Abstract 260: Congruence of Speckle Tracking Echocardiography and Magnetic Resonance Imaging Assessment of Left Ventricular Ejection Fraction

Abstract

Background: Assessment of left ventricular ejection fraction (LVEF) is an important aspect of diagnosing and treating patients with cardiovascular disease. While cardiac MRI is considered the gold standard for assessing LVEF, high cost and patient risk has precluded routine use of MRI in clinical practice. Echocardiography is a common non-invasive, safe, and economic method for assessing LVEF. Traditionally, these measures required making visual determinations, resulting in wide variations in LVEF estimates. Using calculations from global longitudinal peak strain (GLPS) in speckle tracking echocardiography (STE) might reduce variability in estimations of LVEF and enhance treatment decision-making. To examine whether STE GLPS compared to cardiac MRI provides clinically acceptable variation in LVEF estimations, the congruence between GLPS in STE and cardiac MRI was examined. Methods: At a single Midwestern regional referral center, medical records were abstracted for inpatients and outpatients with a first-listed cardiovascular disease diagnosis who were ≥18 years of age and who were evaluated using speckle tracking echocardiogram and magnetic resonance imaging procedures within a 30-day period between January, 2011 and May, 2014. Demographic and relevant clinical variables that might affect echocardiographic and MRI quality and interpretation were extracted from medical records. Overall agreement between LVEF estimates from cardiac MRI and from STE GLPS was assessed using the concordance correlation coefficient, Bland-Altman analysis, and weighted Kappa. The influence of demographic and co-morbidity factors on the agreement was also assessed using multivariable regression analysis. Results: A total of 93 patient medical records (48 males; age 59±18 years) underwent both STE and MRI within a maximum 30-day period. For quantification of LVEF, STE correlated well with MRI (r=0.73; CCC=0.67), but LVEF was significantly underestimated (bias=6.8%), with wide limits of agreement (-17% to 31%). Categorization of LVEF into four classes, ranging from severely impaired to normal, resulted in a weighted kappa of 0.65 (95% CI: 0.53, 0.77). The difference between MRI and STE increased with increasing BMI (b=0.48, P=.01), and this difference was greater for those with hypertension (mean=9.5) than those without hypertension (mean=3.8, P=.02). Discussion and Conclusions: While these two measures indicate moderate to substantial agreement, STE-derived LVEF was underestimated in most patients compared with MRI. The magnitude of STE underestimation was positively associated with both BMI and hypertension. Further research is needed on technological developments in echocardiography, as well as interventions focused on addressing inter-observer agreement variability and treatment decision-making to effectively achieve desired clinical outcomes.