Interobserver, Intraobserver and Intrapatient Reliability Scores of Myocardial Strain Imaging with 2-D Echocardiography in Patients Treated with Anthracyclines

Abstract

Myocardial strain imaging with 2-D echocardiography is a relatively new noninvasive method to assess myocardial deformation. To determine the interobserver, intraobserver and intrapatient reliability scores, we evaluated myocardial strain measurements of 10 asymptomatic survivors of childhood cancer. Ten patients were selected randomly out of a follow-up cohort of childhood cancer survivors. All 10 patients underwent a transthoracic echocardiographic examination. Two-dimensional gray scale images were made in parasternal apical four-chamber, apical two-chamber, midcavity short-axis and basal short-axis views. Offline analysis was performed using software for echocardiographic quantification (Echopac 6.1.0, GE Medical Systems, Horten, Norway). All echocardiographic studies were analyzed offline by three observers, separately (A.M., G.W., M.P.). A custom-made software package was designed for averaging the strain curves of three consecutive cardiac cycles. Values of peak systolic strain, time-to-peak strain and time-to-end systole of the different segments of the left ventricle were used for statistical analysis. Interobserver, intraobserver and intrapatient reliability were expressed as intraclass correlation coefficients (ICCs). Interobserver ICCs of peak strain, time to peak strain and time to aortic valve closure (AVC) were generally good to very good in all views and segments, except for in the two-chamber view. Intraobserver ICCs were rated as very good for almost all segments, except for the longitudinal peak strain values of the two-chamber view. Intrapatient ICCs were generally good for the two-chamber, four-chamber and midcavity short-axis views, but fair to moderate for the segments of the basal short-axis view (SaxMV). We recommend use of the four-chamber view for longitudinal peak strain values, and the basal and midcavity short-axis views for radial and circumferential peak strain values. Furthermore, we strongly recommend using the average of three cardiac cycles for peak strain values in clinical studies. (E-mail: [email protected])

Introduction

Myocardial strain imaging with 2-D echocardiography is a relatively new noninvasive method to assess myocardial deformation (Reisner et al. 2004). It is considered to be less affected by tethering, translational artefacts and traction than Doppler measurements of myocardial velocities (Korinek et al 2005, Leitman et al 2004). Tissue movement is estimated using frame-by-frame image tracking. Ultrasound tissue images contain many small picture elements (speckles) and natural acoustic markers, which move together with the tissue and do not change their pattern significantly in the time between the successive frames. In this way, displacement of tissue can be assessed in 2-D for the whole image. Each speckle can be followed accurately during several consecutive frames. A different displacement between regions reflects deformation of the tissue by contraction or relaxation (Leitman et al. 2004). Strain is the change in size relative to its original size, i.e., the relative deformation. Regional deformation of the myocardium occurs in the three orthogonal directions: longitudinally, radially and circumferentially. Using 2-D strain imaging, two of these components can be determined, depending on the used image plane. Myocardial strain imaging is still used only in an experimental setting and has not yet gained extensive clinical use. The first studies on reproducibility of this technique have been published (Serri et al 2006, Hurlburt et al 2007, Chan et al 2006, Reant et al 2008). We are currently evaluating the use of strain rate imaging (SRI) for the early detection of anthracycline-induced cardiotoxicity in a large study. To determine the interobserver, intraobserver and intrapatient reliability scores, we evaluated myocardial strain measurements of 10 asymptomatic survivors of childhood cancer.