Author + information
- Received March 12, 2015
- Revision received April 20, 2015
- Accepted April 28, 2015
- Published online November 1, 2015.
- Claudia Stöllberger, MD∗∗ (, )
- Birgit Gerecke, MD†,
- Rolf Engberding, MD†,
- Bernhard Grabner, MD∗,
- Cosima Wandaller, MD∗,
- Josef Finsterer, MD, PhD∗,
- Matthias Gietzelt, PhD‡ and
- Andreas Balzereit, MD†
- ∗Krankenanstalt Rudolfstiftung, Vienna, Austria
- †Klinikum Wolfsburg, Wolfsburg, Germany
- ‡Heidelberg University Hospital, Institute of Medical Biometry and Informatics, Heidelberg, Germany
- ↵∗Reprint requests and correspondence:
Univ. Prof. Dr. Claudia Stöllberger, Krankenanstalt Rudolfstiftung, Steingasse 31/18, A-1030 Vienna, Austria.
Objectives The aim of the study was to assess interobserver agreement (IOA) between 3 observers from 2 laboratories.
Background IOA of left ventricular hypertrabeculation/noncompaction (LVHT) in adults has only been studied within single echocardiographic laboratories.
Methods Echocardiographic recordings with and without LVHT were selected and anonymized. The “not-LVHT” cases were matched for age and systolic function. Each observer reviewed the recordings, blinded to the initial diagnosis and the other observers’ results. Pre-defined criteria for LVHT were: 1) >3 prominent trabeculae at end-diastole, distinct from papillary muscles, false tendons, or aberrant bands; 2) a noncompacted part of a 2-layered myocardial structure formed by these trabeculations; 3) a ratio of >2:1 of noncompacted to compacted layer at end-systole; and 4) perfusion of the intertrabecular spaces from the ventricular cavity. IOA was estimated using the kappa measure of concordance.
Results Cine-loops of 100 patients (42 women, ages 16 to 92 years), 50 from each center, and 51 with LVHT as the initial diagnosis, were reviewed. The left ventricular end-diastolic diameter was 32 to 78 mm, and ejection fraction, 4% to 88%. The observers agreed about presence (n = 29) or absence (n = 36) of LVHT and disagreed in 35 cases. Agreement was higher among the 2 observers from the same laboratory (kappa 0.793 [95% confidence interval (CI): 0.672 to 0.915]) than from different laboratories (kappa 0.628 [95% CI: 0.472 to 0.784], kappa 0.669 [95% CI: 0.521 to 0.818]). The observers agreed with the initial report of LVHT-presence in 53% and of absence in 67%. By reviewing the discordant cases, consensus was achieved about LVHT presence (n = 8) or absence (n = 16); in 11 cases, the diagnosis remained questionable. Discordance was due to poor image quality, lack of views in different apical planes, aberrant bands and chordae tendineae, abnormally sized or inserting papillary muscles, and localized calcifications of the endocardium.
Conclusions IOA was substantial for diagnosing LVHT. However, even the application of pre-defined criteria yielded disagreement in 35% of cases; and after mutual review, there were still 11% questionable cases.
Left ventricular hypertrabeculation/noncompaction (LVHT), also termed “spongy myocardium” or “persisting sinusoids,” is a cardiac abnormality of unknown etiology, classified as “primary genetic cardiomyopathy” by the American Heart Association or as “unclassified cardiomyopathy” by the European Society of Cardiology (1,2). The first case of the isolated form of LVHT including the echocardiographic appearance was reported in 1984 (3). LVHT is echocardiographically characterized by an increased number of left ventricular trabeculae and a 2-layered myocardial structure, with an outer compacted and an inner noncompacted layer. Though some attempts have been undertaken to uniformly define LVHT, no agreement about the echocardiographic criteria for LVHT exists (3–5). There are only a few reports about interobserver agreement (IOA) in diagnosing LVHT (Table 1) (6–10). Previous studies have only been carried out with observers from 1 echocardiographic laboratory (6–8) or in a core laboratory (9). Only in 1 study of pediatric cases did the observers come from different echocardiographic laboratories (Table 1) (10). To our knowledge, there is no study about IOA between echocardiographers from different laboratories in the echocardiographic diagnosis of LVHT in adults. The aim of the present study in adults was to assess IOA between 3 echocardiographers from 2 laboratories when diagnosing LVHT according to criteria that had been developed previously in a pilot study (11).
Patients and controls
Two experienced echocardiographers (A.B. and C.W.) who were not involved in the reviewing process selected echocardiographic recordings from LVHT cases from the echocardiographic databases of 2 cardiology departments (Klinikum Wolfsburg and Krankenanstalt Rudolfstiftung). They were unaware of the clinical data of the patients. Additionally, they selected controls (“not-LVHT”) from the same echocardiographic databases on the basis of matching age and degree of systolic function and no mention of LVHT in the echocardiographic report. Excluded from the control group were patients with congenital heart disease, recent myocardial infarction, aortic regurgitation and aortic stenosis, and bedside and emergency echocardiographic investigations. Systematic family screening for cardiomyopathy was not carried out in either LVHT or control patients. Patients and controls were investigated by different echocardiographers between 2009 and 2013 by using the echocardiographic machines VIVID5, VIVID S6, and VIVID 9 (General Electric, Fairfield, Connecticut). The echocardiographic definition of LVHT was the same in both laboratories and had been developed in a pilot study (11).
In order to enable blinded review, names and birth data of the patients were removed from the cine-loops by 2 of the authors (A.B. and B.Gr.) who were not involved in the reviewing process. They stored the anonymized cine-loops on discs by randomly mixing the sequence of LVHT cases and controls. Additionally, they recorded the following demographic and clinical data from LVHT patients and controls: age at echocardiographic examination, sex, left ventricular ejection fraction, and left ventricular end-diastolic diameter.
Review of the cine-loops
Three observers with 20 to 29 years’ experience in diagnosing LVHT (R.E., B.Ge., and C.S.) independently reviewed the anonymized echocardiographic recordings. Two observers (R.E. and B.Ge.) came from the same echocardiographic laboratory. Each observer reviewed the anonymized cine-loops by him or herself at different times without the presence of any other observer and blinded to the initial diagnosis and the other observers’ results.
LVHT was diagnosed if all of the following criteria, developed in a previous pilot study, were fulfilled (11):
1. >3 prominent trabecular formations at the endocardial side of the left ventricular myocardium visible in end-diastole, distinct from papillary muscles, false tendons, or aberrant bands;
2. a noncompacted part of a 2-layered myocardial structure formed by these trabeculations, best visible at end-systole;
3. a ratio of >2:1 of the noncompacted to the compacted myocardial layer measured at end-systole; and
4. perfusion of the intertrabecular spaces from the ventricular cavity at end-diastole on color-Doppler echocardiography or contrast echocardiography.
According to a preset form, the 3 observers assessed whether any of the criteria were present or not. The diagnosis of LVHT was made first by an eyeball evaluation and then, in suspected cases, by off-line measurements of the compacted and noncompacted myocardial layers, documented on printouts.
Review of the discordant and questionable cases
The 3 observers met to review and discuss discordant cases. It was recorded whether they found an agreement or not and which factors might have been responsible for discordant assessments. For cases that remained questionable, the methods by which a definite diagnosis could be achieved were assessed and whether LVHT was confirmed or excluded.
All data were transferred to an Excel version 14.0.7153.5000 (Microsoft, Redmond, Washington) data file for statistical analysis. Agreement between the observers was estimated using the kappa measure of concordance. To calculate kappa, the Software R version 2.10.1 (R Foundation for Statistical Computing, Vienna, Austria) and the package “epibasix” version 1.1 were used. A kappa ranging from 0.21 to 0.40 was classified as “fair,” from 0.41 to 0.60 as “moderate,” from 0.61 to 0.80 as “substantial,” and >0.80 as “almost perfect.” Group comparisons for differences of mean values from noncategorical data were carried out by the Student t test. Categorical data were analyzed by the 2-sided Fisher exact test.
Patients and controls
Cine-loops of 100 patients (42 women, ages 16 to 92 years), 50 from each center, were selected. In 51 of these patients, LVHT had been diagnosed according to the echocardiographic report (Klinikum Wolfsburg n = 31, Krankenanstalt Rudolfstiftung n = 20) and in the 49 controls (not-LVHT), 1 or more of the following diagnoses had been established: heart failure (n = 15), syncope (n = 2), stroke/embolism (n = 4), dilated cardiomyopathy (n = 3), coronary artery disease (n = 16), hypertensive heart disease (n = 19), and hypertrophic cardiomyopathy (n = 1). The left ventricular end-diastolic diameter ranged from 32 to 78 mm, and the left ventricular ejection fraction from 4% to 88%. In none of the LVHT or control cases was a familial cardiomyopathy known.
Agreement between the observers
After reviewing the cine-loops, the 3 observers agreed in 65 cases about the presence (n = 29) or absence (n = 36) of LVHT (Figure 1). The remaining 35 cases were discordantly assessed by the 3 observers. Agreement was classified as “substantial” for all 3 observers and was slightly higher among the observers R.E. and B.Ge., who came from the same echocardiographic laboratory (kappa 0.793 [95% confidence interval (CI): 0.672 to 0.915]), than among R.E. and C.S. (kappa 0.628 [95% CI: 0.472 to 0.784]) and B.Ge. and C.S. (kappa 0.669 [95% CI: 0.521 to 0.818]), who came from different echocardiographic laboratories. By mutually reviewing the discrepant 35 cases, a consensus was achieved about the presence (n = 8) or absence (n = 16) of LVHT. In the remaining cases, no consensus could be achieved, and the diagnosis remained questionable (n = 11).
Agreement with the initial diagnosis of LVHT
In 51 of 100 patients, LVHT had been diagnosed initially. In only 27 of 51 patients with initially diagnosed LVHT was the diagnosis confirmed by all observers (Figure 1). In the remaining 24 patients with initially diagnosed LVHT, all observers excluded the diagnosis (n = 3), or the results were discordant (n = 21). By mutually reviewing these discordant cases, the diagnosis of LVHT was confirmed in 8 cases and excluded in 3 cases. In the remaining 10 cases, the diagnosis remained questionable.
Agreement with the initial diagnosis not-LVHT
In 49 of 100 patients, LVHT had not been diagnosed initially (not-LVHT). In only 33 of 49 not-LVHT cases, absence of LVHT was confirmed by all observers. In the remaining 16 not-LVHT patients, all observers diagnosed LVHT (n = 2), or the results were discordant (n = 14). By mutually reviewing these discordant cases, the diagnosis of LVHT was excluded (n = 13) or remained questionable (n = 1).
Reasons for discordance
By mutually reviewing the discordant cases, the following reasons for discordance were identified: 1) poor image quality, especially in visualizing the apical regions; insufficient imaging planes visualizing the left ventricle in the apical 4-chamber view; 2) problems in differentiation between parts of the posterior wall and apical trabeculations in end-systole in cases with small left ventricles and thickened myocardium (Figure 2); and 3) differentiation between aberrant bands, chordae tendineae; abnormally sized or inserting papillary muscles, and localized calcifications of the endocardium and trabeculations (Figure 3).
Diagnosis of questionable LVHT
In 11 cases, the diagnosis of LVHT remained questionable even after mutually reviewing the cine-loops. Two examples are given in Figures 2 and 3. In 10 of these patients, the initial diagnosis had been LVHT, and in 1 patient, not-LVHT. Patients with questionable LVHT had either a 2-layered structure with a ratio of noncompacted:compacted >2 but <4 trabeculations in 1 imaging plane (n = 4) or >3 trabeculations but a ratio of noncompacted:compacted layer <2 (n = 7). All 3 observers agreed that in these 11 questionable cases, follow-up echocardiographic investigations and additional investigations such as contrast studies, 3-dimensional echocardiography, and cardiac magnetic resonance are necessary for confirmation or exclusion of LVHT.
Diagnostic measures for the questionable cases
All of the 11 questionable cases eventually underwent additional imaging studies. The initial diagnoses of these 11 cases were LVHT in 10 cases and not-LVHT in 1 case. Repeated transthoracic echocardiography with contrast or transesophageal echocardiography confirmed LVHT in 7 cases and excluded LVHT in 1 case with the initial diagnosis of LVHT (Figure 2). In 3 further cases, cardiac magnetic resonance confirmed (n = 2) or excluded (n = 1) LVHT in 1 case with the initial diagnosis of not-LVHT (Figure 3).
Association of clinical parameters with diagnostic agreement
No association was detected between the IOA and patients’ age (area under the curve [AUC]: 0.59). There was a tendency that IOA was lower in patients with smaller left ventricles (AUC for left ventricular end-diastolic diameter 0.68) and better systolic function (AUC for left ventricular ejection fraction 0.69) than in patients with enlarged left ventricles and decreased systolic function.
This study shows that among echocardiographers who have experience for several years with the diagnosis of LVHT, IOA concerning the diagnosis of LVHT is substantial. However, the observers agreed with the initial report of presence of LVHT in only 53% and of the absence in only 67% of the cases. Even after mutually reviewing the discordant cases, in 11%, the diagnosis of LVHT remained questionable.
Our results are similar to previous studies that have found that the reproducibility of diagnosing LVHT by accepted criteria is poor (Table 1). In the study by Saleeb et al. (6), 2 observers from the same echocardiographic laboratory reviewed echocardiograms of 104 children with the initial diagnosis of LVHT and 100 matched controls, and identified only 67% of the LVHT cases correctly. In that study, generally, agreement between observers was better than agreement between either observer and the initial report. A further study from a registry among adults reported that a core laboratory consisting of 2 to 4 independent experts in echocardiography confirmed the diagnosis of LVHT in only 105 of 154 cases (68%) (9).
Improving IOA in the diagnosis of LVHT is not only an academic need. Overdiagnosis of LVHT may become clinically relevant and have an impact on a patient’s career, especially if he or she is an athlete (12). A recently published case report shows that even among experts, the measurements of the myocardial layers and thus establishing the diagnosis of LVHT varies considerably when examining the same patient (13). Overdiagnosis of LVHT may also have an impact on prevalence studies. It has been shown by a study in patients with heart failure that the prevalence of LVHT is dependent on the diagnostic criteria (8).
One of the limitations of our study was that the analysis was restricted to cine-loops. Thus, it was not possible to obtain additional views in the questionable cases. The echocardiographic diagnosis of LVHT requires imaging in several imaging planes. The number of trabeculations can be counted best in apical views in end-diastole, whereas the 2-layered myocardial structure can be visualized and measured best in short-axis views in end-systole, especially in the apical regions. Lack of adequate recording of these multiple planes makes the diagnosis of LVHT difficult. A further limitation is that assessment of intraobserver agreement was not part of the protocol because of the following reasons: 1) the initial diagnosis was obtained by several echocardiographers and not only by the authors of the study; and 2) when jointly reviewing the discordant cases, the observers were aware about their and the others’ diagnoses. A further limitation is that we did not consider for the analysis the anatomic location where the measurements of the noncompacted and compacted myocardial layers were carried out.
Practical implications of our study are that echocardiographers should be aware of the problems in the diagnosis of LVHT and the high interobserver variability. The lack of a reference method as a gold standard for the diagnosis of LVHT complicates the diagnostic process. Even the pathoanatomic diagnosis of LVHT is not always clear-cut (14,15). Uncertain cases should be presented to experienced investigators from other echocardiographic laboratories before establishing the diagnosis. But even joint review of questionable cases cannot solve all diagnostic problems. In these cases, repeat echocardiography using contrast or transesophageal echocardiography may solve the problem. Cardiac magnetic resonance or 3-dimensional echocardiography may be further options, but it has to be kept in mind that the diagnostic criteria for LVHT by cardiac magnetic resonance are not well anatomically controlled and that IOA for these technologies has only been assessed sporadically (16,17).
We conclude that it has to be acknowledged that the echocardiographic diagnosis of LVHT is not always clear-cut. There remain questionable cases in which further investigations are needed. Cooperation between physicians and experts in imaging techniques is warranted to overcome these problems and to develop uniformly accepted and easily applicable diagnostic criteria for LVHT.
COMPETENCY IN MEDICAL KNOWLEDGE: Among echocardiographers who have long-time experience of several years with the diagnosis of LVHT, IOA concerning the diagnosis of LVHT is substantial. However, the observers agreed with the initial report of the presence of LVHT in only 53% and of the absence in only 67% of the cases. Even after mutually reviewing the discordant cases, in 11%, the diagnosis of LVHT remained questionable.
TRANSLATIONAL OUTLOOK: It has to be acknowledged that the echocardiographic diagnosis of LVHT is not always clear cut. There remain questionable cases in which further investigations by different imaging methods such as magnetic resonance imaging or contrast echocardiography are needed, and sometimes no consensus can be achieved. Cooperation between experts in imaging techniques and from different laboratories is warranted to overcome these problems and to develop uniformly accepted and easily applicable diagnostic criteria for LVHT.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- area under the curve
- confidence interval
- interobserver agreement
- left ventricular hypertrabeculation/noncompaction
- Received March 12, 2015.
- Revision received April 20, 2015.
- Accepted April 28, 2015.
- American College of Cardiology Foundation
- Maron B.J.,
- Towbin J.A.,
- Thiene G.,
- et al.
- Elliott P.,
- Andersson B.,
- Arbustini E.,
- et al.
- Brescia S.T.,
- Rossano J.W.,
- Pignatelli R.,
- et al.
- Kohli S.K.,
- Pantazis A.A.,
- Shah J.S.,
- et al.
- Habib G.,
- Charron P.,
- Eicher J.C.,
- et al.,
- Working Groups 'Heart Failure and Cardiomyopathies' and 'Echocardiography' of the French Society of Cardiology
- Stöllberger C.,
- Finsterer J.
- Gati S.,
- Rajani R.,
- Carr-White G.S.,
- Chambers J.B.