain and strain rate in patients with paroxysmal and persistent atrial fibrillation: relationship to left atrial structural remodeling detected by delayed-enhancement MRI, Circ. Cardiovasc. Imaging 3 (2010) 231–239. [15] S. Vijayakumar, E. Kholmovski, M. Haslam, N. Burgon, N. Marrouche, Dependence of image quality of late gadolinium enhancement MRI of left atrium on number of patients imaged: results of multi-center trial DECAAF, J. Cardiovasc. Magn. Reson. 16 (Suppl. 1) (2014) 146. [16] R.J. Hunter, D.A. Jones, R. Boubertakh, et al., Diagnostic accuracy of cardiac magnetic resonance imaging in the detection and characterization of left atrial catheter ablation lesions: a multicenter experience, J. Cardiovasc. Electrophysiol. 24 (2013) 396–403. [17] M. Kubanek, M. Sramko, J. Maluskova, et al., Novel predictors of left ventricular reverse remodeling in individuals with recent-onset dilated cardiomyopathy, J. Am. Coll. Cardiol. 61 (2013) 54–63. [18] J.L. Harrison, H.K. Jensen, S.A. Peel, et al., Cardiac magnetic resonance and electroanatomical mapping of acute and chronic atrial ablation injury: a histological validation study, Eur. Heart J. 35 (2014) 1486–1495. [19] R.J. Kim, D.S. Fieno, T.B. Parrish, et al., Relationship of MRI delayed contrast enhancement to irreversible injury, infarct age, and contractile function, Circulation 100 (1999) 1992–2002.