Study Group

Forty consecutive female patients (mean age, 51 ± 8 years) with breast cancer, scheduled to be treated with epirubicin, were enrolled prospectively in the study. Inclusion criteria were (1) age > 18 years, (2) provision of written informed consent signed, (3) previous untreated HER2-negative breast cancer scheduled for treatment with chemotherapy according to international standardized protocols, and (4) EF > 50%. Exclusion criteria were (1) any history of cardiovascular disease and/or active cardiovascular treatment, (2) diabetes mellitus, and (3) mediastinal irradiation. All subjects were informed about the procedures and provided informed consent. The study protocol was approved by the local ethics committee.

All subjects were evaluated at baseline using conventional echocardiography, TVI, and STE. All measurements were averaged from ≥3 consecutive cardiac beats. The full echocardiographic protocol was repeated in all patients within 24 hours of the end of the third and sixth cycles of chemotherapy (63 ± 6 and 126 ± 12 days after baseline, respectively). Cardiotoxicity was defined as reduction in EF of ≥10% to an EF of <55%, without signs or symptoms of heart failure (HF), compared with baseline, after the sixth cycle of treatment. According to the reduction in EF after the completion of treatment, two groups were defined: patients who developed cardiotoxicity (group I) and patients without cardiotoxicity (group II).

Echocardiography

Examinations were performed using a commercially available ultrasound machine equipped with TVI and STE, using a 2-MHz to 4-MHz transducer (Vivid 7 Dimension; GE Medical Systems, Milwaukee, WI). Digital echocardiographic data were acquired during passively held end-expiration for offline analysis using a dedicated software package (EchoPAC version 9.0.1 for PC; GE Medical Systems). Heart rate and blood pressure were measured just before the study. A one-lead electrocardiogram was recorded simultaneously.

Conventional echocardiography consisted of M-mode, two-dimensional, and Doppler blood flow measurements to assess LV structure and global function. All measurements were taken in accordance with the current guidelines of the American Society of Echocardiography. M-mode tracings from the parasternal long-axis view were used to measure systolic and diastolic septal and posterior wall thickness and LV diameters and to calculate fractional shortening (FS). Two-dimensionally guided M-mode echocardiography of mitral annular motion (medial and lateral) was also performed. The systolic amplitude in the long axis was measured from the onset of annular movement toward the apex, at the peak of the QRS complex, to the maximal systolic displacement; mean displacement (mitral annular plane systolic excursion) was calculated by averaging the two sites. Cross-sectional images were recorded from the apex, and end-diastolic and end-systolic areas and LV lengths were measured from the apical four-chamber and two-chamber views (using the modified biplane Simpson’s method) for the calculation of EF. Diastolic function was assessed using pulsed-wave Doppler of the transmitral flow; the E/A ratio was calculated. LV inflow was recorded using color M-mode echocardiography, and flow propagation velocity was measured.

Online pulsed-wave tissue velocity traces were recorded from the six basal segments (from the apical four-chamber, two-chamber, and three-chamber views) to assess longitudinal function. Peak myocardial velocities in systole (S′) and early diastole (E′) were measured, and their six-segment mean was calculated. The E/E′ ratio was calculated as a marker of LV filling pressure.

STE was used to assess systolic myocardial deformation in all three directions (longitudinal, circumferential, and radial) and rotation, according to current recommendations. Frame rates of 70 to 80 frames/sec were used, with optimal sector width and depth. The frame in which the LV endocardium was best defined was identified, and the border of the endocardium was traced manually to identify the region of interest between the endocardial and epicardial borders. After definition of the region of interest in the long-axis or short-axis views, the postprocessing software automatically divided the ventricle into six equally distributed segments. All LV segments were tracked optimally in each echocardiographic view. Cardiac time intervals were measured using pulsed-wave tracings from the LV outflow tract (aortic valve opening and aortic valve closure). We calculated peak longitudinal strain (LS) and strain rate (LSR) as means of the 18 ventricular segments from the apical views (four-chamber, two-chamber, and three-chamber views), peak circumferential strain and strain rate, and peak radial strain and strain rate from the six ventricular segments, from the short-axis view at the level of the papillary muscles. “LV twist” refers to the rotation of the short-axis views of the left ventricle as viewed from the apical end and is defined as the absolute apex-to-base gradient rotation along the longitudinal axis of the left ventricle, expressed in degrees. We measured apical rotation, a counterclockwise rotation during LV ejection and the most important part of LV twist, using the short-axis apical view well beyond the papillary muscle and basal rotation, which is significantly lower in magnitude than apical rotation and opposite in direction, at the basal short-axis view. Subsequent recoil of twist deformation, so-called untwist, is associated with the release of restoring energy and contributes to LV diastolic relaxation and early diastolic filling. As a parameter of a diastolic mechanic function, we measured early untwist rate (degrees per second) as the mean difference between the untwist curves. All examinations were performed by the same experienced investigator.

The reproducibility and repeatability of all echocardiographic parameters were measured in our laboratory in 20 consecutive patients by two observers with similar experience. Intraobserver and interobserver variability and test-retest variability were calculated as 2SD/√2 and are reported as percentages from the mean value to give the coefficient of variation.

Statistical Analysis

Statistical analysis was performed with SPSS version 19 (SPSS, Inc, Chicago, IL). Results are presented as mean ± SD or as percentages. Two-tailed P values < .05 were considered significant. The minimum number of patients required was calculated to be 20, on the basis of our reproducibility of tissue Doppler and STE, meaning that significant changes of ≥5% can be observed between consecutive measurements when the sample size is ≥20. Differences between the two groups for the baseline characteristics and echocardiographic parameters were calculated using independent t tests for differences in means and using χ ² tests for differences in proportions. Comparisons of echocardiographic parameters within each group versus baseline after the third and sixth cycles of epirubicin were performed using one-way analysis of variance for repeated measurements, with subgroups analysis using the Scheffé test. Univariate Pearson’s correlation was used to investigate the relation between two parametric variables. Multiple linear stepwise regression analysis was used to assess the independent influence of selected parameters on reduction of EF of ≥10% after the sixth cycle of epirubicin. Receiver operating characteristic (ROC) curves were used to assess optimal cutoff points for independent parameters to predict reduction in EF of ≥10% after the sixth cycle of chemotherapy. Because of the relatively small sample size of the group with cardiotoxicity, the differences between baseline and third cycle measurements for S′, LS, LSR, E/A ratio, and flow propagation velocity were used input variables. This allowed the development of binormal ROC curves and the calculation of cutoff values on the basis of optimal choices for sensitivity and specificity from the curves. Analysis of the ROC curves was used to determine cutoff values and to calculate the sensitivity, specificity, positive predictive value, and negative predictive value for reduction (as percentages) for S′, LS, LSR, E/A ratio, and flow propagation velocity after the third cycle of epirubicin. Sensitivity, specificity, positive predictive value, and negative predictive value were calculated assuming a 25% prevalence of cardiotoxicity in the study population. P values < .05 for asymptotic significance with the null hypothesis (true area, 0.5) were used to compare the areas under two ROC curves.