Publication
Article
Author(s):
We performed cineventriculography, unenhanced echocardiography, contrast-enhanced echocardiography, and magnetic resonance imaging to define the presence of regional left ventricular wall motion abnormalities. Interobserver agreement in the analysis of regional wall motion abnormality was highest for contrast-enhanced echocardiography, followed by cineventriculography and cardiac magnetic resonance imaging; it was lowest for unenhanced echocardiography. Contrast-enhanced echocardiography also showed the highest accuracy in the detection of panel-defined regional wall motion abnormalities.
The detection of regional wall motion abnormalities is an integral part of the evaluation of left ventricular function, particularly in patients with coronary artery disease who may have hypokinesia or akinesia. Regional wall motional abnormalities are currently assessed in clinical practice using echocardiography, cineventriculography, and cardiac magnetic resonance imaging (MRI). In a recently published study, we used 4 imaging modalities—cineventriculography, cardiac MRI, unenhanced echocardiology, and contrast-enhanced echocardiography—for the detection of left ventricular regional wall motion abnormalities.1 For each of the imaging techniques, the accuracy as well as the interobserver and intermethod agreement in the evaluation of regional wall motion abnormalities was determined.
Patients and methods
Eight centers participated in the study. Cineventriculography, unenhanced echocardiography, and contrast-enhanced echocardiography were performed in 100 patients. In a subgroup of 56 patients, additional cardiac MRI was performed. The patients were scheduled for cardiac imaging because they had suspected or confirmed cardiovascular disease. Centers were given detailed protocols for performance of imaging examinations and specific criteria for reading the imaging studies.
Echocardiographic images were obtained with apical 4-, 3-, and 2-chamber views, employing unenhanced tissue harmonic imaging and contrast-enhanced low mechanical index (MI) imaging (MI, 0.3). Harmonic imaging (MI, 1.6; gain, 50%; compression, 70%) was used for unenhanced imaging. A low MI of 0.3 was preselected (gain, 60%; compression, 15%) for contrast-specific imaging. Contrast enhancement was obtained by infusion of SonoVue® (Bracco Imaging SpA, Milan, Italy) with an average infusion rate of 1.35 mL/min and bolus injections, when necessary. Cineventriculography was accomplished with biplane technique using the 30° right anterior oblique and the 60° left anterior oblique projections. Cardiac MRI was performed on a 1.5 T system during breathhold with electrocardiographic triggering. Using a temporal resolution 50 ms, 4-, 3-, and 2-chamber views, as well as short-axis views, with a slice thickness of 10 mm in basoapical direction were obtained. Adherence to these guidelines was monitored.
Diagnostic images were read by 2 independent offsite readers and 1 onsite reader blinded to the patients’ diagnoses. To define regional left ventricular function, standard protocols were used with a 16-segment model for MRI and echocardiography, and a 7-segment model for cineventriculography. For each segment, myocardial contraction was scored as normokinetic, hypokinetic, or akinetic. A wall motion abnormality was defined as the presence of hypokinesia or akinesia in at least 1 ventricular segment.
Because no imaging modality may be considered the gold standard in the evaluation of regional wall motion abnormalities, agreement was evaluated based on the diagnosis made by an independent expert panel. The panel consisted of 2 cardiologists who determined the presence or absence of regional wall motion abnormalities according to a diagnostic algorithm that included evaluation of clinical records (history, electrocardiography, and coronary angiography), results of the onsite and offsite readers, and results of the 4 sets of diagnostic images.
Results
Regional wall motion abnormalities were detected in 58% to 86% of patients, depending on the imaging modality and the readers’ subjective evaluations. For unenhanced and contrast-enhanced echocardiographic images, there were similar rates of patients with abnormalities (63%-67%), whereas there was greater variability in these rates when MRI and cineventriculography were assessed. Interobserver agreement among the 3 readers of each imaging modality for the detection of regional wall motion abnormalities was highest for contrast-enhanced echocardiography (mean kappa value = 0.77). MRI and cineventriculography had intermediate levels of interobserver agreement (kappa = 0.43 and 0.56, respectively). Interobserver agreement was lowest with unenhanced echocardiography (kappa = 0.41). Overall, the intermethod agreement regarding the presence of regional wall motion abnormalities was relatively poor (kappa < 0.59) for all the imaging modalities used. The kappa coefficient for intermethod agreement between unenhanced echocardiography and cineventriculography was 0.28, and between unenhanced echocardiography and MRI was 0.29. Application of the contrast agent increased the kappa coefficient for intermethod agreement noticeably, to 0.46 with MRI and 0.59 with cineventriculography.
Considering the panel decisions on the presence of regional wall motion abnormalities, the accuracy of each of the 4 imaging modalities was assessed. The expert panel judged regional wall motion abnormalities to be present in 67 of the 100 patients. Based on the expert panel definition of wall motion abnormalities, contrast-enhanced echocardiography was most accurate for the detection of regional wall motion abnormalities (87.2%), whereas unenhanced echocardiography and cineventriculography were least accurate (82.9% and 82.8%, respectively). Magnetic resonance imaging had an intermediate level of accuracy (84.9%). For all 3 readers of each method, agreement was low between the panel decision and cineventriculography as well as unenhanced echocardiography. Agreement was high with 2 readers for cardiac MRI, whereas it was low for 1 offsite reader because of significant overreading. Agreement between the panel decision and contrast echocardiography was higher than agreement between the panel decision and unenhanced echocardiography or cineventriculography for the 56 cardiac MRI patients.
Discussion
The prognosis and treatment of patients can be significantly influenced by the evaluation of regional systolic function at rest as well as under stress conditions. Whereas regional wall motion abnormalities at rest may indicate scar formation or chronic myocardial ischemia, evaluation of regional wall motion abnormalities during stress conditions may allow the detection of myocardial ischemia or viability. Assessment of regional systolic function is a highly complex process. It relates to the analysis of dynamic images, with consideration of changes in wall thickness as well as endocardial inward motion. Interpretation thresholds for assessing regional left ventricular function as abnormal are subjective. The variability among observers in the interpretation of cardiac imaging tests is well known.2,3
Image quality has been shown to have an important effect on the ability to define regional wall motion abnormalities.3,4 Thus, diagnostic accuracy as well as interobserver agreement in the interpretation of test results relate to image quality. The image quality of MRI is presumed to have advantages because of the high spatial resolution in almost all patients. Because of the usually excellent image quality and full volumetric data acquisition, MRI is known for high accuracy and reproducibility in the assessment of left ventricular volumes and ejection fractions.4,5 However, recent studies have indicated that MRI is far from perfect on interobserver agreement with regard to regional function analysis. According to a study by Paetsch and colleagues, 150 patients undergoing dobutamine (Dobutrex) stress MRI had an average level of reader agreement on test results as defined by a kappa coefficient of only 0.55 between 3 expert readers from different centers.6 Contrast echocardiography has been shown in several studies to significantly improve endocardial border delineation and thereby reader confidence in global left ventricular function analysis.7-9 Contrast enhancement has been shown to improve analysis of global left ventricular function and also to have a positive effect on regional function analysis.
This study permitted the first direct within-patient comparison of cineventriculography, unenhanced echocardiography, contrast-enhanced echocardiography, and cardiac MRI to evaluate the presence of regional wall motion abnormalities.1 Results confirm that contrast echocardiography has a strong effect on assessment of regional left ventricular function. Use of contrast echocardiography was associated with the highest interobserver agreement on test results and significantly improved the accuracy in the detection of consensus-defined regional function abnormalities. However, the study also indicates that even when using high-quality imaging modalities, regional wall motion analysis is characterized by a considerable interobserver variability resulting from different reader thresholds in the evaluation of regional wall motion abnormalities. Thus, objective quantitative measures of regional left ventricular function are desirable and should assist in the evaluation of regional wall motion in the future.
Conclusion
We compared the use of cineventriculography, contrast-enhanced echocardiography, unenhanced echocardiography, and cardiac MRI to define the presence of regional left ventricular wall motion abnormalities. Contrast-enhanced echocardiography allowed good interobserver agreement in regional wall motion analysis and displayed the highest accuracy in detecting panel-defined wall motion abnormalities.