Global longitudinal strain: an early marker for cardiotoxicity in patients treated for breast cancer

Cardio-oncology is an emerging field of interest in cardiology that focuses on the detection, monitoring and treatment of cardiovascular disease occurring as a side-effect of chemotherapy and radiotherapy. Cardiological assessment before starting potentially cardiotoxic cancer treatment is essential and should be continued throughout the oncological treatments, since left ventricular (LV) dysfunction can occur at any time.

The cardiotoxicity of anthracyclines is well-recognised, dose-related and irreversible [1]. Trastuzumab increases the cardiotoxicity of anthracycline treatment and can also cause cardiotoxicity when given on its own, but this cardiotoxicity is reversible [2]. When giving combined treatment with anthracyclines and trastuzumab, LV dysfunction has been reported in up to a third of the patients, with overt heart failure in up to 16% of patients [3].

A recent consensus paper from the European Society of Cardiology proposed a definition of chemotherapy-related cardiotoxicity as an LVEF decrease of > 10 percentage points from baseline to a value < 53% on repeat confirmatory echocardiographic imaging [4]. Nevertheless, routinely used two-dimensional (2D) and three-dimensional (3D) LV ejection fraction (LVEF) measurements may fail to detect subtle subclinical signs of LV dysfunction, which represent the first warning signs for cardiotoxicity. Once LVEF has deteriorated, it may be too late to reverse the anthracycline-induced cardiotoxicity. Monitoring the development of early subclinical cardiotoxicity may therefore be crucial; however, clear guidelines on how to do this are lacking.

Global longitudinal strain (GLS) has been found to be a useful marker to detect subclinical myocardial dysfunction [5, 6]. An early reduction in GLS of 10–15% during cancer therapy appears to be a useful parameter for the prediction of cardiotoxicity [4]. It has previously been shown that early deterioration of GLS can predict LVEF decline 3 months later [5]. GLS-guided introduction of cardioprotective medication been has shown to reduce the fall in LVEF at 1 year compared to standard LVEF guidance [7]. However, in these studies 2D echocardiography was used to assess LVEF, while 3D echocardiography has proven to be superior [8]. The added value of GLS over 3D LVEF is therefore unknown. Also, so far little is known about why some patients develop cardiotoxicity but others do not.

The primary objective of our study is to assess whether subclinical deterioration of GLS is a more reliable early predictor for LV dysfunction than 3D LVEF and can predict 3D LVEF decline later on. Furthermore, we will examine which baseline factors are associated with a high risk of cardiotoxicity in a cohort of women with breast cancer treated with anthracyclines followed by trastuzumab.

In total, 51 females with a mean age at baseline of 54 ± 12 years were included in the study. After review of the echocardiograms, one patient was excluded due to inadequate quality of the echocardiographic imaging.

Baseline characteristics and echocardiographic characteristics at baseline are presented in Table S1 (Electronic Supplementary Material). In total, 12 (26%) patients had cardiac medications at baseline. During a mean follow-up duration of 1.1 ± 0.45 years, 5 patients died, yielding a Kaplan-Meier survival estimate of 70.5 ± 13.9% at 5 years. None of the deaths were related to a cardiovascular event. None of the patients showed signs of clinical heart failure or any cardiac event during follow-up.

While there are several reports in the literature about early detection of cardiotoxicity, ours is the first to compare 3D LVEF and GLS over time using a linear mixed-effects model over an extended follow-up period during trastuzumab therapy. We have demonstrated that GLS deteriorates earlier than 3D LVEF. Therefore, GLS may still be of added value regarding the detection of cardiotoxicity, even in the 3D echocardiography era.

The principal findings from this study indicate that: (1) subclinical LV dysfunction can be detected on average 30 days earlier by a GLS decrease than by a 3D LVEF decline; (2) no baseline characteristics, such as age, hypertension and CRS, could be identified as predictors for cardiotoxicity; (3) in our cohort none of the patients reached a clinical endpoint and LV dysfunction was reversible after pausing trastuzumab.

GLS is an established preclinical marker of cardiac dysfunction [4]. It is well known that GLS provides independent and incremental prognostic information regarding long-term risk of cardiovascular morbidity and mortality [14]. The longitudinal myocardial fibres responsible for longitudinal shortening seem particularly vulnerable at a relatively early stage to a wide range of cardiac diseases, including cardiotoxicity [7]. However, a 15% deterioration of GLS as a cutoff to be used in clinical practice will still be subject to discussion. The cutoff is based on small studies, and the predictive value with respect to heart failure is unknown [5]. In our cohort there were no clinical consequences, and it may be debated whether acting on this cutoff value, instead of acting on a subclinical drop in LVEF, adds to the long-term perspective of these patients.

In contrast to other studies, we were not able to demonstrate with our advanced statistical model that parameters such as increasing age, hypertension, 3D LVEF at baseline or CRS are clinical predictors for the development of cardiotoxicity [15, 16]. This is most likely due to our small sample size. Pilot studies have shown the great potential of a combination of echocardiographic screening and various biomarkers [17, 18]. Future research should be directed at optimising identification of the patients at highest risk of developing cardiotoxicity, who deserve intensified cardiac follow-up, in order to warrant the most efficient screening.

In our study, once subclinical 3D LVEF deterioration occurred, trastuzumab was paused and heart failure therapy was initiated empirically on an individualised basis by the treating cardiologist. Whether initiation of cardioprotective medications improves the outcome in patients with subclinical cardiac dysfunction is not entirely clear. A study by Boekhout et al. did not support the hypothesis that concomitant use of candesartan protects against a decrease in LVEF during or shortly after trastuzumab treatment in breast cancer patients [19]. In the MANTICORE trial, 94 patients were randomised to either placebo, perindopril or bisoprolol to investigate the cardioprotective effect. The authors showed signs of myocardial inflammation early after trastuzumab initiation. Patients on bisoprolol showed no deterioration of GLS after the first cycles in contrast to the other two groups. Both bisoprolol and perindopril were associated with preserved LVEF after completion of trastuzumab treatment [20]. In another trial, 468 trastuzumab-treated patients were randomised to either angiotensin-converting enzyme (ACE) inhibitors or beta-blockers or placebo. The authors found fewer interruptions of trastuzumab treatment in the first two groups, and the therapy prevented cardiotoxicity [21].

An ongoing larger randomised controlled trial with angiotensin-receptor inhibitors and beta-blockers in patients receiving anthracycline and trastuzumab and evaluated with strain measurements is still awaited [22]. The SUCCOUR trial, including 331 anthracycline-treated patients, showed that GLS-guided initiation of cardioprotective medication (ACE inhibitors and beta-blockers) may lead to a smaller reduction in LVEF after 1 year, in patients who received this cardioprotective treatment [7]. We demonstrated that, in all patients with cardiotoxicity in our study, 3D LVEF and GLS did recover after temporarily pausing trastuzumab and/or starting cardioprotective medication, even when trastuzumab was continued after a short pause. These are preliminary observations, which require longer-term follow-up. Some evidence indicates that, several years after chemotherapy, GLS values in asymptomatic cancer survivors remain significantly lower than in the overall population [23].

While the clinical consequence of subclinical cardiotoxicity remains uncertain, interrupting cancer therapy might have deleterious effects with regard to cancer outcome. Two studies aimed to determine whether trastuzumab may be continued despite a subclinical LVEF decline. One small but interesting prospective study suggested that, in patients administered ACE inhibitors and beta-blockers, trastuzumab may be continued safely despite mild cardiotoxicity [24]. However, two patients (10%) developed heart failure with an LVEF < 40%, but completely recovered after discontinuation of trastuzumab. Further randomised controlled trials are required to confirm this outcome in low-risk patients.

The SAFE-HEaRt trial studied LVEF during trastuzumab treatment in patients with an LVEF of 40%–49% prior to trastuzumab initiation [25]. They were treated with beta-blockers and/or ACE inhibitors. In 27 patients, trastuzumab treatment could be completed. However, trastuzumab was discontinued in 3 patients because of heart failure or an LVEF of < 35%. Two of them died due to disease progression. It is unknown whether the cardiac outcome outweighs the cancer outcome with respect to trastuzumab discontinuation.

Myocardial strain imaging using GLS is able to identify subclinical LV dysfunction earlier than 3D LVEF measurement in women undergoing treatment with anthracyclines followed by trastuzumab for breast cancer. After pausing trastuzumab and starting cardioprotective therapy, this LV dysfunction is potentially reversible. It remains to be established whether the use of GLS to guide early cardioprotective therapy results in improved clinical outcomes and which patients are at highest risk of developing cardiotoxicity.