Novel imaging paradigms for characterization of cardiovascular toxicities associated with cancer therapy

Abstract

Since the 1970s, there have been many paradigm shifts in the detection of cardiotoxicity associated with cancer treatment. These have included movement away from late sequelae of cancer treatment such as heart failure or decline in ejection fraction toward earlier detection of subclinical toxicity using echocardiographically derived global longitudinal strain and high sensitivity troponin assays. In addition to this, cardiac MRI has had an expanding role in detecting cardiotoxicity due to its reproducibility in left ventricular ejection fraction measurements and tissue characterization. This article summarizes the strengths and weaknesses of these imaging techniques in characterizing cardiotoxicity in patients receiving cancer therapy.

To some extent, the origin of cardio-oncology dates back to the late 1970s when oncologist Von Hoff noted the association between the cumulative dose of doxorubicin received and the probability of congestive heart failure [1]. More than two decades later, a similar association was observed when both left ventricular (LV) dysfunction and congestive heart failure were considered outcome events, even at doses of doxorubicin previously felt to carry an acceptable risk of cardiotoxicity [2]. One of the first paradigm shifts in detecting cardiotoxicity was the transition from surveillance for heart failure to ejection fraction (EF). Along with that was a paradigm shift in the imaging modalities used to detect LV dysfunction from multigated acquisition scan (MUGA) to 2D echocardiography. CMR started to emerge into mainstream clinical use in the mid-late 1990s and quickly became the gold standard technique for LV volumes and EF. It suffered from its lack of portability and widespread availability. On that backdrop, 3D echocardiography, which compares well to CMR for reproducible LV volumes and EF, evolved to be incorporated into echocardiographic platforms and is now widely available. More recently, there has been yet another paradigm shift toward characterizing subclinical cardiotoxicity. For this, biomarkers, myocardial strain imaging and tissue characterization with CMR have emerged as useful tools.

Global longitudinal strain (GLS) as a measure of subclinical LV dysfunction was proposed as a component of the imaging framework for early detection of cardiotoxicity back in 2014 with the caveat that an interval change in GLS may indicate subclinical LV dysfunction but that there was insufficient data to support the initiation of medical therapy based on this parameter alone [3]. The SUCCOUR trial recently demonstrated that regardless of whether one used a reduction in GLS or LVEF as a trigger to initiate cardioprotective therapy, there was no significant difference between groups in LVEF at 1-year follow-up [4]. Thus, the data still seems insufficient to support the routine implementation of cardioprotective therapy. However, GLS remains an important prognostic indicator. For example, in patients with checkpoint inhibitor myocarditis, in whom 50 % or more may have a normal EF, GLS was reduced in myocarditis cases with both reduced and preserved EF compared to controls and was predictive of major adverse cardiac events [5].

Myocardial strain can also be obtained from CMR using a variety of techniques such as myocardial tagging, Displacement Encoding with Stimulated Echoes (DENSE), strain encoding MRI (SENC), or feature tracking (Fig. 1). Data from a small study of patients receiving anthracycline chemotherapy demonstrated that patients with normal strain, as measured by SENC, in less than 80 % of their myocardium were more likely to experience cardiotoxicity. In those with less than 55 % of the myocardium with normal strain, LV functional recovery was less likely [6]. Noncontrast parametric techniques such as T1 mapping can be used to further delineate the intrinsic properties of tissue by characterizing the signals from the intra- and extra-cellular compartments. Pre- and post-contrast assessment of T1 mapping can be used to determine the extracellular volume (ECV), a measure of diffuse interstitial fibrosis seen in various forms of heart failure. These techniques can be harnessed in cases of chemotherapy-related cardiac dysfunction even if it is subclinical. For instance, increased native T1 times and ECV have been noted in cancer survivors as early as 3 years following anthracycline administration independent of age, sex and co-existing cardiovascular comorbidities. Further study is needed to determine if this finding is associated with onset of heart failure or change in exercise capacity in this population [7]. Additional CMR parameters such as T2-weighted imaging and T2 mapping (which detect myocardial edema/inflammation and late gadolinium enhancement (which detects necrosis and fibrosis) can be harnessed to support the diagnosis of other forms of cardiotoxicity such as ICI myocarditis (Fig. 2). However, there is still much to be learned about the optimal timing of CMR imaging in such patients. A retrospective registry data study showed late gadolinium enhancement was present in only 48 % of ICI myocarditis which appeared to be related to the timing of the study [8]. Cardiotoxicity has historically been defined by clinical heart failure or a decline in EF. Currently, the emphasis has shifted to early detection before a decline in EF occurs. GLS has emerged as a clinically useful tool in that regard. With advancements in cancer therapy and imaging, there is now an expanding role for CMR, which offers multiplanar imaging, strain, and tissue characterization, referred to by some as “virtual histology.” However, regardless of the imaging tool used to detect cardiotoxicity, much work is needed to refine surveillance algorithms, correlate imaging findings with hard long-term outcomes such as clinical heart failure or cardiovascular death, and test preventative and interventional strategies to offset cardiotoxicity associated with cancer therapeutics.

Fig. 1.

Myocardial strain using MRI techniques. In this figure, MRI measures of global longitudinal and circumferential strain are derived using strain encoding (SENC, top) and post-processing of cine images with feature tracking (bottom) in the same 2 patients, A and B. Feature tracking quantifies myocardial deformation through an algorithm that identifies images features in the myocardium that can be reliably identified and tracked on a frame-by-frame basis throughout the cardiac cycle. SENC differs from feature tracking in that it measures strain in the direction perpendicular to the imaging plane: circumferential from the long-axis and longitudinal from the short-axis images.

MRI = magnetic resonance imaging, SENC = strain encoding, GLS = global longitudinal strain, GCS = global circumferential strain. Image modified and used without restriction under the terms of the Creative Commons Attribution 4.0 International License (https://creativecommons.org/licenses/by/4.0/). Original source: Early, J et al. [9]

Fig. 2.

Use of cardiac MRI in cardio-oncology. Some cardiac MRI parameters such as T2 Stir and late gadolinium enhancement, could be harnessed to detect inflammation, edema and necrosis or fibrosis often associated with the diagnosis of other forms of cardiotoxicity such as immune-checkpoint inhibitor myocarditis. These images, obtained in a patient treated with nivolumab, demonstrate increased signal on T2 Stir (panel A, black asterisk) and late gadolinium enhancement (panel B, white arrow) consistent with immune checkpoint myocarditis.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.