Cardiovascular Function Phenotypes in Response to Cardiotoxic Breast Cancer Therapy

Cardiotoxicity, defined by declines in left ventricular ejection fraction (LVEF) and symptomatic heart failure (HF), is a significant complication of commonly used treatments for breast cancer patients (1). Although recent studies have resulted in a better understanding of the changes in LVEF with anthracyclines and trastuzumab, there are numerous unanswered questions pertaining to an individual patient’s response to potentially cardiotoxic cancer therapy. For example, are there distinct patterns of LVEF change with anthracyclines and/or trastuzumab? What is the significance of modest LVEF declines that do not meet the standard criteria for cardiac dysfunction (i.e., LVEF decline of $10% to <50%)? Techniques such as latent class analysis have been used to understand how changes in cardiovascular risk factors inform the development of overt cardiovascular disease and to define new phenotypes in heterogeneous disease states (2). We sought to determine if we could apply such methods to breast cancer patients treated with anthracyclines and/or trastuzumab to understand the potential distinct cardiovascular responses to these cardiotoxic therapies.

The CCT (Cardiotoxicity of Cancer Therapy) study is a prospective cohort study of breast cancer patients from the Rena Rowan Breast Center of the Abramson Cancer Center at the University of Pennsylvania (3). Latent class growth modeling was used to identify subgroups with distinct trajectories of LVEF change over time (4). Absolute change in LVEF from baseline was included as the dependent variable; time, modeled as a cubic spline with 4 degrees of freedom, was the independent variable. A priori, the maximum possible number of LVEF trajectory classes was limited to 5 given the moderate sample size. Bayesian information criterion was used to determine the optimal number of LVEF trajectory classes.

A total of 314 patients contributed 1,437 echocardiograms over a maximum follow-up of 3.5 years. A median of 4 (interquartile range: 3 to 6) echocardiograms were performed per patient. A 3-class model was selected based on the pre-specified goodness of fit criteria. The results of this model revealed 3 major patterns of LVEF change over time (Figure 1A). In Class 1, which included 51% of patients, a stable LVEF trajectory throughout the 3.5 years of followup was observed (mean LVEF change −0.7 ± 1.6%). In Class 2, a modest LVEF decline occurred within the first 6 months; this decline persisted throughout follow-up (mean change −3.7 ± 1.4%). A total of 40% of the cohort comprised this group. In Class 3, which consisted of 9% of the cohort, there was a significant LVEF decline in the first year followed by partial recovery (mean change −8.4 ± 2.1%).

FIGURE 1. LVEF Trajectory Classes and Accompanying Changes in Circumferential Strain.

Distinct classes of left ventricular ejection fraction (LVEF) change over time in breast cancer patients treated with anthracyclines and/or trastuzumab (A), and accompanying changes in circumferential strain according to LVEF trajectory class (B). Solid lines represent estimates of mean change and dotted lines represent 95% confidence intervals.

Further characterization of changes in circumferential strain revealed important differences across the LVEF trajectory classes (Figure 1B). Circumferential strain was not significantly altered in Class 1. However, circumferential strain showed a sustained reduction in Class 2, consistent with a persistent worsening of cardiac function and mechanics. In Class 3, a pronounced reduction in strain was seen for up to 1 year, followed by recovery after 2 years. An increase in HF symptoms was noted across all classes, worst in Class 3, but persistent in Class 2.

In summary, heterogeneity exists in an individual patient’s cardiovascular “response” to potentially cardiotoxic cancer therapy. In a rigorously phenotyped cohort of breast cancer patients treated with anthracyclines and/or trastuzumab, we sought to gain insight into this heterogeneity by determining if there are distinct subgroups of LVEF change. We identified 3 classes, including those who demonstrated stable LVEF over time (Class 1); early, modest, and persistent LVEF decline (Class 2); and early, significant LVEF decline followed by partial recovery (Class 3). Notably, 40% of the patients comprised Class 2, and persistent abnormalities in a sensitive imaging marker of cardiac dysfunction and HF symptoms were observed in this class. Our findings suggest that modest LVEF declines in breast cancer patients exposed to anthracyclines and/or trastuzumab could be more clinically important than previously understood. Careful monitoring and longer-term follow-up will determine the incidence of HF over time.