| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2003 ARRS Executive Council Award 1 |
1 Department of Radiology, Massachusetts General Hospital, Harvard Medical
School, 100 Charles River Plaza, Ste. 400, Boston, MA 02114.
2 Department of Cardiology, Massachusetts General Hospital, 100 Charles River
Plaza, Ste. 400, Boston, MA 02114.
Received March 12, 2003;
accepted after revision September 10, 2003.
Address correspondence to S. Abbara
(sabbara{at}partners.org).
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Twenty-six consecutive patients clinically referred for evaluation of possible ARVD underwent cardiac MRI. Two independent observers reviewed the images retrospectively. Intramyocardial areas (n = 101) that had increased signal intensity relative to normal surrounding myocardium on T1-weighted conventional spin-echo images ("index areas") were identified. The index areas were interpreted for presence of fatty infiltration using two sets of images: The first set was obtained without fat suppression, and the second set was obtained with fat suppression. Agreement between reviewers and confidence of interpretation were determined and compared.
RESULTS. Interobserver agreement was measured using a 5-point scale: 1, definitely not fat; 2, probably not fat; 3, equivocal; 4, probably fat; and 5, definitely fat. The resulting kappa values were 0.35 for non–fat-suppressed images and 0.55 for fat-suppressed images. Interobserver kappa increased from 0.67 without fat suppression to 0.90 with fat suppression using a 3-point scale: 1, not fat; 2, equivocal; and 3, fat. Confidence in the diagnosis increased from 7.2 without fat suppression to 8.8 with fat suppression (p < 0.0001) on a 10-point scale ranging from 1, not confident, to 10, very confident.
CONCLUSION. The use of fat-suppressed in addition to non–fat-suppressed conventional T1-weighted spin-echo imaging increased interobserver agreement and confidence in diagnosis and evaluation of intramyocardial fatty infiltration in patients who were suspected to have ARVD.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Hallmarks of ARVD include fibrofatty infiltration of the right ventricular free wall, right ventricular wall motion abnormalities with or without dilatation, and focal aneurysms in the right ventricular outflow tract, apex, and infundibulum. [3–5]. Clinical presentation of ARVD usually consists of arrhythmias of right ventricular (RV) origin ranging from isolated premature ventricular beats to sustained ventricular tachycardia, ventricular fibrillation, and sudden cardiac death.
The diagnosis of ARVD is difficult and is currently made on the presence of major and minor criteria that include structural, functional, histologic, electrocardiographic, arrhythmic, and genetic factors [6]. The diagnosis of ARVD is based on the presence of two major criteria, one major plus two minor criteria, or four minor criteria. Minor criteria include a family history of premature sudden cardiac death (< 35 years) or suspected ARVD, ECG abnormalities in the right precordial leads (V1–V3), and mild global or segmental right ventricular wall motion abnormalities. Major criteria include family disease confirmed at necropsy or surgery, epsilon waves on ECG, severe segmental or global right ventricular dilatation, right ventricular aneurysms, and fibrofatty replacement of right ventricular myocardium.
Endomyocardial biopsy is unreliable for the diagnosis of ARVD because the patchy distribution of the fibrofatty change may lead to sampling error. The only truly diagnostic gold standard is gross pathology from transplant hearts or postmortem examinations.
The imaging techniques used to evaluate right ventricular abnormalities include echocardiography, CT, conventional angiography, radionuclide angiography, and MRI. Among these, MRI is the most versatile and widely accepted because it may be used to detect right ventricular wall motion abnormalities, thinning of the myocardium, right ventricular dilatation, and fatty infiltration of the right ventricular myocardium [3, 7–9].
Identification of intramyocardial fat on conventional high-resolution T1-weighted spinecho images can be challenging because the right ventricle is a thin structure and areas of affected myocardium can be quite small. In addition, proximity to the surface coil, truncation band artifacts, and various motion-related ghosting artifacts may cause high signal intensities to be projected onto the myocardium and mistaken for fat.
Fat suppression performed during conventional spin-echo cardiac MRI has been used to identify fatty infiltration [10]. Fat-suppressed cardiac-gated conventional spin-echo imaging adds approximately 10–13 min per acquisition, depending on the individual's heart rate and scanning parameters. However, data proving the utility of fat-suppressed MRI in enhancing the interpretation of studies for suspected ARVD have not, to our knowledge, been reported in the literature. The aim of this study was to determine the utility of fat-suppressed spin-echo imaging in improving agreement between reviewers and confidence in the diagnosis of intramyocardial fatty infiltration.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MRI Acquisition
All patients were studied using a Signa CVi 1.5-T clinical system (General
Electric Medical Systems, Milwaukee, WI). They were scanned in the prone
position using a cervical-thoracic-lumbar spine phased array coil. ECG-gated
high-resolution axial and sagittal spin-echo images were acquired with spatial
saturation bands applied superiorly, inferiorly, posteriorly, and over the
atria. A 12–14 cm2 field of view was used. Slice thickness
ranged from 5 to 7 mm, usually with a 1-mm gap. Matrix size was 256 frequency
encoding steps by 192–224 phase encoding steps. The TR was equal to 1
cardiac cycle length, the TE was set at 8 msec, and the number of excitations
was 4.
This sequence was repeated with identical spatial coordinates and time points in the cardiac cycle using identical scanning parameters, except chemical fat saturation replaced the spatial saturation bands.
Image Analysis
The MR images were analyzed retrospectively offline by two independent
observers on a remote Advantage workstation (General Electric Medical
Systems). Intramyocardial areas with increased signal intensity relative to
normal myocardium on conventional T1-weighted spin-echo images were identified
(up to four areas per patient) and labeled as "index areas" (total
n = 101 index areas). Magnification, adjustment of the window level
and contrast settings, and viewing one adjacent slice above and below the
indexed area were permitted. The reviewers assigned a score to each individual
index area initially based only on the conventional spin-echo images acquired
without fat suppression. The index areas were scored on a 5-point scale: 1,
not fat; 2, probably not fat; 3, equivocal; 4, probably fat; and 5, definitely
fat. A 3-point scale was also used: 1, not fat; 2, equivocal; and 3, fat. The
reviewers estimated their level of confidence on a 10-point scale ranging from
1, not confident, to 10, very confident. After the individual index areas were
scored, the reviewers were allowed to review the entire acquisition and were
asked to give an overall patient score on the 3-point scale.
They then scored the index areas again as previously described after viewing the non–fat-suppressed images and corresponding fat-suppressed images together. After the individual index areas were scored, they were allowed to review the entire acquisition, including the non–fat-suppressed and the fat-suppressed images, and were asked to give an overall patient score on the 3-point scale.
Additionally, the reviewers were asked to score the helpfulness of the fat-suppressed images on a 4-point scale: (0, not helpful, to 3, very helpful). Reviewer agreement and confidence in diagnosis before and after fat suppression were determined.
Statistical Analysis
Pearson's correlation coefficients were used to determine the agreement of
scores between the two reviewers. The agreement between the reviewers was
measured using weighted kappa values for the index areas on a 5-point scale
and a 3-point scale, and the agreement on overall interpretations was measured
on a 3-point scale. The change in reviewer confidence was analyzed with a
paired Student's t test.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Scoring of Index Areas
The scores for the 101 individual index areas per patient (mean, 3.88;
range, 3–4) obtained from only the non–fat-suppressed spin-echo
images and the scores obtained when reviewed together with the fat-suppressed
images are depicted in Tables 1
(5- and 3-point scales) and 2
(3-point scale). Figure 1
illustrates the overall distribution of scores on the 5-point scale before and
after the fat-suppressed images were reviewed.
|
|
|
The individual fat scores (5-point scale) from the two reviewers showed a correlation coefficient of r = 0.528 (p < 0.0001) when using only standard spin-echo imaging. After adding the fat-suppressed images, the correlation coefficient between the reviewers increased to r = 0.822 (p < 0.0001). The mean (and SEM) absolute difference in scores between reviewers decreased from 0.90 (0.096) without fat suppression to 0.44 (0.068) with fat suppression.
The kappa values for interobserver agreement on the individual scores were 0.35 without fat suppression and 0.55 with it on the 5-point scale. On the 3-point scale, the same scores were 0.67 and 0.90, respectively.
A typical example of a case of reviewer agreement is illustrated (Fig. 2A, 2B), as well as examples of cases about which the reviewers disagreed on the basis of the standard spin-echo images alone but agreed after including the fat-saturated images in the analysis (Figs. 3A, 3B and 4A, 4B).
|
|
|
|
|
|
Overall Case Interpretation
The interobserver kappa value for agreement on the overall impression of
presence or absence of intramyocardial fatty change (n = 24 patients)
using a 3-point scale is (mean ± SD, 95% confidence interval [CI])
kappa = 0.67 ± 0.018, 0.635–0.705 before fat suppression and
kappa = 0.915 ± 0.0106, 0.894–0.936 with use of fat-suppressed
imaging (Table 2).
There were nine cases of initial disagreement on the overall impression using the 3-point scale. After review of the fat-saturated images, agreement was reached in five of these cases. In two cases, agreement could not be reached. In the other two cases, one reviewer remained indeterminate. In one case, an initial agreement changed to disagreement after viewing of the fat-saturated images.
Confidence Levels
The confidence levels for the overall interpretation (10-point scale)
increased from 7.2 (SD, 1.76; SEM, 0.13) without fat suppression to 8.9 with
fat suppression (p < 0.0001)
(Table 3). The use of
fat-suppressed imaging was not thought to be helpful in 8.1% of patients (for
reviewer 1, 8.0%; for reviewer 2, 8.1%). It was mildly helpful in 20.1% (for
reviewer 1, 19.0%; for reviewer 2, 21.2%). It was moderately helpful in 23.6%
(for reviewer 1, 25.0%; for reviewer 2, 22.2%) and very helpful in 48.3% (for
reviewer 1, 48.0%; for reviewer 2, 48.5%).
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The clinical diagnosis of ARVD currently depends on the presence of major and minor criteria. Definitive identification of intramyocardial fatty change is important because it is one of the major criteria. This study confirms that in patients undergoing MRI evaluation for possible ARVD, the use of fat-suppressed sequences in addition to conventional spin-echo imaging without fat suppression improved interobserver agreement in the diagnosis of intramyocardial fatty infiltration. More important, the use of fat suppression reduced the indeterminate diagnoses in the overall interpretation of the study. The use of fat suppression also improved the confidence of interpretation of intramyocardial fatty infiltrate and was felt to be helpful in most cases.
We achieved fat suppression by using a chemically selective saturation prepulse. Before each excitation in the spin-echo sequence, a narrowband excitation pulse was applied selectively at the Larmor frequency for fat. This pulse is effective because fat-based protons precess more slowly than water-based protons by approximately 225 Hz at 1.5 T. Fat saturation was successfully performed in all 26 of our patients with a chemically selective prepulse.
A major limitation of this study is the absence of histologic confirmation of intramyocardial fatty infiltration. Studies of ARVD are complicated by the absence of a noninvasive diagnostic gold standard. The only generally accepted gold standard for fatty infiltration of the myocardium in ARVD is pathologic findings from either transplant surgeries or from postmortem examinations. Endomyocardial biopsy of the right ventricle poses risks because of its invasive nature. Furthermore, because of the patchy nature of RV fatty infiltration, the technique is not very sensitive. Subtle to moderate ARVD might be missed on biopsy because of sampling error [1, 13].
Cardiac MRI with fat suppression allows the identification of even subtle to mild degrees of intramyocardial fatty infiltration. The significance of such minor fatty changes with respect to the generation of ventricular arrhythmias is uncertain, however. It is unknown whether a threshold exists, below which mild intramyocardial fatty infiltration ceases to be a major criterion for ARVD. The question requires further study.
In conclusion, when patients underwent cardiac MRI for suspected ARVD, fat-suppressed imaging significantly improved interobserver agreement. It also allowed the reviewers to be more definitive and confident in their interpretations. We therefore recommend routinely including fat-suppressed sequences in cardiac MRI studies when ARVD is suspected.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  H. Tandri, E. Castillo, V. A. Ferrari, K. Nasir, D. Dalal, C. Bomma, H. Calkins, and D. A. Bluemke Magnetic Resonance Imaging of Arrhythmogenic Right Ventricular Dysplasia: Sensitivity, Specificity, and Observer Variability of Fat Detection Versus Functional Analysis of the Right Ventricle J. Am. Coll. Cardiol., December 5, 2006; 48(11): 2277 - 2284. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
