Non-Invasive and Diagnostic Cardiology
MR Imaging of Arrhythmogenic Right
Ventricular Cardiomyopathy: Morphologic
Findings and Interobserver Reliability
David A. Bluemked
Richard D. Whitep,1
Elizabeth A. Krupinskib
Lawrence M. Boxte,1
Charles B. Higginsi,1
Victor A. Ferrarig
Dudley J. Pennelln,1
aSarver Heart Center and bDepartment of Radiology, University of Arizona, Tucson, Ariz., cDepartment of
Radiology, University of Pennsylvania, Philadelphia, Pa., dDepartment of Radiology, Johns Hopkins Hospital,
Baltimore, Md., eDepartment of Radiology, Beth Israel Medical Center, New York, N.Y., USA; fDepartment of
Cardiology, Istituto di Clinica Medica e Cardiologia, Florence, Italy; gDepartment of Medicine, University of
Pennsylvania, Philadelphia, Pa., USA; h3rd Department of Internal Medicine, Central Hospital, St. Pölten, Austria;
iDepartment of Radiology, University of California San Francisco, San Francisco, Calif., jDepartment of Radiology,
Mayo Clinic, Rochester, Minn., kDepartment of Radiology, University of Chicago, Chicago, Ill., USA; lDepartment of
Radiology, Children’s and Women’s Health Center of British Columbia, Vancouver, Canada; mDepartment of
Radiology, University of Uppsala, Uppsala, Sweden; nDepartment of Cardiovascular MR, Royal Brompton Hospital,
London, UK; oDepartment of Radiology, University of Minnesota, Minneapolis, Minn., pDepartment of Radiology,
Cleveland Clinic Foundation, Cleveland, Ohio, USA; qDepartment of Cardiology and Angiology, University of
Münster, Münster, Germany
Received: February 11, 2003
Accepted: February 14, 2003
David A. Bluemke, MD, PhD
Department of Radiology
Johns Hopkins Hospital, 600 N Wolfe St
Baltimore, MD 21287 (USA)
Tel. +1 410 955 4062, Fax +1 410 955 9799, E-Mail email@example.com
Fax +41 61 306 12 34
© 2003 S. Karger AG, Basel
Accessible online at:
Readers of the MR studies.
Magnetic resonance imaging W Diagnosis W
Cardiomyopathy W Right ventricle
Background: Magnetic resonance (MR) imaging is fre-
quently used to diagnose arrhythmogenic right ventricu-
lar cardiomyopathy/dysplasia (ARVC/D). However, the
reliability of various MR imaging features for diagnosing
ARVC/D is unknown. The purpose of this study was to
determine which morphologic MR imaging features
have the greatest interobserver reliability for diagnosing
ARVC/D. Methods: Forty-five sets of films of cardiac MR
images were sent to 8 radiologists and 5 cardiologists
with experience in this field. There were 7 cases of defi-
nite ARVC/D as defined by the Task Force criteria. Six
cases were controls. The remaining 32 cases had MR
imaging because of clinical suspicion of ARVC/D. Read-
ers evaluated the images for the presence of (a) right
ventricle (RV) enlargement, (b) RV abnormal morpholo-
gy, (c) left ventricle enlargement, (d) presence of high T1
signal (fat) in the myocardium, and (e) location of high T1
signal (fat) on a Likert scale with formatted responses.
Results: Readers indicated that the Task Force ARVC/D
cases had significantly more (¯2 = 119.93, d.f. = 10, p !
0.0001) RV chamber size enlargement (58%) than either
the suspected ARVC/D (12%) or no ARVC/D (14%) cases.
When readers reported the RV chamber size as enlarged
they were significantly more likely to report the case as
ARVC/D present (¯2 = 33.98, d.f. = 1, p ! 0.0001). When
Bluemke et al.
readers reported the morphology as abnormal they were
more likely to diagnose the case as ARVC/D present (¯2 =
78.4, d.f. = 1, p ! 0.0001), and the Task Force ARVC/D
(47%) cases received significantly more abnormal re-
ports than either suspected ARVC/D (20%) or non-ARVC/
D (15%) cases. There was no significant difference be-
tween patient groups in the reported presence of high
signal intensity (fat) in the RV (¯2 = 0.9, d.f. = 2, p 1 0.05).
Conclusions: Reviewers found that the size and shape of
abnormalities in the RV are key MR imaging discrimi-
nates of ARVD. Subsequent protocol development and
multicenter trials need to address these parameters.
Essential steps in improving accuracy and reducing vari-
ability include a standardized acquisition protocol and
standardized analysis with dynamic cine review of re-
gional RV function and quantification of RV and left ven-
Copyright © 2003 S. Karger AG, Basel
Arrhythmogenic right ventricular cardiomyopathy/
dysplasia (ARVC/D) is a condition in which the right ven-
tricular muscle undergoes replacement by fatty and/or
fibrous tissue that is interspersed among the remaining
muscular fibers [1, 2]. This forms a substrate for reentrant
ventricular arrhythmias. ARVC/D presents most fre-
quently in men between the ages of 15–40 with premature
ventricular beats and nonsustained or sustained ventricu-
lar tachycardia of left bundle branch block morphology
. Though rare, sudden cardiac death may be the first
manifestation of this disease. It must be included in the
differential diagnosis of right ventricular outflow tract
(RVOT) tachycardia, which is the most common cause of
ventricular tachycardia in young people without obvious
structural heart disease. It is important to establish the
correct diagnosis since RVOT tachycardia is generally not
a progressive condition, seldom a cause of arrhythmic
death and not known to be transmitted genetically . In
contrast, ARVC/D may be progressive and may cause ar-
rhythmic death. In addition, since ARVC/D may be
familial (30–50% of cases) , it is important to correctly
diagnose this disease in the proband in order to know if
familial screening should be undertaken.
The diagnosis of ARVC/D is based on the identifica-
tion of structural abnormalities, fatty replacement of
the right ventricular myocardium, electrocardiographic
changes, arrhythmias of right ventricular origin, and his-
tory of familial disease. Because of the difficulty in diag-
nosing ARVC/D with certainty, a task force was con-
vened that proposed diagnostic criteria based on these
abnormalities . Individual criteria were listed as major
or minor and combinations of these were judged to indi-
cate the presence of this disease. Included among the
major criteria were severe or marked structural abnor-
malities of the right ventricle (RV) and fibro-fatty re-
placement of the myocardium on endomyocardial biopsy.
The Task Force stated that experience with MR imaging
in the diagnosis of ARVC/D was limited and required fur-
MR imaging is often used in the diagnostic evaluation
of ARVC/D because it is noninvasive and readily depicts
morphologic abnormalities of the RV, including right ven-
tricular chamber size or aneurysm, as well as functional
abnormalities of the RV using MR cine techniques [7–
12]. MR cine images, however, lack standardization and
many MR centers do not have the capability to transmit
or display cine images for purposes of referral to tertiary
medical centers. Morphologic MR imaging using spin-
echo technique is widely available and offers excellent
soft-tissue contrast resolution. In particular, fat is de-
picted with high signal intensity on T1-weighted MR
images, whereas the myocardium has an intermediate sig-
nal. The presence of myocardial fat identified by MR
imaging has been reported to be strongly associated with
ARVC/D [5, 7–12]. The reliability and interobserver
agreement of these MR imaging features, however, has
never been evaluated in a multi-center trial. The purpose
of this study was to determine the reliability of various
morphologic MR imaging features for diagnosing ARVC/
D and to determine interobserver variability for these
Hardcopy films of 39 patients were obtained from the ARVC/D
registry at the University of Arizona Health Sciences Center. Only
‘static’ MR images were evaluated since cine ‘functional’ MR images
were not available from most of the referring hospitals. Twelve of the
MR examinations were performed at university medical centers, 21
at private hospitals and 6 at imaging centers. The MR studies were
performed in 17 states and in Canada between the years 1994 and
2000 (median 1998). In addition, hardcopy films of 6 cardiac MR
examinations that were not performed for suspected ARVC/D were
also randomly selected from the film library at the University of Ari-
zona Medical Center. These cases were designated as control cases
for the blinded readings. The indications for performing these MR
studies included anomalous coronary artery, cardiac arrest, aortic
dissection, hypertension, stroke, and pectus excavatum. The MR
scanners were 1.5 T General Electric scanners (34 imagers), a 1.5 T
Philips scanner (1), 1.5 T Siemens scanners (6), and 0.5 T General
Evaluating MR Exams of Arrhythmogenic
Right Ventricular Cardiomyopathy
Table 1. Questions and possible answers on the pre-formatted response sheet
1 Is the study diagnostic for ARVC/D?
2 Is there evidence of fat in the myocardium?
Definitely no fat
Probably no fat
Possibly no fat
3 Circle areas where fat is located.
4 RV chamber size?
5 RV configuration, morphology, shape (aneurysm, muscle bundles, scalloping, trabeculation)?
Definitely normal Probably normal
6 LV chamber size?
7 How is overall image quality?
8 Are motion artifacts present?
9 Are there other artifacts?
Electric scanners (4). All MR scanners had ECG gating capability.
None of the MR scanners were ‘dedicated’ cardiovascular MR scan-
Each set of 45 MR films was sent to 8 radiologists and 5 cardiolo-
gists with expertise in cardiovascular MR imaging. No clinical infor-
mation was provided, other than that the reader was to evaluate the
films for the MR imaging findings of ARVC/D. All readings were
conducted independently. A preformatted response sheet accompa-
nied each case. There were nine questions with Likert-scale format-
ted responses to choose from (table 1). The readers circled the appro-
priate response based on their interpretation of the images. Five mor-
phologic descriptors previously described to be associated with
ARVC/D were evaluated by the readers: (a) enlargement of the RV;
(b) morphologic abnormality of the RV (e.g., aneurysm, scalloping,
increased trabeculation); (c) enlargement of the left ventricle (LV);
(d) presence of high T1 signal (fat) in the myocardium, and (e) loca-
tion of high T1 signal (fat). An overall impression as to the presence
or absence of ARVC/D was recorded. Overall image quality and the
presence or absence of image artifacts was also recorded.
Adjudication of Patient Diagnoses
All patients referred for the evaluation of ARVC/D were evaluat-
ed according to ARVC/D Task Force guidelines by a single reviewer
(F.M.) [7, 13, 14]. Patients were then classified as either ‘suspected
ARVC/D’ or ‘definite’ ARVC/D patients. Suspected ARVC/D pa-
tients had been referred for possible ARVC/D but these patients did
not meet Task Force guidelines and ultimately were not diagnosed as
ARVC/D. Definite ARVC/D patients met Task Force guidelines for
the diagnosis. Finally, a ‘control’ group of patients who had under-
gone MR imaging for reasons other than ARVC/D was included. MR
imaging results were not used for adjudication to classify patients
into any of these categories.
Of the 45 patients evaluated, 6 (13%) were control cases, 32
(71%) were suspected ARVC/D cases and 7 (16%) were definite
ARVC/D cases. Since there were 13 readers evaluating the 45 cases
(585 total reads), 78 reads were of the control cases, 416 reads were of
the suspected ARVC/D cases and 91 reads were of the definite
ARVC/D cases. Readers were allowed to not score MR findings if
they felt they were unable to assess that specific finding (e.g., due to
image artifacts). Thus, in the Results section below, the total number
of reads in each category is less than above totals.
Bluemke et al.
Fig. 1. Percentage of readings that identified increased T1-signal
intensity (corresponding to fat signal) in the RV in the control group,
the suspected and the definite ARVC/D group. There is no signifi-
cant difference between the groups (p 1 0.05).
A z-test for proportions  was used to compare the proportions
of ARVC/D image findings present versus ARVC/D findings absent
for the control cases, suspected ARVC/D cases and definite ARVC/D
cases. ¯2 analysis  was used to compare the presence or absence of
morphologic features between the three groups of cases. p values less
than 0.05 were defined as statistically significant. Î analysis was per-
formed to measure the overall rate of diagnostic agreement between
the 13 readers on all 45 cases . All statistical analyses were carried
out using StatView version 5.0 software (SAS Institute Inc., Cary,
MR Imaging Findings
The presence of high signal (equal to fat signal intensi-
ty) on T1-weighted images in the RV has been reported to
be associated with ARVC/D [5, 8–13]. Readers reported
no significant difference between patient groups in the
reported presence of high signal intensity in the RV (fig. 1;
32/78 control cases, 159/397 suspected ARVC/D cases,
41/90 definite ARVC/D cases; ¯2 = 0.9, d.f. = 2, p 1 0.05).
The location of the reported high T1 signal (e.g., anterior
wall, right ventricular outflow tract, apex) was not related
to whether the case was reported as positive or negative
for ARVC/D (¯2 = 22.98, d.f. = 14, p 1 0.05). When a
reader detected high T1 signal in the RV, this was highly
associated with that reader’s final diagnosis of ARVC/D
present (188/217 reads, 87%) rather than ARVC/D ab-
sent (44/348 reads, 13%; ¯2 = 304, d.f. = 1, p ! 0.0001).
RV chamber enlargement and morphology were as-
sessed by the readers. Readers found both of these mor-
phologic features to be present more frequently in definite
ARVC/D cases than in control or suspected ARVC/D
groups (fig. 2, 3). RV enlargement was present in 11/78
(14%) control reads, 50/394 (13%) suspected and 52/89
(58%) definite ARVC/D reads (¯2 = 96.4, d.f. = 2, p !
0.0001). Abnormal right ventricular morphology was
present in 12/76 (16%) control reads, 84/395 (21%) sus-
pected and 42/89 (47%) definite ARVC/D reads (¯2 =
29.83, d.f. = 2, p ! 0.0001).
When a reader diagnosed right ventricular enlarge-
ment, this was associated with that reader’s final diagno-
sis of ARVC/D present (70/218 reads, 32%) rather than
ARVC/D absent (43/343 reads, 13%; ¯2 = 31.60, d.f. = 1,
p ! 0.0001). Similarly, abnormal right ventricular mor-
phology detected by the reader was also associated with
that reader’s final diagnosis of ARVC/D present (105/218
reads, 48%) rather than ARVC/D absent (33/342, 10%;
¯2= 105.1, d.f. = 1, p ! 0.0001).
Readers also assessed the size of the LV. Left ventricu-
lar size was judged by the readers to be similar in all
groups (fig. 4). Normal LV size was present in 63/76
(83%) control reads, 356/393 (90%) suspected and 80/87
(92%) definite ARVC/D reads (¯2 = 4.22, d.f. = 2, p 1
Reader Agreement: Final ARVC/D Diagnoses
Based on all of the morphologic MR findings (func-
tional cine images were not available), readers were asked
for a final classification of ARVC/D as present or absent
(fig. 5). For control cases, readers correctly classified 52/
78 (67%) of cases as ARVC/D absent versus 26/78 (33%)
as ARVC/D present (z = 2.86, p ! 0.01). For suspected
ARVC/D cases, readers classified 157/416 (38%) of cases
as ARVC/D absent versus 259/416 (62%) as ARVC/D
present (z = 4.75, p ! 0.01). For definite ARVC/D cases,
readers classified 41/91 (47%) of cases as ARVC/D absent
versus 50/91 (53%) as ARVC/D present (z = 0.57, p 1
0.05). Overall reader agreement in the final classification
of ARVC/D present or absent was very poor (Î = 0.017).
Reader agreement for individual MR imaging features
was evaluated (fig. 6–9). In these figures, the minimum
and maximum rates of agreement with the final clinical
diagnosis of each case are shown within each category
(control, suspected ARVC/D, definite ARVC/D). For ex-
ample, for high T1 signal in the myocardium, figure 6
shows that for the control group, 12/13 (92%) readers
Evaluating MR Exams of Arrhythmogenic
Right Ventricular Cardiomyopathy
Fig. 2. Percentage of readings that identified enlargement of the RV
in the control group, the suspected and the definite ARVC/D group.
The differences between the groups was statistically significant (p !
Fig. 3. Percentage of readings that identified abnormal morphology
of the RV in the control group, the suspected and the definite ARVC/
D group. The differences between the groups was statistically signifi-
cant (p ! 0.0001).
Fig. 4. Percentage of readings that identified normal left ventricular
(LV) size in the control group, the suspected and the definite ARVC/
D group. The differences between the groups was not statistically sig-
nificant (p 1 0.05).
Fig. 5. Percentage of readings that defined ARVC/D as absent, in the
control group, the suspected and the definite ARVC/D group.
agreed that fat was present in the RV myocardium in 1 of
the patients. In another patient in the control group, only
1/13 readers (8%) identified the presence of fat. Since the
maximum agreement rates between the control group and
the patient group are very high and quite similar for fat
signal in the myocardium (92% control group and sus-
pected ARVC/D groups, 85% definite ARVC/D group),
this suggests that (1) a fat signal was present in all the
groups and (2) the fat signal did not discriminate between
subgroups. The differences between the maximum and
Bluemke et al.
Fig. 6. Percentage reader agreement for identification of increased
T1-signal intensity (corresponding to fat signal) in theRV. The mini-
mum (o) and maximum (d) rates of agreement on a per-patient
basis are shown for the control group, the suspected and the definite
ARVC/D group. For example, for one patient in the control group,
12/13 (92%) readers identified increased T1-signal intensity in the
RV, while in another control patient, only 1/13 (8%) readers identi-
fied a fat signal.
Fig. 7. Percentage reader agreement for identification of enlarge-
ment of the RV. The minimum (o) and maximum (d) rates of
agreement on a per-patient basis are shown for the control group, the
suspected ARVC/D group and the definite ARVC/D group. For
example, for 1 patient in the control group, 5/13 (38%) readers iden-
tified enlargement of the RV, while in another control patient, 0/13
(0%) readers described right ventricular enlargement.
Fig. 8. Percentage reader agreement for abnormal morphology of the
RV. The minimum (o) and maximum (d) rates of agreement on a
per-patient basis are shown for the control group, the suspected and
the definite ARVC/D group. For example, for 1 patient in the control
group, 5/13 (38%) readers identified abnormal right ventricular mor-
phology, while in another control patient, 0/13 (0%) readers de-
scribed abnormal right ventricular morphology.
Fig. 9. Percentage reader agreement for identification of normal size
of the LV. The minimum (o) and maximum (d) rates of agreement
on a per-patient basis are shown for the control group, the suspected
and the definite ARVC/D group. For example, for 1 patient in the
control group, 13/13 (100%) readers identified normal LV size, while
in another control patient, 8/13 (62%) readers described a normal
Evaluating MR Exams of Arrhythmogenic
Right Ventricular Cardiomyopathy
minimum rates of agreement suggest that right ventricu-
lar chamber size and morphology show lower degrees of
variability between readers compared to high T1 signal
(fat) in the myocardium for suspected ARVC/D and defi-
nite ARVC/D patients.
MR Image Quality
There was no relationship between image quality (¯2 =
6.88, d.f. = 3, p 1 0.05), the presence or severity of motion
artifacts (¯2 = 2.087, d.f. = 3, p 1 0.05), or the presence and
severity of other artifacts (¯2 = 1.62, d.f. = 3, p 1 0.05) and
whether a case was reported as ARVC/D present or
absent. The image quality, motion and other artifact
scores are summarized in table 2.
MR imaging is frequently performed when the diagno-
sis of ARVC/D is considered. MR imaging readily evalu-
ates the morphology and size of the RV using spin-echo
images. Additionally, MR has the potential to identify fat
signal within the myocardium [5, 7–12]. While these find-
ings have been reported to be key MR features of patients
with ARVC/D, their reproducibility has previously not
been determined in a multi-center study. The results of
this study indicate that identification of a fat signal is less
reliable than identification of right ventricular enlarge-
ment or morphologic changes (fig. 1–3). Further, the iden-
tification of the fat signal by the readers and the final diag-
nosis of ARVC/D were highly associated (p ! 0.0001),
suggesting readers relied on this finding for their final
diagnosis of ARVC/D. We conclude that for MR imaging
performed on conventional MR scanners (without dedi-
cated cardiovascular MR software), identification of fat
signal using conventional MR scanners is not a reliable
predictor of ARVC/D, and is not reproducibly identified
by expert MR readers.
There are several reasons that may account for these
results. The RV has a complex geometry, is asymmetric
and highly trabeculated . In normal individuals, the
mean right ventricular free-wall thickness is only 2.7 B
0.4 mm, and only 1.9 B 1.1 mm at the anterior right ven-
tricular apex . Epicardial fat is usually present, espe-
cially in association with the right coronary artery and left
anterior descending artery. Tongues of epicardial fat may
extend into the myocardium in normal individuals [14,
17]. Pathologically, these individuals have been found to
have fatty infiltration but no fibrosis of the right ventricu-
lar myocardium. This is in contrast to patients with
Table 2. Percent rating responses for image quality, motion artifacts
and other artifacts
n = 7
n = 32
n = 6
7 7 9
All figures are percentages.
Mild = Artifact(s) do not interfere with interpretation; moder-
ate = artifact(s) interfere somewhat with diagnosis; severe = arti-
fact(s) interfere extensively with diagnosis, may render image unin-
ARVC/D, who usually do have fibrosis in addition to fat
in the RV [14, 17]. MR identification of a right ventricu-
lar fat signal requires high resolution images without sig-
nificant artifacts. Unfortunately, patients with ARVC/D
are likely to have frequent premature ventricular beats
that can cause motion artifacts and diminish image quali-
ty. Image quality for patients in this study was considered
excellent in less than 10% of cases (table 2). We note that
morphologic findings and overall assessment of right ven-
tricular size by MR imaging are less dependent on very
high resolution detail and correspondingly had better
interobserver agreement as well as better agreement with
the final patient diagnosis.
An additional reason for reader variability may be the
lack of a standardized imaging protocol for the MR exam-
ination of patients with suspected ARVC/D. Thus some
of the views obtained may not be optimal. In addition,
other aspects of the test were not standardized. These
included whether MR scanning was performed in the
prone or supine position, the use of a body coil or a sur-
face coil (i.e., phased-array thoracic coil) and the orienta-
tion of phase encoding and read-out gradients. Thus, in
contrast to our results, a single-center study showed great-
er than 80% agreement between three independent ob-
Bluemke et al.
Table 3. Proposed optimal acquisition and analysis protocol for cardiovascular MR imaging in ARVC/D
Pharmacologic control of arrhythmia (e.g., beta-blocker).
1.5 Tesla MR scanner.
Thoracic torso or cardiac coil (phased array receiver coil).
Sequences (field of view preferred 24-26cm, but less than 32 cm). Anterior surface coil may be used to reduce
‘wrap around’ artifact.
1. Ventricular anatomy, wall thinning, fatty infiltration.
a. ECG-gated spin echo, transaxial plane, to cover from above the pulmonary valve to the diaphragm with
suppression of blood pool signal (such as preparatory double inversion pulse) with either 3-mm multi-slice,
4 NEX, 5-mm gap, respiratory compensation or 5-mm single slices, 1 NEX, 3-mm gap with multiple
acquisitions during breath holds. Fast spin-echo technique recommended only for breath-hold images.
In-plane pixel resolution should be ^1.5 ! 1.5 mm. Motion/artifact control techniques such as the use of a
saturation band over the anterior chest wall is encouraged.
b. Same spin-echo sequence in the short axis plane from base to apex of both ventricles.
2. Ventricular function (regional and global): Breath-hold cine gradient-echo sequences (steady-state free
precession) sequences preferred such as TrueFISP, balanced fast-field echo, Fiesta) encompassing the
entire right and LVs in both transaxial and short axis planes (planes to correspond with the spin-echo
acquisitions above to allow direct comparison). Ten-millimeter gap between center of slices and optimal
thickness according to local practice, but not less than 5 mm.
3. Other planes and sequences optional; for example fat suppression double inversion recovery fast spin echo
Analysis to include:
1. Quantification of global RV and LV volumes, function and mass.
2. Examination for regional wall-motion abnormalities on dynamic cines.
3. Examination for RV anatomical abnormalities such as thinning, bulging, abnormal trabeculation.
4. Examination for possible fatty infiltration, excluding known sources of fat such as around the right coro-
nary artery and left anterior descending coronary artery, atrial lipomatous hypertrophy and obesity.
5. Direct comparison of possible fatty infiltration sites with regional wall motion in corresponding areas on
6. Examination for LV involvement (clinically apparent in up to 15% of cases).
servers  for MR imaging of ARVC/D. The authors of
this study suggest an MR imaging protocol for ARVC/D
(table 3) that includes imaging in at least 2 planes, insur-
ing adequate coverage of the RV for known areas of pre-
dominance of ARVC/D (RV inflow and outflow tracts
and apex). Other alternative protocols have been used [5,
7, 12] and optimal protocol selection will require addi-
MR images obtained and interpreted at a single center
also have the advantage of evaluating cine MR images
 that aid in the diagnosis of ARVC/D and improve the
reliability of MR imaging. For example, myocardial thin-
ning on static MR images together with a wall motion
abnormality in the same region of the RV on the cine loop
has a high specificity for ARVC/D . Cine images dis-
play right ventricular global function and regional wall
motion abnormalities. A major deficiency of many MR
scanners is the inability to distribute, e.g. on CD-ROM,
cine loops of the acquired images. This deficiency is being
addressed on newer MR scanners, and CD-ROM or vid-
eotape capability is more likely to be present on those MR
scanners that have dedicated software and hardware for
cardiovascular MR imaging. PACS technology is increas-
ingly available and may further aid in the distribution of
cine MR images. Thus, our results are only applicable to
the use of static (generally spin-echo) images for ARVC/D
diagnosis as performed at general MR imaging centers in
the United States and Canada. Since cine loops of the car-
diac cycle were not available in this study, we cannot fully
assess the accuracy of cine MR imaging for ARVC/D
Except for the control patients, all other patients evalu-
ated in this study had idiopathic ventricular tachycardia,
and some of these patients likely had right ventricular out-
flow tract (RVOT) tachycardia. This could confound our
results, because there are reports that patients with RVOT
tachycardia have abnormalities on MR imaging. White et
al.  reported right ventricular abnormalities on MR
Evaluating MR Exams of Arrhythmogenic
Right Ventricular Cardiomyopathy
imaging in 76% of 46 patients with RVOT tachycardia,
including fixed thinning, reduced wall thickening or re-
duced wall motion. The most frequent location of these
abnormalities was in the right ventricular anterior wall,
but there was often caudal extension into the subparietal
region in the anterior wall of the RV. Fatty replacement
was observed in 8 (25%) patients, including those with an
indeterminate diagnosis. Similar findings were reported
by others [19–22]. In contrast, Grimm et al.  reported
no cardiac MR imaging abnormalities in 14 patients with
RVOT tachycardia. This wide range of results may be
partly attributed to differences in patient selection, diag-
nostic criteria for RVOT tachycardia and definition of
abnormal hyperintense signals in MR images as well as
incomplete diagnostic evaluation, particularly by the lack
of angiographic confirmation of the abnormalities .
Therefore the specificity of hypokinesis, akinesis, or even
fatty infiltration by MR imaging to differentiate patients
with RVOT tachycardia from those with ARVC/D re-
mains unclear, particularly when these changes are lim-
ited to the body of the RV and are not present in the poste-
rior subtricuspid region.
In this study, we did not quantitate ventricular func-
tion or size. ARVC/D is readily diagnosed when advanced
structural abnormalities of the RV are present. The dis-
ease, however, is more difficult to diagnose with certainty
when only minimal structural changes of the RV are
present. Quantitative analysis of the RV is likely more
reliable than visual inspection of chamber diameter or
area and can be compared with published normal values
using cardiovascular MR imaging . Minor increases
in RV volumes in the absence of other findings should be
treated with caution as a possible early sign of ARVC/D
which, in the correct clinical context, should lead to peri-
odic reassessment. Normal volumes in the absence of oth-
er signs of ARVC/D is highly reassuring.
In conclusion, radiologists and cardiologists should
recognize that the diagnosis of ARVC/D is presently
based on the fulfillment of ARVC/D Task Force criteria.
In this regard, MR imaging is useful for delineating the
morphology of the RV. Using conventional MR scanners
that are most widely available in the community setting,
identification of right ventricular enlargement and abnor-
mal morphology are more reliable than identification of a
fat signal in the right ventricular myocardium. Dedicated
cardiovascular MR scanners are now available at special-
ized centers, and evaluation of interobserver variability
using this technology needs to be undertaken using a stan-
dardized imaging protocol and analysis method with
experienced MR readers.
1 Marcus FI, Fontaine GH, Guiraudon G, et al:
Right ventricular dysplasia: A report of 24
adult cases. Circulation 1982;65:384–398.
2 Marcus FI, Fontaine G: Arrhythmogenic right
ventricular dysplasia/cardiomyopathy. A Re-
view. PACE 1995;18:1298–1314.
3 Corrado D, Fontaine G, Marcus FI, McKenna
WJ, Nava A, Thiene G, Wichter T: Arrhyth-
mogenic right ventricular dysplasia/cardio-
myopathy: The need for an International Re-
gistry. Circulation 2000;101:e101–e106/J Car-
diovasc Electrophysiol 2000;11:827–832.
4 Marcus FI: Is arrhythmogenic right ventricular
dysplasia, Uhl’s anomaly and right ventricular
outflow tract tachycardia a spectrum of the
same disease? Cardiol Rev 1997;5:25–29.
5 Boxt LM: Radiology of the right ventricle. Ra-
diol Clin North Am 1999;37:379–400.
6 McKenna WJ, Thiene G, Nava A, Fontaliran
F, Blomstrom-Lundqvist C, Fontaine G, Ca-
merini F: Diagnosis of arrhythmogenic right
ventricular dysplasia/cardiomyopathy. On be-
half of the Task Force of the Working Group
Myocardial and Pericardial Disease of the Eu-
ropean Society of Cardiology and of the Scien-
tific Council on Cardiomyopathies of the Inter-
national Society and Federation of Cardiology,
supported by the Schoepfer Association. Br
Heart J 1994;71:215–218.
7 Blake LM, Scheinman MM, Higgins CB: MR
features of arrhythmogenic right ventricular
dysplasia. Am J Roentgenol 1994;162:809–
8 Midiri M, Finazzo M, Brancato M, Hoffmann
E, Indovina G, Maria MD, Lagalla R: Arrhyth-
mogenic right ventricular dysplasia: MR fea-
tures. Eur Radiol 1997;7:307–312.
9 Wichter T, Lentschig MG, Reimer P, et al:
Magnetic resonance imaging; in Nava A, Rossi
L, Thiene G (eds): Arrhythmogenic Right Ven-
tricular Cardiomyopathy/Dysplasia. Amster-
dam, Elsevier Science BV, 1997, pp 269–284.
10 Friedrich MG: Magnetic resonance imaging in
cardiomyopathies. J Cardiovasc Magn Reson
11 Ricci C, Longo R, Pagnan L, Dalla Palma L,
Pinamonti B, Camerini F, Bussani R, Silvestri
F: Magnetic resonance imaging in right ventric-
ular dysplasia. Am J Cardiol 1992;70:1589–
12 Kayser, HWM, van der Wall, EE , Sivanan-
than, MU, Plein, S, Bloomer TN, and de Roos,
A: Diagnosis of arrhythmogenic right ventricu-
lar dysplasia: A review. Radiographics 2002;
13 Richardson PJ, McKenna WJ, Bristow M, et al:
Report of the 1995 World Health Organiza-
tion/International Society and Federation of
Cardiology Task Force on the definition and
classification of cardiomyopathies. Circulation
14 Fontaine G, Fontaliran F, Hebert JL, Chemla
D, Zenati O, Lecarpentier Y, Frank R: Ar-
rhythmogenic right ventricular dysplasia.
Annu Rev Med 1999;50:17–35.
15 Kanji GK: 100 Statistical Tests. Thousand
Oaks, Sage Publications, 1993, pp 75–76.
16 Cohen J: A coefficient of agreement for nomi-
nal scales. Educ Psychol Meas 1960;20:37–46.
Bluemke et al.
17 Burke AP, Farb A, Tashko G, Virmani R: Arr-
hythmogenic right ventricular cardiomyopathy
and fatty replacement of the right ventricular
myocardium. Are they different diseases? Cir-
18 White RD, Trohman RG, Flamm SD, et al:
Right ventricular arrhythmia in the absence of
arrhythmogenic dysplasia: MR imaging of
myocardial abnormalities. Radiology 1998;
19 Globits S, Kreiner G, Frank H, Heinz G, Klaar
U, Frey B, Gossinger H: Significance of mor-
phological abnormalities detected by MRI in
patients undergoing successful ablation of right
ventricular outflow tract tachycardia. Circula-
20 Proclemer A, Basadonna PT, Slavich GA,
Miani D, Fresco C, Fioretti PM: Cardiac mag-
netic resonance imaging findings in patients
with right ventricular outflow tract premature
contractions. Eur Heart J 1997;18:2002–2010.
21 Markowitz SM, Litvak BL, Ramirez de Arella-
no EA, et al: Adenosine-sensitive ventricular
tachycardia. Right ventricular abnormalities
delineated by magnetic resonance imaging.
22 Carlson MD, White RD, Trohman RG, et al:
Right ventricular outflow tract ventricular
tachycardia: Detection of previously unrecog-
nized anatomic abnormalities using cine mag-
netic resonance imaging. J Am Coll Cardiol
23 Grimm W, List-Hellwig E, Hoffman J, et al:
Magnetic resonance imaging and signal-aver-
aged electrocardiography in patients with repe-
titive monomorphic ventricular tachycardia
and otherwise normal electrocardiogram.
24 Auffermann W, Wichter T, Breithardt G, Joa-
chimsen K, Peters P: Arrhythmogenic right
ventricular disease: MR imaging vs. angiogra-
phy. AJR 1993;161:1549–1555.
25 Lorenz CH, Walker ES, Morgan VL, Klein SS,
Graham TP: Normal human right and left ven-
tricular mass, systolic function and gender dif-
ferences by cine magnetic resonance imaging. J
Cardiovasc Magn Reson 1999;1:7–21.