Intervendor consistency and reproducibility of left ventricular 2D global and regional strain with two different high-end ultrasound systems.

Abstract

We aimed to assess intervendor agreement of global (GLS) and regional longitudinal strain by vendor-specific software after EACVI/ASE Industry Task Force Standardization Initiatives for Deformation Imaging. Fifty-five patients underwent prospective dataset acquisitions on the same day by the same operator using two commercially available cardiac ultrasound systems (GE Vivid E9 and Philips iE33). GLS and regional peak longitudinal strain were analyzed offline using corresponding vendor-specific software (EchoPAC BT13 and QLAB version 10.3). Absolute mean GLS measurements were similar between the two vendors (GE -17.5 ± 5.2% vs. Philips -18.9 ± 5.1%, P = 0.15). There was excellent intervendor correlation of GLS by the same observer [r = 0.94, P < 0.0001; bias -1.3%, 95% CI limits of agreement (LOA) -4.8to 2.2%). Intervendor comparison for regional longitudinal strain by coronary artery territories distribution were: LAD: r = 0.85, P < 0.0001; bias 0.5%, LOA -5.3 to 6.4%; RCA: r = 0.88, P < 0.0001; bias -2.4%, LOA -8.6 to 3.7%; LCX: r = 0.76, P < 0.0001; bias -5.3%, LOA -10.6 to 2.0%. Intervendor comparison for regional longitudinal strain by LV levels were: basal: r = 0.86, P < 0.0001; bias -3.6%, LOA -9.9 to 2.0%; mid: r = 0.90, P < 0.0001; bias -2.6%, LOA -7.8 to 2.6%; apical: r = 0.74; P < 0.0001; bias -1.3%, LOA -9.4 to 6.8%. Intervendor agreement in GLS and regional strain measurements have significantly improved after the EACVI/ASE Task Force Strain Standardization Initiatives. However, significant wide LOA still exist, especially for regional strain measurements, which remains relevant when considering vendor-specific software for serial measurements.