Schaeffter, Bernet Kato, C. Aldo Rinaldi, Michael Cooklin, Reza Razavi, Mark D. O'Neill and
Aruna Arujuna, Rashed Karim, Dennis Caulfield, Benjamin Knowles, Kawal Rhode, Tobias
Irreversible Atrial Injury After Catheter Ablation: Evidence From Magnetic Resonance
Acute Pulmonary Vein Isolation Is Achieved by a Combination of Reversible and
Print ISSN: 1941-3149. Online ISSN: 1941-3084
Copyright © 2012 American Heart Association, Inc. All rights reserved.
Avenue, Dallas, TX 75231
is published by the American Heart Association, 7272 Greenville
Circulation: Arrhythmia and Electrophysiology
2012;5:691-700; originally published online May 31, 2012;
Circ Arrhythm Electrophysiol.
World Wide Web at:
The online version of this article, along with updated information and services, is located on the
is online at:
Circulation: Arrhythmia and Electrophysiology
Information about subscribing to
Information about reprints can be found online at:
Permissions and Rights Question and under Services. Further information about this process is available in the
permission is being requested is located, click Request Permissions in the middle column of the Web page
Clearance Center, not the Editorial Office. Once the online version of the published article for which
can be obtained via RightsLink, a service of the Copyright
Circulation: Arrhythmia and Electrophysiology
Requests for permissions to reproduce figures, tables, or portions of articles originally published
by guest on October 27, 2013 by guest on October 27, 2013 by guest on October 27, 2013 by guest on October 27, 2013 by guest on October 27, 2013 by guest on October 27, 2013 by guest on October 27, 2013 by guest on October 27, 2013 by guest on October 27, 2013 by guest on October 27, 2013 by guest on October 27, 2013 http://circep.ahajournals.org/http://circep.ahajournals.org/http://circep.ahajournals.org/http://circep.ahajournals.org/ http://circep.ahajournals.org/ http://circep.ahajournals.org/http://circep.ahajournals.org/ http://circep.ahajournals.org/http://circep.ahajournals.org/ http://circep.ahajournals.org/http://circep.ahajournals.org/Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from Downloaded from
gin,1 an observation that has led to the emergence of pul-
monary vein isolation (PVI) as an effective treatment for
AF. Typically, ablation is performed at the left atrial (LA)-
pulmonary vein (PV) junction,2,3 with the intention of causing
acute tissue necrosis to eliminate conduction between the LA
and PVs. Clinical recurrences of AF after catheter ablation are
common, and recovery of LA-PV conduction is ubiquitous
in patients with and without documented AF during follow-
up.4 Single-procedure success rates are modest, suggesting
aroxysmal atrial fibrillation (AF) is often triggered by
spontaneous ectopic beats of pulmonary venous ori-
that the factors which contribute to acute PVI are not well
Clinical Perspective on p 700
Delayed enhancement (DE) magnetic resonance imaging
(MRI) after the administration of gadolinium has been used
extensively to image ventricular scar after myocardial infarc-
tion, secondary to coronary occlusion.6 More recent work has
demonstrated the potential use of cardiac magnetic resonance
imaging (CMRI) for assessment of atrial fibrosis before abla-
tion and of atrial injury after ablation.7,8 Although gadolinium
© 2012 American Heart Association, Inc.
Circ Arrhythm Electrophysiol is available at http://circep.ahajournals.org DOI: 10.1161/CIRCEP.111.966523
Received August 8, 2011; accepted May 12, 2012.
From the Division of Imaging Sciences and Biomedical Engineering, King’s College London, London, United Kingdom (A.A., R.K., D.C., B.K.,
K.R., T.S., B.K., C.A.R., R.R., M.D.O., J.G.); and Department of Cardiology, Guy’s and St. Thomas’ NHS Foundation Trust, London, United Kingdom
(A.A., D.C., C.A.R., M.C., R.R., M.D.O., J.G.).
Some of the data in this manuscript were presented in oral and poster abstract form at the Heart Rhythm Society sessions in 2011.
Correspondence to Dr Jaswinder Gill, MD, FRCP, Division of Imaging Sciences and Biomedical Engineering, The Rayne Institute, Lambeth Wing,
St. Thomas’ Hospital, Lambeth Palace Rd, London SE1 7EH, United Kingdom. E-mail firstname.lastname@example.org
Background—Pulmonary vein reconnection after pulmonary vein isolation is common and is usually associated with
recurrences of atrial fibrillation. We used cardiac magnetic resonance imaging after radiofrequency ablation to investigate
the hypothesis that acute pulmonary vein isolation results from a combination of irreversible and reversible atrial injury.
Methods and Results—Delayed enhancement (DE; representing areas of acute tissue injury/necrosis) and T2-weighted
(representing tissue water content, including edema) cardiac magnetic resonance scans were performed before,
immediately after (acute), and later than 3 months (late) after pulmonary vein isolation in 25 patients with paroxysmal
atrial fibrillation undergoing wide-area circumferential ablation. Images were analyzed as pairs of pulmonary veins
to quantify the percentage of circumferential antral encirclement composed of DE, T2, and combined DE+T2 signal.
Fourteen of 25 patients were atrial fibrillation free at 11-month follow-up (interquartile range, 8–16 months). These
patients had higher DE (71±6.0%) and lower T2 signal (72±7.8%) encirclement on the acute scans compared with
recurrences (DE, 55±9.1%; T2, 85±6.3%; P<0.05). Patients maintaining sinus rhythm had a lesser decline in DE between
acute and chronic scans compared with recurrences (71±6.0% and 60±5.8% versus 55±9.1% and 34±7.3%, respectively).
The percentage of encirclement by a combination of DE+T2 was almost similar in both groups on the acute scans (atrial
fibrillation free, 89±5.4%; recurrences, 92±4.8%) but different on the chronic scans (60±5.7% versus 34±7.3%).
Conclusions—The higher T2 signal on acute scans and greater decline in DE on chronic imaging in patients with recurrences
suggest that they have more reversible tissue injury, providing a potential mechanism for pulmonary vein reconnection,
resulting in arrhythmia recurrence. (Circ Arrhythm Electrophysiol. 2012;5:691-700.)
Key Words: ablation ◼ atrial fibrillation ◼ magnetic resonance imaging ◼ pulmonary vein reconnection
◼ reversible tissue injury
Acute Pulmonary Vein Isolation Is Achieved by a
Combination of Reversible and Irreversible Atrial Injury
After Catheter Ablation
Evidence From Magnetic Resonance Imaging
Aruna Arujuna, MBChB, MRCP; Rashed Karim, PhD; Dennis Caulfield, MBBS, MRCP;
Benjamin Knowles, PhD; Kawal Rhode, PhD; Tobias Schaeffter, PhD; Bernet Kato, PhD;
C. Aldo Rinaldi, MD, FRCP; Michael Cooklin, MD, FRCP; Reza Razavi, MD, FRCP;
Mark D. O’Neill, MRCP, DPhil, FHRS; Jaswinder Gill, MD, FRCP
692 Circ Arrhythm Electrophysiol August 2012
diffuses into the intracellular space after the loss of cell
membrane integrity associated with acute tissue destruction,
it can also accumulate acutely in the increased extracellular
space created by myocardial edema, which may represent a
reversible form of cardiac injury and is, therefore, not specific
to necrotic tissue.9 An alternative method to visualize myocar-
dial edema uses the linear relationship between T2 relaxation
time and myocardial water content and may be a more sensi-
tive in vivo marker of myocardial edema than DE MRI.10
The aim of the study was to use DE+T2-weighted CMRI
to characterize the tissue effect of left atrial ablation and to
relate the pattern of acute atrial injury to clinical outcome. We
hypothesize that acute PVI is caused by a combination of irre-
versible tissue destruction and reversible tissue injury at the
Twenty-five patients (17 men; mean age, 55±11 years) with symp-
tomatic, drug-refractory paroxysmal AF undergoing their first PVI
completed the study. Twenty-nine patients agreed to the study, but
4 were excluded (3 because of claustrophobia with failure to com-
plete scan and 1 because of an ineffective respiratory navigator). All
scans used for the purposes of data analysis were deemed of adequate
quality for analysis by an experienced CMR operator. Therapeutic
anticoagulation with an international normalized ratio >2 for at least
4 weeks before the procedure was mandated. The study was approved
by Guy’s and St Thomas’ Hospitals Research Ethics Committee.
Acute procedural success was defined as PVI confirmed using a
circumferential mapping catheter. Clinical outcomes are reported at
6-month follow-up. Patients were followed up in a clinic to assess
symptoms. Twenty-four-hour Holter monitors were used at 6 months.
Every effort was made to obtain ECG recordings of symptomatic
recurrences. Recurrences were defined on the basis of (1) symptoms
with ECG evidence of the presence of AF/flutter/tachycardia; or (2)
the presence of symptomatic or asymptomatic episodes of atrial ar-
rhythmia lasting for >30 seconds on ambulatory cardiac monitoring.
All participants underwent MRI in a 1.5-Tesla Philips Achieva MR
system (Philips Healthcare, Best, the Netherlands), using either a
32-channel surface coil (Invivo, Orlando, FL) or a large 2-channel
T2-weighted images were acquired using a multislice turbo spin
echo acquisition technique, with a double inversion recovery pre-
pulse for black-blood imaging. Spatial presaturation with inversion
recovery fat suppression was applied. The echo time used was set at
120 ms, with linear profile ordering. This enabled the image resolution
to be set at 1.5×1.5 mm2, with a slice thickness of 5 mm. The num-
ber of slices was set to provide complete coverage of the left atrium
(20–25 slices). Diaphragmatic motion was tracked, and respiratory
motion correction was applied to minimize motion blurring and dif-
ferences in respiratory phase between slices during image acquisition.
To visualize DE, a 3-dimensional ECG-triggered, free-breathing
inversion recovery turbo field echo scan with respiratory navigator
motion correction was performed with a pixel resolution of 1.3×1.3×4
mm3, which was then reconstructed to 1.3×1.3×2 mm3. Data were
acquired at mid-diastole, with a 150-ms acquisition window and a
low-high k-space ordering, as well as spatial presaturation with in-
version recovery fat suppression. The inversion recovery delay time
was determined from a Look-Locker sequence and was set at a TI
intermediate between the optimal TIs to null myocardium and blood.
Previous work has validated this method for reproducible visualiza-
tion of the late enhancement signal from necrotic tissue.11 DE scans
were performed 20 minutes after contrast agent administration. The
number of slices was set for complete atrial coverage (30–40 slices).
To optimize visualization of the PVs, slice orientation was performed
in the 4-chamber view. Images obtained with this method appear to
reflect the PVs at their maximal size. Similar MR sequences were
used for images acquired (1) before ablation, (2) within 24 hours of
ablation, and (3) 3 to 6 months after ablation.
A 6F decapolar catheter was placed in the coronary sinus to provide a
reference for electroanatomic mapping and to enable LA pacing. Two
transseptal punctures were made, and access to the left atrium was
obtained using 8.5F nondeflectable long sheaths (St. Jude Medical
Inc, St. Paul, MN). After the first transseptal puncture, intravenous
heparin was administered to achieve an activated clotting time be-
tween 300 and 400 seconds. A 3-dimensional geometry of the left
atrium was created using either NavX (St. Jude Medical Inc) or
CARTO XP (Biosense Webster Inc, Diamond Bar, CA). A circular
mapping catheter (Inquiry Optima; St. Jude Medical Inc) was then
placed in each PV in turn, whereas the corresponding LA-PV an-
trum was targeted with wide-area circumferential ablation. Energy
was delivered through a NaviStar ThermoCool 3.5-mm irrigated tip
catheter (Biosense Webster Inc), with flow limited to 17 mL/min,
power limited to 30 W on the anterior wall and 25 W on the pos-
terior wall, and temperature limited to 50°C. Ablation lesions
were marked on the LA geometry when there was an 80% reduc-
tion in the local electrogram voltage or after 30 seconds of en-
ergy delivery. One tag was applied to the shell per 30 seconds of
radiofrequency (RF) energy delivery, and a standard tag size was
used throughout the study. If LA-PV conduction persisted despite
wide-area circumferential ablation, additional lesions were deliv-
ered along the original ablation line at sites of earliest activation on
the circular mapping catheter until entry block in all 4 veins was
confirmed by observing the elimination or dissociation of PV po-
tentials. Exit block was not routinely assessed. Neither adenosine
nor isoprenaline was routinely administered to test the integrity of
PVI or to search for non-PV triggers of AF.
Image Processing, Analysis, and Its Validation
Using CMRI, this study sought to quantify the extent of PV antral
encirclement as demonstrated by DE+T2-weighted CMRI, individu-
ally and combined. To achieve this, an automated 3-dimensional
method (Figure 1) for visualizing and quantifying myocardial injury
(Figure 2) after ablation was used, which has previously been de-
scribed in detail.11
All 3-dimensional MR reconstructions were analyzed indepen-
dently twice by 2 experienced readers, blinded to clinical outcome
and to the timing of the scan after catheter ablation. T2 and DE
signal circumferential quantification was performed by reconstruct-
ing all CMR scans into individual left atrial shells (Figure 3). PVs
were analyzed as ipsilateral pairs for each of the 25 patients at
3 time points, permitting analysis of 150 PV pairs. For each PV pair,
T2 and DE were quantified as occupying a percentage of the antral
circumference. The percentage of pulmonary vein encirclement by
delayed gadolinium enhancement (DE), high T2-weighted signal
(T2), and combination of delayed gadolinium enhancement and T2
(DE+T2) was determined independently by both readers, and a con-
sensus was reached. A high degree of interobserver agreement was
seen on a Bland Altman test, with a maximum observed difference of
10% seen. The mean±SD interobserver error for DE, T2, and DE+T2
was 1.5±2.5%, 1.5±3.5%, and 0.8±2.2%, respectively, which was ac-
ceptable for the purposes of data analysis.
Summaries for continuous variables are expressed as mean±CI.
Follow-up times are reported as the median and interquartile range.
Categorical variables were compared between recurrence and non-
recurrence groups using a χ2 test. The percentage of circumferential
encirclement by DE, T2, and DE+T2 groups was compared to test
for differences between group means. Statistical analyses were per-
formed using Stata (StataCorp 2009). A linear regression model with
Arujuna et al Reversible and Irreversible Atrial Injury 693
predictor (code 1 for nonrecurrence and 0 for recurrence) and out-
come T2, DE, T2 and DE, DE/(T2 and DE), respectively, was applied
and run in Stata. We used the vce (cluster subject) option in Stata12 to
allow for intersubject dependence (left PV [LPV] and right PV [RPV]
measurements from the same patient). Analyses for acute and chronic
PV findings on CMRI were performed separately. P<0.05 was con-
sidered statistically significant.
Patient and Procedural Data
The Table outlines the baseline study population demo graphics.
Successful PV isolation was achieved in all patients. Median
follow-up time was 11 (interquartile range, 8–16) months.
A 3-month blanking period was observed, during which
arrhythmia recurrences were treated with antiarrhythmic
drugs or direct current cardioversion. No repeat ablation was
performed within the blanking period. Clinical recurrence
of AF was documented in 11 (46%) patients, with a median
time to recurrence of 94 (interquartile range, 45–166) days.
Patients with recurrences had significantly larger LA size and
longer duration of AF. Seven of 11 patients with recurrences
underwent a redo procedure; 2 patients are awaiting a redo
procedure, and 2 declined further intervention. Procedural
complications include 2 femoral venous hematomas, which did
not require intervention. No stroke, tamponade, or esophageal
fistula occurred in this study. The absence of both turbulence
on MR angiography and luminal narrowing in comparison with
the prescans confirmed no PV stenosis on follow-up MRI scans.
The circumferential burden of DE+T2-weighted signal
detected before any ablation was low in comparison with
acute postablation imaging (Figure 2) and did not occupy
>5% of the PV circumference. The median time of image
acquisition in relation to the procedure was 3 (interquartile
range, 1–10) days. Preablation DE signal localized to the
mitral annulus, a common finding because of the fibroelastic
nature of cardiac tissue at this site. T2 signal was largely
observed around the atrial roof, and this is likely explained
by the imperfections arising from the MR sequence. In
black-blood sequence, residual bright blood signal is
observed in areas of slow through-plane flow (eg, in the
apex of the ventricles). This problem has been reported in
acute edema assessment in the ventricles after acute myo-
cardial infarction.13 Overall, the amount of T2 signal preab-
lation was very small.
All acute imagings were performed between 18 and 24 hours
after catheter ablation. Figure 2 demonstrates the typical
T2-weighted (Figure 2A) and DE (Figure 2B) appearances in
2 patients before and after catheter ablation. The left atrial
burden of DE+T2-weighted signal was significantly increased
after catheter ablation in comparison with baseline (Figure 2).
On the acute scans, DE signal was concentrated in the PV
Figure 1. A, Raw magnetic resonance (MR) scan image of the left atrium (LA) and pulmonary veins (PVs) showing areas of delayed
enhancement. B, Fusion of the MR-derived 3-dimensional LA shell into the delayed enhancement image. The red arrows indicate
the direction in which the maximum intensity projection is taken. C, Projection of the MR signal intensities onto the surface shell.
The surface shell color is set within a range going from green to yellow to red, corresponding with low- to high-signal intensity.
D, 3-dimensional color LA shell harvested from the delayed enhancement MR image.
694 Circ Arrhythm Electrophysiol August 2012
antral region, whereas T2 signal was more widely distributed
in the atrium, remote from sites of ablation.
Individual analysis of the circumferential extent of both sig-
nal types revealed that the T2-weighted signal occupied 100%
of the antral circumference in 5 of 50 PV pairs, whereas the
DE signal did not achieve complete encirclement of any vein
pair. There was no significant difference between the circum-
ferential extent of the DE signal around the LPVs (mean, 65%;
CI, 56.4–73.6%) and the RPVs (63%; 55.4–70.6%; P=0.67).
Similarly, although the circumferential extent of the T2 signal
was greater, there was no significant difference between LPVs
(75%; 66.6–83.4%) and RPVs (80%; 73.0–87.0%; P=0.31).
Combined analysis of DE+T2 signals, using reconstructed
shells codisplaying both signal types, revealed areas of T2
enhancement to overlap and interdigitate with those areas of
high DE signal intensity (Figure 3). Hence, the sum of
DE+T2 is ≤100%. For the LPVs, the circumferential extent
of the DE signal, T2 signal, and the combination of both
Figure 2. A, A series of T2 signal images of the left atrium (LA) and pulmonary veins in 2 patients, with arrows pointing toward regions of
hyperenhancement in column 2. Baseline images in the first column show no significant T2 enhancement (tissue edema) compared with the
acute postablation images in the second column. The late scans in the third column shows the T2 signal becoming almost similar to base-
line in the preablation scans in column 1. B, A series of delayed enhancement (DE) images of the left atrium and pulmonary veins in 2 patients
with arrows pointing toward regions of hyperenhancement in both columns 2 and 3. Baseline images in the first column show no significant
DE signal (tissue injury/necrosis) compared with acute postablation images in the second column. The late scans in the third column shows
that areas of DE signal become less diffuse and more defined with sharper borders in comparison with the acute scans. AO indicates aorta.
Arujuna et al Reversible and Irreversible Atrial Injury 695
Figure 3. A and B, A series of reconstructed 3-dimensional left atrial shells to visualize T2 and the delayed enhancement (DE) signal in
patients shown in Figure 2. The 3 columns represent the 3 time points: preprocedure scans (prescans; column 1), acute postprocedure
scans performed within 18 to 24 hours (column 2), and the late scans performed later than 3 months (column 3). Quantification of these
enhancements was performed as percentage encirclements of the left and right pulmonary vein antra. Row A depicts the raw intensi-
ties mapped on to the shells from the T2 and DE magnetic resonance scans. Row B shows the corresponding T2 and DE 3-dimensional
shells that have been thresholded semiautomatically. Red areas signify delayed enhancement, and blue areas signify T2 signal intensity.
In row C, the combined enhancement of T2 and DE is seen together. On the acute scans seen in column 2, gaps present within areas of
red (DE) are filled in by areas of blue (T2). In column 3, the blue (T2) and red (DE) signals resolve, with a greater effect seen for T2 versus
DE signal. C, The electroanatomical maps in relation to the corresponding acute and late postprocedure 3-dimensional left atrial shell for
the 2 patients.
696 Circ Arrhythm Electrophysiol August 2012
signal types was 65% (56.4–73.6%), 75% (66.6–83.4%), and
90% (86.1–94.9%), respectively. For the RPVs, the circumfer-
ential extent of the DE signal, T2 signal, and the combination of
both signal types was 63% (55.4–70.6%), 80% (73.0–87.0%),
and 92% (86.4–97.6%), respectively. Compared with DE alone,
the combined DE+T2 signal was significantly greater for both
LPV (P=0.009) and RPV (P=0.027). Complete antral encircle-
ment with combined DE+T2-weighted signals was seen in 17 of
50 (34%) PV pairs at the acute scan.
As seen on the chronic follow-up scans, the T2 signal had
largely resolved (Figure 4), whereas a decline in the extent
of the DE signal was seen. For the LPVs, the circumferential
extent of the DE signal decreased from 65% (56.4–73.6%) to
51% (42.8–59.2%; P=0.016); for the RPVs, the circumferen-
tial extent of DE decreased from 63% (55.4–70.6%) to 46%
(37.5–54.5%; P=0.002). Discontinuities in areas of the DE
signal could be seen.
Recurrences of AF: Relationship to MR Assessment
Both acute and late scan data were divided into 2 groups
according to the respective clinical outcome—those with and
without AF recurrences. One hundred pairs of PVs (50 acute,
50 late) analyzed previously were divided into 2 groups
according to the presence or absence of a clinical recurrence
of AF. Figure 5 summarizes the percentage of circumferential
encirclement of DE, T2-weighted signal, and the combination
of DE+T2 around the LPV and RPV pairs by clinical outcome
for both the acute and late scans.
Table. Patient Demographics Categorized Into No
Recurrences and Recurrences at 6-Month Clinical Follow-Up
Duration of AF, mo
58±10.7 49±12.4 55±10.8 0.46
LA size, cm
History of smoking
Thyroid disease (%)
Atrial flutter (%) 5 (20)2 (14) 3 (27)>0.10
AF indicates atrial fibrillation; LA, left atrium; LVEF, left ventricular ejection
Figure 4. This scatter boxplot shows a comparison of pre-, acute, and late T2, delayed enhancement (DE), and combined T2 and DE for
both left and right pulmonary vein (PV) antrum. Each individual scatter plot represents the raw data for that specific group. The dots within
each group have been dispersed horizontally to optimize visualization and clarity. The boxplots on the other hand represent median
(red line), 95% CIs (yellow box), and 1 SD (blue box). An overall higher enhancement is seen in all 6 groups on the acute scans compared
with the 6 groups on the chronic scans. The percentage of encirclement by T2 signal diminishes from >75% to ≈5% in keeping with
reversible injury. The percentage of encirclement by DE signal diminishes to a much lesser extent. Using a combination of DE+T2 signal,
the percentage of encirclement decreases from 90% at the acute scans to ≈50% at the follow-up scan. The categories on the X axis are
composed of 3 elements: The first corresponds to the measure on the MRI scan (T2 or DE), The second is the time of the scan (P indi-
cates pre; A, acute scan after ablation; L, late scan after ablation). The third represents left-sided veins (L) or right sided veins (R).
Arujuna et al Reversible and Irreversible Atrial Injury 697
On the acute scans, there was no difference in the combined
DE+T2 signal between both groups, with a mean percent-
age encirclement of 89% (83.6–94.4%) in the nonrecurrence
group and 92% (87.2–96.8%) in the recurrence group. When
the DE signal was analyzed, a significantly higher mean per-
centage encirclement was noted in the AF-free group (n=14;
28 pairs of PVs) compared with the recurrence group (mean,
71%; CI, 65.0–77.0% versus 55%; 45.9–64.1%, respectively;
P= 0.016). Conversely, the T2 signal was noticeably lower
in the AF-free group compared with the recurrence group
(mean, 72%; CI, 64.2–79.8% versus 85%; 78.7–91.3%,
respectively; P=0.038). With the combined areas of DE+T2
forming almost complete rings around the PVs, ratios of
DE to (DE+T2) were calculated (Figure 6). Patients with no
recurrences had a higher mean DE/(DE+T2) ratio compared
with the recurrence group (0.82±0.12 versus 0.58±0.20;
On the late scans, DE was the predominant signal type
seen and was significantly greater in the AF-free group
compared with the recurrence group (mean, 60%; CI, 54.3–
65.7% versus 34%; 26.7–41.3%, respectively; P<0.0001).
A comparison of the acute and late scan DE data in both
groups showed a lower regression of this signal type in the
AF-free group (from mean, 71%; CI, 65.0–77.0% to 60%;
54.3–65.7%; P=0.03) relative to the group with arrhyth-
mia recurrences (from mean, 55%; CI, 45.9–64.1% to 34%;
The findings of this article are as follows: (1) acute PV iso-
lation is not associated with complete circumferential injury
as determined by CMRI performed at a median of 20 hours
after ablation; (2) increased DE+T2-weighted signals are
both seen within 24 hours of left atrial catheter ablation; (3)
the T2-weighted signal has largely resolved by 3 months of
follow-up, supporting its use as a marker of acute, reversible
atrial injury; and (4) in patients with clinical recurrences, a
greater proportion of the acute circumferential antral injury is
accounted for by the T2-weighted signal than in arrhythmia-
Previous works evaluating the role of CMRI in LA assess-
ment after catheter ablation has focused on delayed enhance-
ment imaging delineating areas of scar before and after
ablation.7,8,14,15 However, MRI of acute, reversible atrial injury
after catheter ablation has only been recently reported.16,17
There is evidence from animal studies that tissue edema
causes right atrial wall thickening after linear ablation in the
right atrium.18 Left atrial edema most likely occurs during and
immediately after AF ablation, as evidenced by an increase in
atrial wall thickness, and resolves within 1 month.19 During
late-gadolinium MRI performed immediately after ablation,
both nonenhancing and hyperenhancing tissue types are seen,
the former being a poor predictor of scar visualized at 3-month
follow-up.17 This is likely to reflect ablated but not necessar-
ily necrotic tissue confirming previous work, including that
from our own laboratory, that DE MRI overestimates the acute
Figure 5. This scatter boxplot shows a comparison of percentage of pulmonary vein (PV) encirclement according to clinical outcome,
no recurrence (NR) versus recurrence (R), of atrial fibrillation (AF) accounted for by T2 signal, delayed enhancement (DE) signal, and com-
bined T2 and DE at 3 time points: preablation (pre), immediately after (acute), and follow-up scans (late). Each individual scatter plot
represents the raw data for that specific group. The dots within each group have been dispersed horizontally to optimize visualization and
clarity. The boxplots on the other hand represent median (red line), 95% CIs (yellow box), and 1 SD (blue box). The absolute decline in
DE is less for patients with no AF recurrence. CMR indicates cardiac magnetic resonance.
698 Circ Arrhythm Electrophysiol August 2012
extent of tissue injury after left atrial catheter intervention
by virtue of the accumulation of gadolinium in extravascular
water associated with acute inflammation. Although there is
a good correlation between endocardial voltage-defined scar
and T2-weighted signal immediately after ablation, there is
a poor correlation with the DE MRI–defined scar at 3-month
follow-up,20 further supporting the transient nature of at least
part of the ablation injury process.
T2 signal was found in the acute CMR scans, remote
from the ablation sites. Similar observations have previously
been described.20 This is most likely related to a cytokine
(interleukin-6)-mediated inflammatory response after RF
ablation.21 Another possible mechanism giving rise to this
observation may be related to sheer/rotational force of the
catheters against the atrial wall during catheter manipulation.
Acute PVI and Atrial Ablation Injury
The data presented in this article demonstrate a high circumfer-
ential extent of each T2 and DE signal within 24 hours of abla-
tion. Although this is consistent with a high degree of overlap
of the 2 imaging signal types, there are also some areas where
the T2 signal can be detected in the absence of the DE signal
and vice versa (Figure 3). By overlaying DE+T2-weighted
images on the same anatomical shell, we have demonstrated
that the circumferential extent of ablation injury is greater
when both signal types are summated, reaching ≈90% (Fig-
ure 4). Although 100% circumferential extent of combined T2
and DE signals was seen in only 17 of 50 PV pairs, it is well
known that acute PV isolation can be achieved using a seg-
mental, electrogram-guided approach, rather than a circumfer-
ential ablation approach, the former not necessarily resulting
in ablation of the entire PV circumference.22 This may explain
the finding that PV isolation can be readily achieved without
circumferential MR evidence of ablation injury.
Atrial Scar and Arrhythmia Recurrence
The MR data at the follow-up scan demonstrate near abolition
of the T2 signal, whereas the DE signal is reduced and
continues to occupy only 60% of the circumferential extent
of both pairs of PVs. This is in keeping with the finding that
chronic PV reconnection is ubiquitous after conventional
wide-area circumferential ablation and, indeed, was seen
in all AF recurrences in this article.4,23 In the present study,
a greater extent of circumferential DE signal at the 24-hour
scan was predictive of freedom from AF (Figures 4 and 5),
whereas the extent of T2 signal was greater in the arrhythmia
recurrence group. Although there is a lack of clarity of what
DE+T2 signals truly represent in the immediate aftermath of
a catheter ablation procedure, the presence of a DE signal
beyond 3-month follow-up is likely to represent permanent
The formation of scar after ablation and its representa-
tion by DE+T2 sequences are not completely understood.
However, there are some similarities with findings of a DE
signal observed in serial CMRI (acute and late) after acute
myocardial infarction, with regression of delayed enhance-
ment areas over time. From this small sample of 25 patients,
we can qualitatively say that DE regions generally become
more distinct and smaller with time but that T2 regions did
not predictably become regions of DE, although there is some
overlap. Recognizing this as a hiatus in our knowledge of
atrial characterization by MR, an animal study is underway
Figure 6. Mean delayed enhancement (DE)/(T2 and DE) ratios quantified on the acute scans for patients with no recurrences versus those
with recurrences. An overall higher DE/(T2 and DE) ratio is seen in patients free from atrial fibrillation.
Arujuna et al Reversible and Irreversible Atrial Injury 699
to investigate further the temporal course of atrial injury/scar
after catheter ablation.
The decline in circumferential extent of DE signal between
acute and follow-up scans was less in patients with no AF
recurrence than in those with AF recurrence. This supports our
hypothesis that the greater the contribution of T2 signal, repre-
senting reversible injury to acute PVI, the higher the likelihood
of PV reconnection after resolution of tissue edema.
Although preliminary work has demonstrated a qualitative
correlation between discontinuities in areas of high DE signal
and conduction gaps on electrophysiology study,8,25 PV recon-
nection also occurs in patients without clinical arrhythmia
recurrence,26 and therefore, caution must be exercised in rely-
ing on the use of MR-defined scar as a surrogate for electro-
Potential Clinical Significance
It has been previously shown that durable RF lesion formation
is dependent on parameters, including catheter tip electrode
size, power, catheter tip temperature, and contact force. The
presented data suggest that there is an element of reversible
myocardial injury during ablation. Ablation strategies and
techniques that favorably alter the necrosis/edema ratio, such
as alternative energy sources, contact pressure sensing, and
improved catheter stability, may minimize reversible myocar-
There are significant limitations to MRI of the left atrium after
catheter ablation, with no widely accepted standardization of
technique between laboratories.
Although there is evidence from animal studies that gado-
linium is predominantly a marker for tissue necrosis, by virtue
of its kinetics, it also accumulates in extracellular water that is
seen in acute inflammation. In addition, although T2 MRI can
preferentially represent myocardial edema, there is currently
no robust histological evidence corroborating this in the atria
after RF ablation.
Although the DE+T2 signals recorded acutely after abla-
tion almost certainly include some double counting of edema
and necrosis by both techniques, the near-complete resolu-
tion of T2 signal at follow-up indicates that, at the very least,
T2 predominantly represents some form of reversible atrial
The annotation of lesions on an electroanatomic map is
subjective and likely does not accurately reflect the site of
atrial injury, which may explain, in part, the unanticipated
MR finding of PV encirclement in only 36% of PV pairs. We
attempted to mitigate this by using a point-by-point technique,
with RF applications of 30 seconds and 1 tag per application.
Detection of asymptomatic recurrences of AF without the
use of continuous monitoring is impossible. Because of the
frequency of monitoring, it is likely that the incidence of
asymptomatic AF is underreported in the current study.
This is a small, hypothesis-generating study, and the use of
necrosis and edema imaging as a predictor of long-term clini-
cal outcome would require a larger study for validation.
Acute PV isolation is achieved by a combination of
reversible and irreversible circumferential tissue injury at the
PV-LA junction. The greater the ablation extent accounted
for by reversible injury, the higher is the incidence of AF
Sources of Funding
This work was supported, in part, by the Technology Strategy Board
under grant 17352, in part, by the Philips Healthcare, Best, The
Netherlands, and, in part, by the EPSRC under grant EP/D061474/1.
Dr Arujuna was supported by St. Jude Medical and Dr Caulfield by an
EU Heart Grant. We acknowledge support from the UK Department of
Health via the National Institute for Health Research Centre award to
Guy’s and St. Thomas’ NHS Foundation Trust in partnership with King’s
College London and King’s College Hospital NHS Foundation Trust.
1. Haïssaguerre M, Jaïs P, Shah DC, Takahashi A, Hocini M, Quiniou G,
Garrigue S, Le Mouroux A, Le Métayer P, Clémenty J. Spontaneous ini-
tiation of atrial fibrillation by ectopic beats originating in the pulmonary
veins. N Engl J Med. 1998;339:659–666.
2. Di Biase L, Elayi CS, Fahmy TS, Martin DO, Ching CK, Barrett C, Bai
R, Patel D, Khaykin Y, Hongo R, Hao S, Beheiry S, Pelargonio G, Dello
Russo A, Casella M, Santarelli P, Potenza D, Fanelli R, Massaro R, Wang
P, Al-Ahmad A, Arruda M, Themistoclakis S, Bonso A, Rossillo A, Ra-
viele A, Schweikert RA, Burkhardt DJ, Natale A. Atrial fibrillation abla-
tion strategies for paroxysmal patients: randomized comparison between
different techniques. Circ Arrhythm Electrophysiol. 2009;2:113–119.
3. Oral H, Chugh A, Good E, Sankaran S, Reich SS, Igic P, Elmouchi D,
Tschopp D, Crawford T, Dey S, Wimmer A, Lemola K, Jongnarangsin K,
Bogun F, Pelosi F Jr, Morady F. A tailored approach to catheter ablation
of paroxysmal atrial fibrillation. Circulation. 2006;113:1824–1831.
4. Rajappan K, Kistler PM, Earley MJ, Thomas G, Izquierdo M, Sporton
SC, Schilling RJ. Acute and chronic pulmonary vein reconnection after
atrial fibrillation ablation: a prospective characterization of anatomical
sites. Pacing Clin Electrophysiol. 2008;31:1598–1605.
5. Weerasooriya R, Khairy P, Litalien J, Macle L, Hocini M, Sacher
F, Lellouche N, Knecht S, Wright M, Nault I, Miyazaki S, Scavee C,
Clementy J, Haissaguerre M, Jais P. Catheter ablation for atrial fibrilla-
tion: are results maintained at 5 years of follow-up? J Am Coll Cardiol.
6. Saeed M, Wendland MF, Takehara Y, Higgins CB. Reversible and ir-
reversible injury in the reperfused myocardium: differentiation with con-
trast material-enhanced MR imaging. Radiology. 1990;175:633–637.
7. Peters DC, Wylie JV, Hauser TH, Kissinger KV, Botnar RM, Essebag
V, Josephson ME, Manning WJ. Detection of pulmonary vein and left
atrial scar after catheter ablation with three-dimensional navigator-gat-
ed delayed enhancement MR imaging: initial experience. Radiology.
8. Badger TJ, Daccarett M, Akoum NW, Adjei-Poku YA, Burgon NS,
Haslam TS, Kalvaitis S, Kuppahally S, Vergara G, McMullen L, Ander-
son PA, Kholmovski E, MacLeod RS, Marrouche NF. Evaluation of left
atrial lesions after initial and repeat atrial fibrillation ablation: lessons
learned from delayed-enhancement MRI in repeat ablation procedures.
Circ Arrhythm Electrophysiol. 2010;3:249–259.
9. Ibrahim T, Hackl T, Nekolla SG, Breuer M, Feldmair M, Schömig A,
Schwaiger M. Acute myocardial infarction: serial cardiac MR imaging
shows a decrease in delayed enhancement of the myocardium during the
1st week after reperfusion. Radiology. 2010;254:88–97.
10. Friedrich MG. Myocardial edema–a new clinical entity? Nat Rev Car-
11. Knowles BR, Caulfield D, Cooklin M, Rinaldi CA, Gill J, Bostock J,
Razavi R, Schaeffter T, Rhode KS. 3-D visualization of acute RF abla-
tion lesions using MRI for the simultaneous determination of the patterns
of necrosis and edema. IEEE Trans Biomed Eng. 2010;57:1467–1475.
12. Rogers W. Regression standard errors in clustered samples. Stata Tech
700 Circ Arrhythm Electrophysiol August 2012 Download full-text
13. Kellman P, Aletras AH, Mancini C, McVeigh ER, Arai AE. T2-pre-
pared SSFP improves diagnostic confidence in edema imaging in acute
myocardial infarction compared to turbo spin echo. Magn Reson Med.
14. McGann CJ, Kholmovski EG, Oakes RS, Blauer JJ, Daccarett M, Seger-
son N, Airey KJ, Akoum N, Fish E, Badger TJ, DiBella EV, Parker D,
MacLeod RS, Marrouche NF. New magnetic resonance imaging-based
method for defining the extent of left atrial wall injury after the ablation
of atrial fibrillation. J Am Coll Cardiol. 2008;52:1263–1271.
15. Badger TJ, Oakes RS, Daccarett M, Burgon NS, Akoum N, Fish EN,
Blauer JJ, Rao SN, Adjei-Poku Y, Kholmovski EG, Vijayakumar S, Di
Bella EV, MacLeod RS, Marrouche NF. Temporal left atrial lesion forma-
tion after ablation of atrial fibrillation. Heart Rhythm. 2009;6:161–168.
16. Knowles BR, Caulfield D, Cooklin M, Rinaldi CA, Gill J, Bostock J,
Razavi R, Schaeffter T, Rhode KS. 3-d visualization of acute rf ablation
lesions using mri for the simultaneous determination of the patterns of
necrosis and edema. IEEE Trans Biomed Eng. 2010;57:1467–1475.
17. McGann C, Kholmovski E, Blauer J, Vijayakumar S, Haslam T, Cates J,
DiBella E, Burgon N, Wilson B, Alexander A, Prastawa M, Daccarett M,
Vergara G, Akoum N, Parker D, MacLeod R, Marrouche N. Dark regions
of no-reflow on late gadolinium enhancement magnetic resonance imag-
ing result in scar formation after atrial fibrillation ablation. J Am Coll
18. Schwartzman D, Ren JF, Devine WA, Callans DJ. Cardiac swelling
associated with linear radiofrequency ablation in the atrium. J Interv
Card Electrophysiol. 2001;5:159–166.
19. Saeed M, Wendland MF, Masui T, Higgins CB. Reperfused myocardial
infarctions on T1- and susceptibility-enhanced MRI: evidence for loss
of compartmentalization of contrast media. Magn Reson Med. 1994;
20. Vergara GR, Marrouche NF. Tailored management of atrial fibrillation
using a LGE-MRI based model: from the clinic to the electrophysiology
laboratory. J Cardiovasc Electrophysiol. 2011;22:481–487.
21. Brueckmann M, Wolpert C, Bertsch T, Sueselbeck T, Liebetrau C, Kaden
JJ, Huhle G, Neumaier M, Borggrefe M, Haase KK. Markers of myocar-
dial damage, tissue healing, and inflammation after radiofrequency cath-
eter ablation of atrial tachyarrhythmias. J Cardiovasc Electrophysiol.
22. Hocini M, Sanders P, Jaïs P, Hsu LF, Takahashi Y, Rotter M, Clémenty J,
Haïssaguerre M. Techniques for curative treatment of atrial fibrillation.
J Cardiovasc Electrophysiol. 2004;15:1467–1471.
23. Nanthakumar K, Plumb VJ, Epstein AE, Veenhuyzen GD, Link D,
Kay GN. Resumption of electrical conduction in previously isolat-
ed pulmonary veins: rationale for a different strategy? Circulation.
24. Segerson NM, Daccarett M, Badger TJ, Shabaan A, Akoum N, Fish EN,
Rao S, Burgon NS, Adjei-Poku Y, Kholmovski E, Vijayakumar S, Di-
Bella EV, MacLeod RS, Marrouche NF. Magnetic resonance imaging-
confirmed ablative debulking of the left atrial posterior wall and septum
for treatment of persistent atrial fibrillation: rationale and initial experi-
ence. J Cardiovasc Electrophysiol. 2010;21:126–132.
25. Reddy VY, Schmidt EJ, Holmvang G, Fung M. Arrhythmia recur-
rence after atrial fibrillation ablation: can magnetic resonance imag-
ing identify gaps in atrial ablation lines? J Cardiovasc Electrophysiol.
26. Cappato R, Negroni S, Pecora D, Bentivegna S, Lupo PP, Carolei A,
Esposito C, Furlanello F, De Ambroggi L. Prospective assessment of late
conduction recurrence across radiofrequency lesions producing electri-
cal disconnection at the pulmonary vein ostium in patients with atrial
fibrillation. Circulation. 2003;108:1599–1604.
Single ablative therapy for paroxysmal atrial fibrillation has moderate success, and many patients present with recur-
rent arrhythmia. We proposed that the structure of the radiofrequency lesion applied during ablation is important in
determining recurrences. The nature of the radiofrequency lesion was studied using magnetic resonance imaging with
gadolinium-enhanced imaging and high-signal T2-weighted imaging. Twenty-five patients underwent magnetic resonance
imaging scans for delayed enhancement (DE) and T2 at 3 time points: before ablation, within 24 hours, and 6 months after
ablation. Patients were divided into those with (n=11) and without (n=14) recurrent arrhythmia. Levels of DE+T2 were
low in preprocedural scans but rose dramatically immediately after the procedure. Acute DE was greater in patients with-
out recurrences compared with those with recurrences. Conversely, T2 levels were lower in patients without recurrences
and higher in those with recurrences. On the late scans, T2 reduced to baseline. DE, however, remained and was greater
in patients without recurrences. We, therefore, propose that acute radiofrequency ablation injury is composed of 2 types of
tissue damage. DE infers largely necrotic tissue injury, which lasts longer and causes persistent conduction block. T2 is a
transitory phenomenon coexisting with DE, causing acute conduction block. We propose that resolution of the T2 signal
is associated with recurrences of pulmonary vein connection and, therefore, arrhythmia recurrences. Modifications in our
ablative techniques to achieve more DE at the acute ablation would potentially be important in conferring a better ablation
outcome. These data potentially provide a mechanistic explanation as to why pulmonary veins reconnect after wide area