Together, cancer and cardiac disease are the leading causes of death in North America. Recognizing this, in 2010 a group of oncologists saw the need for cardiologists and oncologists to work together and suggested that this interdisciplinary collaboration be termed “cardio-oncology” or “onco-cardiology.” Since then, this concept has evolved into a clinical discipline that focuses on the intersection of cardiac, oncologic, and hematologic disease. As a relatively young discipline, it has attracted a lot of recent attention. However, the need to understand this overlap between cancer and cardiovascular disease is not new but rather newly recognized, coinciding with the enhanced survival offered by modern cancer therapy. Acute as well as long-term cardiac sequelae of established and evolving cancer therapy are now both clinically apparent and relevant. Although cardiotoxicity associated with cancer therapy can include a variety of concerns, such as hypertension, ischemia, or arrhythmia, it is left ventricular (LV) dysfunction that has been a major focus of investigation in this area. After all, the latter is a clinical entity for which we have well-established diagnostic tools, such as echocardiography, along with a well-defined and validated therapeutic strategy in the form of angiotensin-converting enzymes and β-blockers. However, it is clear that overt LV dysfunction, as evidenced by a reduced ejection fraction (EF), is a rather late manifestation of cardiotoxicity. Therefore, more recent attention has focused on harnessing speckle-tracking echocardiography to evaluate myocardial mechanics, with global longitudinal strain (GLS) as an indicator of subclinical LV dysfunction. Indeed, a recently published expert consensus document on the imaging evaluation of patients during and after cancer therapy, defines cardiotoxicity as a symptomatic or asymptomatic decline in LV EF of >10% to a value < 53%. However, in addition to LV EF determination, this document also suggests incorporating GLS in the echocardiographic protocol, with a reduction in GLS of >15% from baseline being meaningful in a patient at risk for cardiotoxicity.

Patients with breast cancer are a fitting population in which to expand our understanding of cardiotoxicity, because they frequently receive agents associated with LV dysfunction, and many patients with early-stage disease have sufficiently long survival during which to assess potential cardiovascular outcomes. Chemotherapy agents with demonstrated potential for cardiotoxicity can be broadly classified into type I agents such as the anthracyclines and mitoxantrone, and type II agents such as trastuzumab. Taxanes, commonly part of the multidrug breast cancer regimen, have also been implicated, but little attention has been focused on these drugs to date because their cardiotoxic effects are deemed comparatively less significant. Anthracyclines are commonly used in the treatment of hematologic malignancies and breast cancer. They are associated with a dose-dependent cardiotoxicity, triggering myocyte apoptosis and necrosis. Myocyte changes can be appreciated ultrastructurally on endomyocardial biopsy. Trastuzumab, also known as Herceptin (Genentech, South San Francisco, CA), is a HER2 monoclonal antibody predominantly used in breast cancer treatment. When used concurrently or sequentially with anthracycline therapy, there can be greater risk for cardiotoxicity. Interestingly, the use of trastuzumab is not associated with ultrastructural changes in myocytes. Cardiotoxicity secondary to this agent appears to be reversible in the majority of patients.

Both anthracyclines and trastuzumab have significantly improved outcomes in breast cancer, allowing many patients to live longer and decreasing the risk for cancer recurrence. Although the majority of cardiac toxicity observed in patients receiving these agents is asymptomatic LV systolic dysfunction, we know from population data that its attendant natural history is not benign. The veiled challenge therefore is to prevent the “cancer” patient from also becoming a “cardiac” patient. Early recognition of overt or subclinical LV dysfunction allows the identification of patients who may benefit from cardiology consultation for symptom surveillance or initiation of heart failure therapy in accordance with accepted practice standards. The goal of these efforts is to usher patients safely and expeditiously through the remainder of their treatment for cancer, the most immediate threat to survival, while seeking to improve their cardiovascular prognoses.

Although the evidence base on which to determine the optimal parameters and surveillance schema for detecting LV dysfunction associated with cancer therapy is still evolving, echocardiography is likely to play a central role. Multiple studies have explored the use of Doppler-derived LV systolic and diastolic indices and/or myocardial deformation parameters, such as strain and strain rate, to determine if any of these parameters could be early indicators of LV dysfunction. On the basis of the available data, it is unclear whether early changes in diastolic parameters are able to reliably predict subsequent LV dysfunction. Data have been more compelling for the use of GLS and longitudinal strain rate as predictors of impending LV dysfunction. A recent systematic review summarizing the evidence for the use of myocardial deformation parameters in the evaluation of cardiotoxicity showed that changes in systolic myocardial deformation (especially GLS) occur earlier than changes in LV EF. Importantly, these changes may be detected with anthracycline doses that are significantly lower than the cumulative lifetime dose of 550 mg/m ² , previously thought to be a reliable threshold for increased cardiotoxicity risk.

The interesting study by Tan et al . reported in this issue of JASE was a relatively small prospective multicenter study of first-time breast cancer patients who were treated with anthracyclines, taxanes and trastuzumab and then were followed for approximately 2 years using repeated speckle-tracking-based analysis of myocardial deformation to assess cardiotoxicity. Cardiotoxicity was defined as a decline in LV EF of ≥5% to a value < 55% with symptoms of heart failure or an asymptomatic decline in LV EF of ≥10% to < 55%. The authors report that LV volumes increased and EF, longitudinal strain, and strain rate decreased at the end of treatment and remained altered, suggesting that a measurable degree of myocardial damage induced by the cancer therapy did not resolve after 2 years of follow-up. Because the chosen patient population was a relatively young one with a low cardiovascular risk profile, the observed alterations are more likely the effect of chemotherapy on the normal heart as opposed to a heart already “primed” for toxicity by underlying cardiovascular disease. It is important to note that patients had been treated with anthracycline for the first 3 months only and then given trastuzumab along with taxanes for 3 months, followed by trastuzumab alone for the remaining 9 months. Because there was no 3-month echocardiographic assessment (after anthracycline but before trastuzumab), we cannot discern if the persistent changes in LV size and function started with the anthracycline or if they were actually due to the additive effect of trastuzumab on top of anthracycline (a double hit). Future studies of patients comparing strain, LV volumes, and EF in patients receiving both agents with those receiving a trastuzumab-containing regimen without anthracycline may shed light on this issue.

An important lesson from this study is that monitoring for cardiotoxicity needs to extend beyond the treatment period, as 17% of patients who did not initially meet criteria for cardiotoxicity at the end of the treatment did later develop cardiotoxicity. Moreover, none of the patients reported any symptoms of heart failure at any point during the study; thus, the observed LV dysfunction might otherwise have gone undetected. Another important lesson is learned from the long follow-up of observation by Tan et al . Whereas most, but not all, prior studies examining the myocardial deformation indices in recipients of anthracyclines and trastuzumab have had follow-up periods ≤ 15 months, Tan et al . followed LV functional parameters, including strain and strain rate, for a median of 2 years. Thus, we are able to get a relatively rare glimpse at the durability of strain alterations caused by chemotherapy. The investigators found that longitudinal strain declined after treatment in more than one half of their patients, and although it recovered at follow-up in 44% of these patients, it did not recover completely in the majority. With regard to strain rate, the percentage of patients with a decline was even greater, while recovery was less apparent.

This study brings up some important and as yet unresolved issues. For instance, what is the prognostic significance of a mild decline in LV EF or an increase in LV volume that appears to remain within the normal range or is at most mildly abnormal, as seen in this study, in young patients at otherwise low cardiovascular risk? Likewise, the decline in peak systolic longitudinal strain, although significant when compared with baseline measurements, remained within the normal reference range quoted by the authors. According to the recently published expert consensus document on the imaging evaluation of patients after cancer therapy, a decline in GLS of >15% from baseline is clinically meaningful, while a decline of <8% is not meaningful. In this study seven of the 16 patients, in whom longitudinal strain was persistently low after chemotherapy, had a recovery in peak systolic longitudinal strain of ≥10% relative to baseline. Would the initiation of standard heart failure therapy upon detecting a reduction in strain alter the long-term cardiovascular outcomes for such patients?

Finally, there is the concern that radiation therapy concomitant with anthracycline-containing chemotherapy may further increase the risk for cardiotoxicity. Forty-one percent of patients in the study by Tan et al . received radiation to the left chest. The authors acknowledge radiation therapy as a potential confounder in their study; however, because of the small study population, it would be difficult to estimate the presence or magnitude of the effect. The timeline for radiation-related cardiac damage can be insidious, and therefore the temporal relationship of any radiation effect to the observed functional abnormalities in this study cannot be ascertained.

The contributions of Tan et al . and other investigators in this evolving discipline have made it increasingly apparent that larger multicenter prospective cardiovascular outcome studies are needed to define the long-term prognostic significance of these and other emerging parameters of potential relevance. Such studies will need to be of adequate size and have follow-up periods sufficiently long to answer these questions, while also examining the modifying effects of radiation and conventional cardiovascular risk factors. Shared data from registries associated with a growing number of cardio-oncology programs may lend additional insight. Ultimately, the goal should be to develop evidence-based clinical guidelines to ensure the safest delivery of cancer therapy and aid in long-term survivorship care. We must applaud Tan et al . for methodically attempting to fill in knowledge gaps and highlight areas for future study in a field in which data to develop practice standards are sorely needed.