Author + information
- Received May 26, 2018
- Revision received August 28, 2018
- Accepted August 29, 2018
- Published online January 21, 2019.
- Masateru Takigawa, MD, PhDa,b,∗∗ (, )
- Jatin Relan, MD, PhDa,c,∗,
- Ruairidh Martin, MDa,d,∗,
- Steven Kim, MD, PhDc,
- Takeshi Kitamura, MDa,
- Ghassen Cheniti, MDa,
- Konstantinos Vlachos, MD, PhDa,
- Xavier Pillois, PhDa,
- Antonio Frontera, MDa,
- Grégoire Massoullié, MDa,
- Nathaniel Thompson, MDa,
- Claire A. Martin, MD, PhDa,
- Felix Bourier, MD, PhDa,
- Anna Lam, MDa,
- Michael Wolf, MDa,
- Josselin Duchateau, MDa,
- Nicolas Klotz, MDa,
- Thomas Pambrun, MDa,
- Arnaud Denis, MDa,
- Nicolas Derval, MDa,
- Julie Magat, PhDa,
- Jérôme Naulin, MSca,
- Mathilde Merle, MSca,
- Florent Collot, MSca,
- Bruno Quesson, PhDa,
- Hubert Cochet, MD, PhDa,
- Mélèze Hocini, MDa,
- Michel Haïssaguerre, MDa,
- Frédéric Sacher, MD, PhDa and
- Pierre Jaïs, MDa
- aCHU Bordeaux, IHU Lyric, Université de Bordeaux, Bordeaux, France
- bHeart Rhythm Center, Tokyo Medical and Dental University, Tokyo, Japan
- cAbbott, St. Paul, Minnesota
- dInstitute of Genetic Medicine, Newcastle University, Newcastle-upon-Tyne, United Kingdom
- ↵∗Address for correspondence:
Dr. Masateru Takigawa, CHU Bordeaux, IHU Lyric, Université de Bordeaux, Avenue de Magellan, 33604 Bordeaux-Pessac, France.
Objectives This study sought to evaluate the relation between bipolar electrode spacing and far- and near-field electrograms.
Background The detailed effects of bipolar spacing on electrograms (EGMs) is not well described.
Methods With a HD-Grid catheter, EGMs from different bipole pairs could be created in each acquisition. This study analyzed the effect of bipolar spacing on EGMs in 7 infarcted sheep. A segment was defined as a 2-mm center-to-center bipole. In total, 4,768 segments (2,020 healthy, 1,542 scar, and 1,206 in border areas, as defined by magnetic resonance imaging [MRI]) were covered with an electrode pair of spacing of 2 mm (Bi-2), 4 mm (Bi-4), and 8 mm (Bi-8).
Results A total of 3,591 segments in Bi-2 were free from local abnormal ventricular activities (LAVAs); 1,630 segments were within the MRI-defined scar and/or border area. Among them, 172 (10.6%) segments in Bi-4 and 219 (13.4%) segments in Bi-8 showed LAVAs. In contrast, LAVAs were identified in 1,177 segments in Bi-2; 1,118 segments were within the MRI-defined scar and/or border area. Among them, LAVAs were missed in 161 (14.4%) segments in Bi-4 and in 409 (36.6%) segments in Bi-8. In segments with LAVAs, median far-field voltage increased from 0.09 mV (25th to 75th percentile: 0.06 to 0.14 mV) in Bi-2, to 0.16 mV (25th to 75th percentile: 0.10 to 0.24 mV) in Bi-4, and to 0.28 mV (25th to 75th percentile: 0.20 to 0.42 mV) in Bi-8 (p < 0.0001). Median near-field voltage increased from 0.14 mV (25th to 75th percentile: 0.08 to 0.25 mV) in Bi-2, to 0.21 mV (25th to 75th percentile: 0.12 to 0.35 mV) in Bi-4, and to 0.32 mV (25th to 75th percentile: 0.17 to 0.48 mV) in Bi-8 (p < 0.0001). The median near-/far-field voltage ratio decreased from 1.67 in Bi-2, to 1.43 in Bi-4, and 1.23 in Bi-8 (p < 0.0001).
Conclusions Closer spacing better discriminates surviving tissue from dead scar area. Although far-field voltage systematically increases with spacing, near-field voltages were more variable, depending on local surviving muscular bundles. Near-field EGMs are more easily observed with smaller spacing, largely due to the reduction of the far-field effect.
Bipolar mapping of ventricular scar has previously been validated with both focal ablation catheters and multipolar catheters (1–3). Multipolar mapping is increasingly being used for substrate delineation. It offers a faster and higher density map (4–6) that is expected to be superior to that obtained using ablation catheters. However, the impact of interelectrode spacing on bipole electrograms (EGMs) in the ventricle has not been systematically examined. This seems particularly important because ablation targets are determined by the signal recorded during the mapping phase, whether it is an activation or voltage map. The aims of the present study were to delineate the effect of increasing interelectrode spacing on local (far- and near-field) EGMs, with reference to scar regions defined by magnetic resonance imaging (MRI).
Experimental myocardial infarction
Experimental protocols were conducted in compliance with the Guiding Principles in the Use and Care of Animals published by the National Institutes of Health (NIH Publication No. 85-23, Revised 1996). In addition, the study was approved by the institutional animal use and care committee, and conformed to the guide for the care and use of laboratory animals. Seven female sheep (age: 4.1 years; weight: 54.9 ± 6.9 kg) were sedated with an intramuscular injection of ketamine hydrochloride (20 mg/kg), acepromazine (0.1 mg/kg), and buprenorphine (20 μg/kg). After an intravenous injection of propofol (2 mg/kg), the sheep were intubated, and anesthesia was maintained with 2% to 3% isoflurane. Sheep were ventilated with a respirator (CARESTATION/Carescape GE, Chicago, Illinois), using room air supplemented with oxygen. An intravenous catheter was placed in the internal jugular vein for infusion of drugs and fluids. Arterial blood gases were monitored periodically, and ventilator parameters were adjusted to maintain blood gases within physiological ranges. A sheep myocardial infarction model was created by an experienced invasive cardiologist using a selective ethanol injection (1 to 2 ml) in the distal one-third of the left anterior descending artery.
MRI myocardial scar segmentation
Late gadolinium-enhanced cardiac MRI was performed 2 to 3 months after the creation of myocardial infarction. Image processing was performed by 2 trained technicians using MUSIC software (EQUIPEX MUSIC, Liryc, Universite ́ de Bordeaux/INRIA, Sophia Antipolis, France). Segmentation was performed as described previously (7,8). Briefly, the cardiac chambers, ventricular epicardium, ascending aorta, and coronary sinus were segmented using semi-automatic methods (8). From left ventricular (LV) wall segmentation, adaptive histogram thresholding was applied to segment dense scar (threshold set at 50% maximum signal intensity) and the gray zone (from 35% to 50%) (9–11).
Electrophysiological study and mapping with HD-Grid catheter
Electrophysiological study with the HD-Grid catheter (Abbott, St. Paul, Minnesota) (Figure 1A) was performed 1 to 3 days after the MRI examination in surviving post-infarct sheep, with identical sedation, analgesia, intubation, and ventilation protocols for mapping of the scar using 3-dimensional electroanatomical mapping (EnSite Precision system, Research Version, Abbott St. Paul, Minnesota) (Figure 1A). A diagnostic catheter (Quadripolar Inquiry, Abbott) was placed in the right ventricle and coronary sinus. The LV was mapped with the research-use HD-Grid catheter (3-mm interelectrode spacing and 1-mm electrode size) via a retrograde aortic approach and/or transseptal approach. A steerable long sheath (Agilis, Abbott) was used if required. The internal projection setting was set at 7 mm with 10-mm interpolation. Field-scaling was applied for all maps. On the point acquisition, contact with the LV endocardial surface was confirmed by fluoroscopy and by the proximity indicator on the EnSite Precision system. The research version of this system also includes 2 features to prove the contact; one feature was to show the catheter placement in each beat, and the other was to allow the system to display whether the multipolar catheter electrodes were in contact with the tissue based on the impedance information (Online Figure 1). The beat was not acquired when the catheter was distorted because it was trapped in a particular space (e.g., the papillary muscles, valves, trabeculations) or when adjacent 2 splines touched each other, which caused noises.
For the registration, all segmentations of MRI imaging were exported as meshes and loaded into the EnSite system. The registration algorithm with the Ensite Fusion module allowed dynamic molding of the 3-dimensional electroanatomical geometry onto the MRI surface (12). After primary registration, the registered model was refined using a second set of fiducial points (e.g., left atrium, coronary sinus, aortic root, LV apex, and mitral annulus) judiciously placed in a stepwise fashion to further align both surfaces at sites of local mismatch (Figure 1B).
With a HD-Grid catheter, different bipolar pairs could be selected. To assess the effect of bipolar spacing, a specific configuration was created using electrodes 1 to 5 in each spline (Figure 1A). The 2-mm bipole pair (Bi-2; 2 mm center-to-center, 1 mm edge-to-edge) was created from electrodes 1 to 2, 2 to 3, 3 to 4, and 4 to 5 in each spline, which was defined as the smallest segment. Bi-2 was compared with the 4-mm bipole pair (Bi-4; 4 mm center-to-center, 3 mm edge-to-edge) created from electrodes 1 to 3 and 3 to 5 in each spline. We also used the 8-mm bipole pair (Bi-8; 8 mm center-to-center, 7 mm edge-to-edge) created from electrodes 1 to 5 (Figure 1A). In each beat acquisition, EGMs created by 2-, 4-, and 8-mm center-to-center spacing that covered exactly the same site were compared; 4 Bi-2 segments were compared with 2 Bi-4 and 1 Bi-8 recordings. Bipolar EGMs were filtered at 30 to 300 Hz. HD-32 (4 splines) was used in sheep 1 to 5, and HD-56 (7 splines) was prepared for sheep 6 and sheep 7. Importantly, HD-32 and HD-56 have the same electrode spacing (and the same electrode size; the only difference is the number of splines). Again, each portion between adjacent electrodes on the spline was defined as 1 segment, and the region covered between electrodes 1 and 5 was defined as 1 site (Figure 1B). We arbitrarily defined Bi-2 as the reference, and compared the following among 3 different bipolar spacings in each MRI-defined regions: 1) detection of local abnormal ventricular activities (LAVAs); 2) changes in EGMs at sites where far- and near-field potentials were separately measured; and 3) changes in single potentials in purely healthy and purely scar areas.
Our definition of LAVA has been described previously (5). LAVAs are defined as sharp, high-frequency ventricular potentials, possibly of low amplitude (but not always), that are distinct from the far-field ventricular EGMs. They sometimes display fractionation, double, or multiple components separated by very low-amplitude EGMs or an isoelectric interval, and are poorly coupled to the rest of the myocardium. In cases in which the discrimination of far- and near-field EGMs was difficult, programmed ventricular stimulation or local pacing with decremental output was performed. LAVA analysis was performed manually by 2 independent, blinded physicians.
Data are expressed as mean ± SD and/or median (25th to 75th percentile) for normally distributed and skewed data, respectively. For the comparison of the 2 groups, Steel-Dwass analysis was performed. A p value <0.05 was considered statistically significant.
EGMs were collected during sinus rhythm in 5 sheep and during right ventricular pacing in the remaining 2 because of frequent mechanical ventricular rhythm during sinus rhythm (Table 1). In total, 4,768 segments were analyzed by Bi-2, Bi-4, and Bi-8. A total of 2,020 segments were defined by MRI as a healthy area, 1,542 segments were in the MRI-defined scar area, and 1,206 segments were in the MRI-defined gray areas (border area).
Overlooking true scars with increased bipole spacing
A total of 3,591 segments were free from LAVAs with Bi-2 (Figure 2A). Yet, in 181 (5.0%) segments analyzed with Bi-4 and 228 (6.3%) with Bi-8, LAVAs were recorded as a consequence of the larger antenna of the bipoles. In the MRI-defined scar area (Figure 2D), single component EGMs without LAVAs were identified in 818 segments with Bi-2. However, in 113 (13.8%) segments analyzed with Bi-4 and 116 (14.2%) segments with Bi-8, LAVAs were recorded for the same reason. In the MRI-defined border areas (Figure 2C), 812 segments were free from LAVAs in Bi-2, of which 59 (7.3%) segments in Bi-4 and 103 (12.7%) segments in Bi-8 showed LAVAs. In the MRI-defined healthy areas (Figure 2A), 1,961 segments were free from LAVAs with Bi-2, of which 9 (0.5%) segments with Bi-4 and 9 (0.5%) segments with Bi-8 showed multiple components.
Overlooking LAVAs with increased bipole spacing
LAVAs were identified in 1,177 segments in Bi-2 (Figure 3A). Two hundred two (17.2%) of these segments in Bi-4 and 449 (38.1%) in Bi-8 missed these LAVAs. In the MRI-defined scar areas (Figure 3D), LAVAs were identified in 724 segments with Bi-2. Seventy-nine (10.9%) of these segments in Bi-4 and 236 (32.6%) segments in Bi-8 missed these LAVAs. In the MRI-defined border areas (Figure 3C), LAVAs were identified in 394 segments in Bi-2. One hundred (25.4%) of these segments in Bi-4 and 173 (43.9%) segments in Bi-8 missed these LAVAs. In the MRI-defined healthy areas (Figure 3A), multiple component EGMs were identified in 59 segments in Bi-2, mostly around the mitral valve, whereas 23 (39.0%) of these segments in Bi-4 and 40 (67.8%) segments in Bi-8 showed single component potentials.
Near-field/far-field voltage ratio decreases with spacing
In total, 1,177 segments in Bi-2, 1,156 segments in Bi-4, and 956 segments in Bi-8 showed both far-field and near-field EGMs. In general, the mean far-field voltage significantly increased as the spacing increased (Figure 4A[a]) (p < 0.0001). The tendency was similar in the MRI-defined healthy area, the MRI-defined border area, and the MRI-defined scar area (Figures 4B[a], 4C[a], and 4D[a]). In general, mean near-field voltage became significantly larger as the spacing increased (Figure 4A[b]) (p < 0.0001), but the extent of the increase was smaller than the far-field voltage, and the far-field voltage showed more proportional increases with the spacing. The increase did not reach significance in Bi-2 versus Bi-4 and Bi-4 versus Bi-8 in the MRI-defined border area (Figure 4C[a]).
In general, the mean near-field/far-field voltage ratio significantly decreased as the spacing increased (Figure 4A[c]) (p < 0.0001). The tendency was similar in the MRI-defined border and scar areas (Figure 4C[c] and 4D[d]).
Relation between spacing and voltage in purely healthy and purely scarred areas
We defined 1,856 segments as purely healthy where electrodes 1 to 5 in 1 spline were all simultaneously in an MRI-defined healthy area, without LAVAs. We defined 228 segments as purely scarred where electrodes 1 to 5 in 1 spline were all simultaneously in an MRI-defined scar area without LAVAs.
In the purely scarred area, the median voltage significantly (p < 0.0001) increased from 0.13 to 0.26 mV in Bi-4 and to 0.51 mV in Bi-8 (Figure 5A). Adding the voltage of 2 or 4 adjacent 2-mm bipoles (Bi-2) included in the corresponding Bi-4 or Bi-8 resulted in almost the same voltage (R2 = 0.99, R2 = 0.98, respectively) (Figures 5B and 5C).
In the purely healthy area, median signal voltage increased from 1.18 mV in Bi-2 to 1.81 mV in Bi-4, and to 2.58 mV in Bi-8 (Figure 6A). Adding the voltage of 2 or 4 adjacent 2-mm bipoles (Bi-2) included in the corresponding Bi-4 or Bi-8 was more weakly correlated to the voltage in Bi-4 (R2 = 0.82) and Bi-8 (R2 = 0.64) (Figures 6B and 6C).
In the present study, we examined EGMs in the same beat with the different spacing bipolar pairs and demonstrated the following findings. 1) Widely spaced bipoles might fail to discriminate the local tissue characteristics; increased spacing might result in missing LAVA areas and in mistakenly identifying scar or normal tissue areas as having LAVAs. 2) Although, generally, both far- and near- field voltage might increase as the spacing increases, a near-field increase might be less proportional. 3) The near-field/far-field voltage ratio increased as the spacing decreased due to the predominant reduction of far-field effect. 4) In pure scar areas, local voltages were additively increased as the spacing increased; however, local voltages in the healthy tissue were more heterogenous.
LAVA detection accuracy with bipolar spacing
In the present study, we demonstrated that LAVA potentials detected by the closest spacing (Bi-2) might be missed in Bi-4 by 17.2% and by 38.1% in Bi-8. This false-negative rate was generally higher in the MRI-defined border area than that in the dense scar area. LAVA timing was reported to be earlier in the border zone than that in the dense scar area (13), and early LAVA in sinus rhythm might be associated with channels entering into the scar (14). In these border zones, far-field EGMs were larger due to adjacent healthy tissue. With widely spaced bipoles, some LAVAs detected with Bi-2 might be masked, and therefore, might be missed by large far-field EGMs (Figure 3E).
In contrast, when single potentials were detected with the smallest bipolar spacing (Bi-2), 5% with Bi-4 and 6% with Bi-8 had LAVAs with far- and near-field EGMs, particularly in the MRI-defined scar areas, where the incidence of LAVAs increased to 13.8% by Bi-4 and to 14.2% by Bi-8. This did not suggest that larger spacing bipoles detected LAVAs better, but rather that widely spaced bipoles indiscriminately detected LAVA EGMs across the entire bipole, which smaller bipole pairs were able to pinpoint more accurately (Figure 2E). Even if larger bipoles are not as accurate as smaller ones, they may still be useful, for example, in identifying an interesting EGM that is nearby but which is not right under the spline.
Voltage changes in far- and near-field potentials
Two previous studies demonstrated that bipolar voltages, in general, increase with spacing (13,14). However, neither study compared identical sites simultaneously, nor did they discriminate between near-field and far-field EGMs (15,16). Furthermore, the comparison between different catheters might have affected the results because of the different electrode sizes and shapes (16). In the present study, by comparing EGMs with a different spacing at exactly the same location (site) and beat, the exact effect of bipolar spacing was clearly described. We demonstrated that both far- and near-field voltages generally increased with spacing, and that this rule was generally consistent in the MRI-defined scar and border areas. However, the voltage increased more in far-field EGMs with longer spacing; thus, the near-field/far-field voltage ratio decreased, which resulted in worse discrimination between near-field and far-field EGMs.
Correlation between bipolar spacing and voltage
In purely scarred areas, the additive voltage of Bi-2 was equal to the voltages of Bi-4 and Bi-8 with a good correlation. The (electrode center-to-center) spacing might be additively associated with far-field voltages, and the correlation coefficient was very high. In contrast, in purely healthy areas, although there was a correlation between the additive voltage of the smallest bipolar spacing (Bi-2) and the other spacing, the correlation coefficient was not as directly proportional as in the purely scarred area. This discrepancy might be because of the fact that single components in the purely scarred area might represent the pure far-field EGMs, whereas single component EGMs in the purely healthy areas might represent both far-field and near-field components, which would not simply be summed over a longer bipole.
In segments where both far- and near-field EGMs were identified, far-field EGMs tended to increase with spacing. However, the extent was not as directly proportional as in the completely scarred areas. This might be due to the presence of some near-field component in the far-field signal. Another hypothesis was that in the border zone, the far-field effect could be produced both from the horizontal direction and the vertical direction (subendocardial surviving tissues), whereas in the complete scar zone, the far-field effect from the vertical direction might be less.
The change in the near-field voltages according to bipolar spacing was less proportional compared with the far-field EGMs, especially in the border zone. In some cases, the voltages were similar even when the spacing was increased (Figure 2E). If there was only a single piece of surviving muscle, covered by 1 segment with Bi-2, this signal voltage might not change dramatically in Bi-4 or Bi-8 when the other segments were all scar. However, when the surviving muscular bundle was large enough to extend beyond a single 2-mm segment, the near-field signal would increase with a longer bipolar spacing. In addition, the near-field voltage was more sensitive to the relationships among bipolar orientation, tissue orientation, and activation direction, which could affect the near-field EGMs more (17).
Mapping during ventricular tachycardia (VT) can be performed in only 30% to 40% of cases (18). Therefore, detection and characterization of myocardial scar during sinus rhythm may be a prerequisite for substrate-based scar modification in some patients. For this substrate-based scar modification, several techniques have been reported, such as scar de-channeling (19), late potential ablation (20), scar homogenizing (21), and LAVA elimination (5). In all of these techniques, residual abnormal muscle bundles in the scar that are potentially associated with the initiation or maintenance of VT are targeted. These tissues usually have small, sharp EGMs, which are occasionally masked by the far-field component (5). Detection and discrimination of these near-field EGMs from the far-field EGMs is mandatory for these techniques. We demonstrated that catheters with smaller bipolar spacing might achieve the precise discrimination of the local tissue structures. In addition, we elucidated the impact of bipolar spacing on both far- and near-field EGMs. Recently, industry has provided new technologies, including several types of multipolar mapping catheters with different electrode size and different interelectrode spacing (e.g., PentaRay, electrode size: 3.14 mm2, center-to-center interelectrode spacing: 3 mm; Biosense-Webster Inc., Diamond Bar, California; and Orion™, electrode size: 0.4 mm2, center-to-center interelectrode spacing: 2.5 mm; Boston Scientific, Marlborough, Massachusetts). Although larger electrode and larger inter-electrode spacing might increase the far-field effect, the impact on the EGMs in these different technologies should be examined. In addition, a specific scar threshold should be determined in each different catheter platform, although <0.5 and >1.5 mV have been, so far, generally used for the threshold of dense scar and healthy voltages, respectively.
First, LAVA analysis was performed manually. To minimize bias when identifying LAVAs, analysis was performed by 2 independent, blinded physicians with adjudication by a third physician when there was disagreement. Second, contact force is known to affect recorded EGMs, but no multipolar mapping catheters were capable of contact force measurement. Intracardiac echocardiography might be a better method to minimize this limitation. Third, although histology and pathology were not compared with the MRI-defined scar in the present study, the well-known cutoff of MRI-defined scar was used for the analysis (9–11). Fourth, because EGMs during sinus rhythm and right ventricular pacing were mixed for the analysis, local voltages might have been affected. Finally, this study did not include epicardial mapping, which might have resulted in different EGM characteristics due to the impact of epicardial fat.
Closer spacing is superior for identifying surviving tissue in the scar. Far-field voltage increases additively with spacing, whereas near-field voltages are more sensitive to the tissue heterogeneity. Near-field EGMs were more easily identified with smaller spacing, particularly due to the reduction of the far-field effect in scar.
COMPETENCY IN MEDICAL KNOWLEDGE: High-resolution mapping with smaller spacing provides a more accurate characterization in the infarcted scar associated with the VT substrate. Better understanding of the substrate will provide the appropriate ablation strategy without unnecessary radiofrequency applications. Understanding of the effect of interelectrode spacing on both far- and near-field EGMS must be required in the era when the advancing technology in the industry might provide different types of catheters.
TRANSLATIONAL OUTLOOK 1: Although the high-resolution mapping catheter will certainly provide more precise characteristics of substrates of VT, the clinical impact should be examined in further studies.
TRANSLATIONAL OUTLOOK 2: Lack of contact force in mapping catheters may limit the advantage of high-resolution mapping system. A larger bipole, but with contact force versus a smaller bipole without contact force, should be examined.
↵∗ Drs. Takigawa, Relan, and Martin contributed equally to this work and are joint first authors.
This study was supported by Equipex MUSIC ANR-11-EQPX-0030, IHU LIRYC ANR-10-IAHU-04, and a research grant from Abbott. Drs. Haïssaguerre, Hocini, Jaïs, and Sacher have received lecture fees from Biosense Webster and Abbott; and Drs. Denis, Derval, Jaïs, and Sacher have received speaking honoraria/consulting fees from Boston Scientific. Mrs. Jatin and Mrs. Kim are Employees of Abbott. 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
- left ventricle
- local abnormal ventricular activity
- magnetic resonance imaging
- ventricular tachycardia
- Received May 26, 2018.
- Revision received August 28, 2018.
- Accepted August 29, 2018.
- 2019 American College of Cardiology Foundation
- Callans D.J.,
- Ren J.F.,
- Michele J.,
- Marchlinski F.E.,
- Dillon S.M.
- Reddy V.Y.,
- Wrobleski D.,
- Houghtaling C.,
- Josephson M.E.,
- Ruskin J.N.
- Patel A.M.,
- d’Avila A.,
- Neuzil P.,
- et al.
- Jaïs P.,
- Maury P.,
- Khairy P.,
- et al.
- Komatsu Y.,
- Cochet H.,
- Jadidi A.,
- et al.
- Amado L.C.,
- Gerber B.L.,
- Gupta S.N.,
- et al.
- Olimulder M.A.,
- Kraaier K.,
- Galjee M.A.,
- et al.
- Roes S.D.,
- Borleffs C.J.,
- van der Geest R.J.,
- et al.
- Tung R.,
- Mathuria N.S.,
- Nagel R.,
- et al.
- Tung R.,
- Kim S.,
- Yagishita D.,
- Vaseghi M.,
- et al.
- Berte B.,
- Relan J.,
- Sacher F.,
- et al.
- Anter E.,
- Josephson M.E.
- Tung R.,
- Vaseghi M.,
- Frankel D.S.,
- et al.
- Berruezo A.,
- Fernandez-Armenta J.,
- Andreu D.,
- et al.
- Silberbauer J.,
- Oloriz T.,
- Maccabelli G.,
- et al.
- Di Biase L.,
- Burkhardt J.D.,
- Lakkireddy D.,
- et al.