Abstract 17216: Limited Concordance of Core Lab Left Ventricular Ejection Fraction With Automated Assessments: A Sub-Study From VALOR-HCM

Abstract

Introduction: The performance of automated left ventricular ejection fraction (LVEF) assessment in patients with altered cardiac chamber geometry, such as those with obstructive hypertrophic cardiomyopathy (HCM), is not well described. Aims: Accurate assessment of LVEF is required for dose titration and continuation of myosin inhibitors in obstructive HCM. From the VALOR-HCM trial (NCT04349072), the agreement of two commercially available software packages was compared with blinded core lab assessment of LVEF. Methods: From baseline and week 16, among 112 patients randomized to either mavacamten or placebo, 178 echocardiograms were analyzed. Representative clips were manually selected, and automated LVEFs were calculated. Comparisons of LVEFs were made using Pearson’s correlation coefficients and Bland-Altman analysis. Results: With core lab assessment, the overall median LVEF was 68% (Q1 66%, Q3 70%). The correlation with vendor 1 was moderate (r = 0.47, 95% CI 0.35-0.58, p <0.001) and weak with vendor 2 (r =0.21, 95% CI 0.07-0.35, p =0.004). Correlation was also weak between vendor 1 and 2 (r = 0.14, 95% CI 0.00-0.28, p = 0.06). Vendors 1 and 2 had an underestimation bias of LVEF of 8% (95% CI 6-9%) and 8% (95% CI 7-9%), respectively, compared to core lab LVEF (Figure 1). Conclusions: In VALOR-HCM, the concordance of LVEF between different commercially available software packages with each other and with a core lab assessment was limited. Further study is needed to evaluate and improve the performance of automated LVEF tools in obstructive HCM patients who have altered cardiac chamber geometry.