P783 Global longitudinal strain detects trastuzumab induced cardiotoxicity early in a clinical population

Abstract

Abstract Background Trastuzumab is highly effective in the treatment of breast cancer, and is often used as an adjuvant therapy. Due to its potential cardiotoxicity, serial monitoring of cardiac function is vital. Ejection fraction (EF) by two-dimensional echocardiography is routinely used but has limitations in measurement variability. Myocardial deformation imaging, in particular Global Longitudinal Strain (GLS), can detect pre-clinical myocardial dysfunction. However, its use is not yet adopted into routine clinical practice. Aims Our aim was to ascertain if a clinically significant reduction in GLS (≥11% from baseline) occurred before the onset of EF reduction, in patients who developed cardiotoxicity whilst receiving trastuzumab. Methods Between January 2014 and January 2019, 235 consecutive patients received trastuzumab and underwent serial echocardiography at 3 monthly intervals at our institute. Cardiotoxicity is defined as a ≥10% EF reduction from baseline or an EF <50%. Women who developed cardiotoxicity as defined by this change in EF were retrospectively studied. Two-dimensional speckle tracking was used to derive peak longitudinal strain in each myocardial segment from the apical four-, three- and two-chamber view images. GLS was taken as the average value of all these segments. The median time to ≥11% GLS reduction and ≥10% EF reduction or EF <50% was compared. Results Thirteen women (mean age 53 ± 9.5 years) developed cardiotoxicity. EF was 61.8 ± 4.4% at baseline and 45.7 ± 7.5% following therapy (p = 0.00). A ≥11% reduction in GLS from baseline was observed in all patients: GLS -20.2 ± 1.5% and -15.6 ± 2.1%, p = 0.00. The median time to cardiotoxicity as defined by EF and GLS was 6 months and 3 months, respectively (p = 0.031), as shown in Table 1. Repeatability analysis showed both EF and GLS measurements in our cohort have good measurement reproducibility. Inter-observer intraclass correlation (ICC) for EF and GLS were 0.912 and 0.913, respectively. Intra-observer ICC for EF and GLS were 0.925 and 0.900, respectively. Conclusion Cardiotoxicity developed in a significant portion (6%) of our patients receiving trastuzumab. As a reduction in GLS was detectable early and preceded that of EF by 3 months, this may represent a therapeutic window for initiation of cardio-protective medication, if and when the use of GLS is incorporated into routine practice for cardiotoxicity surveillance. Table 1: Summary of Findings. Baseline Cardiotoxicity Paired t-test p value Median time of onset EF 61.8 ± 4.4% 45.7 ± 7.5% P < 0.01 6 months* GLS -20.2 ± 1.5% -15.6 ± 2.1% P < 0.01 3 months* *denotes exact sign-test p < 0.05.