ABC | Volume 110, Nº3, March 2018

Original Article Kang et al Multi-layer strain analysis in lymphoma survivors Arq Bras Cardiol. 2018; 110(3):219-228 ECG, arrhythmia, a previous history of heart failure and/or coronary artery disease. The following data were collected including date of completion of chemotherapy, duration of follow-up, cumulative doses of anthracyclines, symptoms and signs of heart failure, and cardiac medications. Twenty‑eigh age-matched and gender-matched referred to our hospital for non-specific chest pain or palpitation, but with no history of cardiovascular disease patients and with completely normal electrocardiograms, echocardiograms, treadmill stress exercises and 24-hour, continuous ambulatory electrocardiograms were selected as controls (control group). Echocardiographic imaging Images were obtained in the left lateral decubitus position with Vivid E9 (GE Healthcare, Horton, Norway) ultrasound systems. Standard two-dimensional images were acquired according to recommendations of the American Society of Echocardiography. 9 Depth was minimized to optimize the frame rate. At least three beats were digitally stored for offline analysis. Left ventricular ejection fraction was calculated using the modified Simpson's biplane method. Left ventricular mass index, relative wall thickness, transmitral peak early (E) and peak late (A) diastolic filling velocities were also measured. Tissue Doppler echocardiography was performed with the sample volume positioned at the basal LV free wall and septum at the mitral annular junction to obtain lateral and septal mitral annular systolic (S’) and early diastolic myocardial tissue velocities (E’). Multi-layer speckle tracking echocardiography Gray scale images for offline speckle tracking analysis were acquired at frame rate of 53 to 84 MHz. Echopac version 11.1 (GE Healthcare, Horton, Norway) was used for multi-layer strain analysis. The automatic tracking analysis was performed in the apical 4-chamber, 2-chamber, long-axis apical view for longitudinal strain and in the parasternal short-axis view at basal, mid-papillary and apical level for circumferential and radial strain according to the vendor's instructions. The endocardial border was manually traced at end-diastole. The ROI (region of interest) for strain analysis was adjusted manually. The locations of the tracking points were adjusted when necessary so that the region of interest extended from endocardial to epicardial borders to approximate the myocardium, which was divided into subendocardial, middle and subepicardial myocardium layers of equal thickness. Peak circumferential (CS) and radial strain (RS) measurements were obtained from the basal, mid-segments of the septal, lateral, inferior, anterior, anteroseptal, posterior walls, apical segments of anterior, inferior, septal, lateral walls, totally 18 segments. Peak longitudinal strain (LS) measurements were obtained from the basal, mid- and apical segments of the anterior, inferior, anteroseptal, anterolateral, inferoseptal, inferolateral walls, totally 16 segments. In each segment, the subendocardial, middle and subepicardial LS and CS were calculated automatically. Regional strain values were averaged to determine global longitudinal/circumferential/radial strain (GLS, GCS, GRS), global subendocardial, middle and subepicardial LS (LS-ENDO, LS-MID, LS-EPI) and CS (CS-ENDO, CS-MID, CS-EPI). Transmural strain gradient was calculated as differences of peak systolic strain between the subendocardial and subepicardial layers. Strain values of each level were calculated. Reproducibility Intra- and inter-observer reproducibility was assessed by calculating the difference between the values of 10 randomly selected patients measured by one observer twice and by a second observer. Statistical analysis Continuous variables with normal distributionwere expressed as the mean ± standard deviation. Continuous variables with non-normal distribution were expressed as median and interquartile range. Differences between two groups were determined using independent samples t test for continuous variables with normal distribution and Kruskal Wallis test for with non-normal distribution. One sample K-S test was used in determining the normality of data. One way ANOVA test was used to compare the differences between strain values of different layers and different levels within each group. Relations between strain values and cumulative anthracycline dose were determined by Pearson correlation analysis. Interobserver and intraobserver reproducibility of strain values were assessed using intraclass correlation coefficients (ICCs) and Bland-Altman analysis. Data were analyzed by SPSS version 16.0 (SPSS, Inc, Chicago, IL, USA). A value of p < 0.05 was considered significant. Results Three patients and one healthy volunteer were excluded from the analysis because of poor image quality (defined as > 2 non-visualized segments). Forty-two patients, 18 males, ranging in age from 22 to 77 years (mean age 55.83 ± 17.92 years), and 27 healthy volunteers, 14 males, ranging in age from 32 to 77 years (mean age 51.39 ± 13.40 years), were finally included in the statistical analysis. Table 1 shows the two groups clinical characteristics. In all patients, the cumulative dose of epirubicin was 319.67 ± 71.71 mg/m 2 (ranging from 150.94 mg/m 2 to 440.00mg/m 2 ). Patients have not received radiotherapy or other cardiotoxic agents. No patient complained of cardiovascular related symptoms. EKG remained normal in all patients. The time from last dose of epirubicin to the echocardiographic examination was 52.92 ± 22.32 months (ranging from 24 months to 104 months). Conventional echocardiographic parameters Table 2 summarized the echocardiographic findings in the two groups. Conventional parameters of systolic and diastolic function, including body surface area indexed left ventricular end-diastolic volume (LVEDV/BSA), body surface area indexed left ventricular end-systolic volume (LVESV/BSA), left ventricular ejection fraction, E velocity, A velocity, E/A ratio, deceleration time, S’ velocity, E/E’, isovolumic relaxation time showed no significant difference between two groups. 220