Journal of the American College of Cardiology
Use of Myocardial Strain Imaging by Echocardiography for the Early Detection of Cardiotoxicity in Patients During and After Cancer ChemotherapyA Systematic Review
Author + information
- Paaladinesh Thavendiranathan, MD∗,†∗∗ (dinesh.thavendiranathan{at}uhn.ca),
- Frédéric Poulin, MD∗,
- Ki-Dong Lim, MD∗,
- Juan Carlos Plana, MD‡,
- Anna Woo, MD∗ and
- Thomas H. Marwick, MD§∗ (tom.marwick{at}utas.edu.au)
- ∗Division of Cardiology, Peter Munk Cardiac Center, Toronto General Hospital, University Health Network, University of Toronto, Toronto, Ontario, Canada
- †Cardiac Conditions in Oncology Program, Mount Sinai Hospital, University of Toronto, Toronto, Ontario, Canada
- ‡Cardio-Oncology Center, Section of Cardiovascular Imaging, Department of Cardiovascular Medicine, Cleveland Clinic, Cleveland, Ohio
- §Menzies Research Institute Tasmania, Hobart, Australia
- ↵∗Reprint requests and correspondence:
Dr. Thomas H. Marwick, Menzies Research Institute of Tasmania, 17 Liverpool Street, Hobart T7000, Australia. - ↵∗∗Dr. Paaladinesh Thavendiranathan, 4N-490 Toronto General Hospital, Peter Munk Cardiac Center, 585 University Avenue, Toronto, Ontario M5G 2N2, Canada.
Abstract
The literature exploring the utility of advanced echocardiographic techniques (such as deformation imaging) in the diagnosis and prognostication of patients receiving potentially cardiotoxic cancer therapy has involved relatively small trials in the research setting. In this systematic review of the current literature, we describe echocardiographic myocardial deformation parameters in 1,504 patients during or after cancer chemotherapy for 3 clinically-relevant scenarios. The systematic review was performed following the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines using the EMBASE (1974 to November 2013) and MEDLINE (1946 to November 2013) databases. All studies of early myocardial changes with chemotherapy demonstrate that alterations of myocardial deformation precede significant change in left ventricular ejection fraction (LVEF). Using tissue Doppler-based strain imaging, peak systolic longitudinal strain rate has most consistently detected early myocardial changes during therapy, whereas with speckle tracking echocardiography (STE), peak systolic global longitudinal strain (GLS) appears to be the best measure. A 10% to 15% early reduction in GLS by STE during therapy appears to be the most useful parameter for the prediction of cardiotoxicity, defined as a drop in LVEF or heart failure. In late survivors of cancer, measures of global radial and circumferential strain are consistently abnormal, even in the context of normal LVEF, but their clinical value in predicting subsequent ventricular dysfunction or heart failure has not been explored. Thus, this systematic review confirms the value of echocardiographic myocardial deformation parameters for the early detection of myocardial changes and prediction of cardiotoxicity in patients receiving cancer therapy.
The mortality rate among patients with cancer has decreased over the past 20 to 30 years (1,2). However, cardiac toxicity (cardiotoxicity) from cancer therapy has become a leading cause of morbidity and mortality in survivors (3,4). In patients who develop heart failure (HF) from cancer therapy, the mortality rate is as high as 60% by 2 years (5). Therefore, contemporary management of patients with cancer should include careful consideration of potential cardiotoxicity during therapy, with a focus on early detection and intervention (6).
Historically, several definitions of cardiotoxicity have been proposed (7). The most commonly used definition is a ≥5% reduction in symptomatic patients (or ≥10% reduction in asymptomatic patients) in the left ventricular ejection fraction (LVEF) from baseline to an LVEF <55% (8). Early detection of cardiotoxicity has predominantly relied upon serial cardiac imaging to identify a reduction in left ventricular (LV) function without signs or symptoms of heart failure (stage B HF) (9). The use of LVEF has important limitations. First, the measurement of LVEF is subject to technique-related variability, which can be higher than the thresholds used to define cardiotoxicity (8,10). Second, the reduction in LVEF is often a late phenomenon, with failure to recover systolic function in up to 58% of patients despite intervention (11–15). Hence, there has been a growing interest in markers of early myocardial changes (i.e., changes with normal LVEF) that may predict the development of subsequent LVEF reduction or the progression to HF, so that preventive strategies with established cardioprotective medications such as beta-blockers, angiotensin-converting enzyme inhibitors, or dexrazoxane could be implemented.
Myocardial deformation can now be readily measured during routine echocardiography, and its value in detecting subclinical ventricular dysfunction, as well as its prognostic value, has been demonstrated in several clinical scenarios (16). A growing body of literature supports the use of myocardial deformation parameters to detect early myocardial injury and to forecast ventricular dysfunction (cardiotoxicity) in patients receiving cancer therapy. This systematic review seeks to summarize the existing data for the following clinically relevant scenarios: 1) detection of early myocardial changes; 2) prediction of subsequent cardiotoxicity; and 3) detection of late consequences of therapy (>1 year post-treatment).
Methods
Search strategy
The search method adhered to the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement for reporting systematic reviews (17). An EMBASE (1974 to November 7, 2013) and MEDLINE (1946 to November 7, 2013) search was performed by an experienced information specialist using the terms “antineoplastic agents,” “radiotherapy,” “cardiac toxicity,” “echocardiography,” and their variations as key words in the OVID search engine without language or species limitations (Fig. 1). References of all selected papers and reviews were screened to identify additional studies.
Literature Search Flow Diagram
Inclusion and exclusion criteria
Any prospective or retrospective study of at least 10 patients that used echocardiographic (echo)-based myocardial deformation parameters as the primary method to detect cardiotoxicity during or after cancer therapy was included. In order to be included in this systematic review, studies had to provide data on changes in deformation parameters and LVEF during therapy. Studies that did not provide data on the type of chemotherapy or the timing of imaging were excluded.
Myocardial deformation
Echocardiographic measures of LV strain have become a robust method to measure myocardial deformation (16,18). Strain is a dimensionless index reflecting the total deformation of the ventricular myocardium during a cardiac cycle as a percentage of its initial length (reported as percentage). Strain rate (SR) is the rate of deformation or stretch (reported as s−1) (18). Both strain and SR can be measured in the longitudinal, radial, and circumferential directions (Fig. 2) (16,18). A key advantage of strain or SR measurement is its ability to differentiate active versus passive movement within a myocardial segment, allowing for the analysis of regional myocardial deformation independent of the translational motion of the heart. Although neither LV strain nor SR are load independent, peak systolic SR correlates well to load-independent indexes of contractility and, hence, provides valuable information about intrinsic contractile function (18,19). LV torsion is a measure of the maximum instantaneous difference in the rotation of the base of the heart in comparison to the apex (20). This is then followed by untwisting contributing to ventricular filling. Peak systolic twisting velocity measures the peak positive rate of torsional deformation during the ejection phase, whereas peak diastolic untwisting velocity measures the peak negative rate of torsional deformation during early diastole (18,20,21). Currently myocardial deformation can be measured using tissue Doppler imaging (TDI) (Fig. 3) and 2- and 3-dimensional speckle tracking echocardiography (STE) (Fig. 2) (18).
Speckle Tracking Echocardiography-Based Peak Systolic Strain Measurements in a Patient With Breast Cancer Prior to Initiation of Cytotoxic Chemotherapy
(A) Global longitudinal strain (GLS), (B) global radial strain (GRS), and (C) global circumferential strain (GCS). The left panels show the direction (arrows demonstrate the direction) in which various strain parameters are being measured. The middle panels demonstrate the segmental strain values (except for circumferential strain). The right panels illustrate the regional strain curves. Circumferential strain curves in the bottom right panel highlight the segmental variability in measurements, illustrating the challenges with this specific strain measurement. Reported normal value for GLS is −19.7% (95% confidence interval [CI]: −20.4% to −18.9%), for GRS 47.3% (95% CI: 43.6% to 51.0%), and GCS −23.3% (95% CI: −24.6% to −22.1%) (68). AVC = aortic valve closure.
Tissue Doppler-Based Strain Assessment
(Left) Tissue Doppler image of the interventricular septum and the region of interest (yellow oval) from where the strain and strain rate values were obtained. (Middle) Longitudinal strain rate curve; and (right) longitudinal strain of the basal interventricular septum. The arrows illustrate the points on the curve where the peak systolic longitudinal strain rate and longitudinal strain would be recorded.
Outcomes
Outcomes of interest were absolute and percentage reductions in myocardial deformation parameters during or after therapy and performance of these parameters in predicting subsequent cardiotoxicity (defined in the preceding text).
Data extraction
All relevant data were extracted using a standard data form by one reviewer (P.T.) and verified by a second reviewer (F.P.). All discrepancies were mutually reviewed and resolved by consensus. Multiple authors were contacted for clarification of data in various publications. We excluded 2 studies without data on timing of echocardiography, as we were unable to make contact with the authors for this additional data (22,23). The following data were extracted: year of publication, number of patients, cancer type, age, sex, chemotherapy used and doses, type and timing of imaging, changes in deformation parameters, and prognostic data.
Results
Detection of early myocardial changes during cancer chemotherapy
Thirteen peer-reviewed publications, involving approximately 384 patients treated with anthracycline-containing regimens, assessed various echo-based myocardial deformation parameters to detect early myocardial changes without providing data on prognosis (24–36). These were single-center cohort studies that primarily focused on breast and hematological malignancies. The mean age ranged from 49 to 70 years (56% to 100% female) in the adult studies, and from 9 to 15 years (23% to 48% female) in the pediatric studies. Earlier work used TDI-based strain, whereas the more contemporary studies have generally used 2-dimensional (2D) STE (Table 1). Despite heterogeneity in the data with respect to patient age, types of cancer, strain techniques, and timing of follow-up, the studies all uniformly demonstrate that changes in myocardial deformation occur earlier than a change in LVEF and at anthracycline doses lower than what was historically thought to be cardiotoxic (e.g., 200 mg/m2 of epirubicin). The degree of change in myocardial deformation parameters amongst the studies has depended on the technique used (2D STE vs TDI) and the type of strain measured.
Summary of Studies That Have Used Advanced Myocardial Mechanics to Illustrate Early Myocardial Injury During Cancer Chemotherapy
2D-based strain
In the absence of a reduction in LVEF, a 2D STE–measured reduction in peak systolic global longitudinal strain (GLS) between 9% and 19% seems to be common either during or immediately after anthracycline therapy (Table 1). Although a reduction in peak systolic global radial strain (GRS) of 6% to 17% (34,37–40) or peak systolic global circumferential strain (GCS) of 11% to 16.7% (38,40,41) may also indicate early myocardial changes, these changes have been less consistent (30,34,41,42). An important limitation of both GRS and GCS is the lower reproducibility of these measurements, which makes the identification of changes from pre- to post-chemotherapy more challenging. Similarly, SR measurements using STE have important technical limitations. Although rotational myocardial deformation and early diastolic SR are potential markers of early myocardial changes (30,33,42), neither of these parameters are currently sufficiently feasible and reliable for routine clinical application.
TDI-based strain
When using TDI-based strain, longitudinal SR of the basal interventricular septum consistently demonstrates a reduction (ranging from 9% to 20%) between pre-therapy and low doses of anthracyclines (e.g., 200 mg/m2 of epirubicin). In contrast, changes in longitudinal strain (LS) have not been a reliable measure of early injury, especially when measurements are obtained only from the basal interventricular septum (25,28,29). However, when multiple septal segments or all 18 myocardial segments are used, reductions in LS of 15% and 17% were seen after first dose of anthracycline (26) and 6 cycles of liposomal doxorubicin (27), respectively. Radial strain parameters are known to be variable. However, a fall in radial SR of 13% to 28% (26,27,35,43) or radial strain of 24% to 35% of the mid inferior-lateral walls also seems to detect early myocardial changes, although the latter has not been consistent (43).
Changes in strain values appear to be regional, although the segmental variation has been inconsistent among studies (31,34). Unfortunately, although biopsy changes have been documented with early injury, the clinical application of biopsy is neither relevant nor feasible, so whether the changes in myocardial deformation truly represent cardiac injury cannot be proven. However, several studies have shown a positive association between higher doses of anthracycline (30–32) or serum markers such as reactive oxygen species levels and troponins (25,28,29,36,40,41) and larger reductions in strain or SR measurements, suggesting that there is biological plausibility for these findings.
Prognostic value of myocardial deformation parameters to detect cardiotoxicity
Although the early detection of myocardial changes appears to be conceptually important, the real value of these changes lie in their ability to prognosticate clinically-relevant outcomes such as subsequent LVEF reduction or the development of HF. The prognostic value has been evaluated in 8 studies (Table 2) involving approximately 452 patients (age range from 47 to 51 years, 58% to 100% women) (37–44). Published studies have either been single-center (37–39,42,44) or multicenter (40,41) cohort studies and, other than the 3 recent studies (37–39), all have only included patients with breast cancer. Most (40–44) have included patients with human epidermal receptor 2 overexpressing breast cancers, with all patients receiving trastuzumab and the majority receiving anthracyclines. However, important differences between studies (Table 2) include differences in duration of follow-up (6 months vs. 12 to 15 months), treatment regimens (proportion receiving anthracycline and radiotherapy, cumulative epirubicin dose, and use of taxanes), the definition of the “baseline” echo (pre- vs. post-anthracyclines), and the number of apical views used to measure strain (all 3 views versus the basal and mid segments of just 2 views). The definition of cardiotoxicity was, however, similar between the studies and the incidence of cardiotoxicity ranged between 13% and 32%, likely relating to differences in baseline cardiac risk factors, treatment regimens, and duration of follow-up.
Summary of Studies That Have Used Early Changes in Advanced Myocardial Mechanics to Predict Subsequent Cardiotoxicity
An early fall in GLS by STE between 10% and 15% predicts subsequent cardiotoxicity (including both asymptomatic and symptomatic LV dysfunction) (37,39–42,44) (Fig. 4, Online Videos 1, 2, and 3). The 95% confidence interval for the optimal GLS cutoff extends from 8.3% to 14.6% (42). The reported sensitivity and specificity of GLS to predict cardiotoxicity (Table 3) is likely optimistic, given the small sample sizes and few cardiotoxicity events. In patients where a relative change in GLS was unavailable, absolute levels of GLS >−19% and −20.5% early during therapy have been associated with cardiotoxicity (40,42). In contrast, GRS was not predictive of cardiotoxicity in the 2 larger studies (40,42), whereas GCS was not predictive in any studies. However, a combined parameter of GLS and LV twist (GLS × LV twist) appears to be the best predictor of subsequent cardiotoxicity, with test characteristics superior to even GLS (Table 3) (39). This latter parameter provides a combined assessment of LV subendocardial function (GLS) and subepicardial function (LV twist), potentially providing a more sensitive measure of early myocardial changes, although this needs confirmation in other studies. A summary of myocardial strain and SR cutoff values to predict cardiotoxicity from the preceding studies is provided in Table 3.
The Utility of Early Strain Changes to Predict Subsequent Cardiotoxicity
The images demonstrate a “bull’s eye” plot of strain values for each of the 17 myocardial segments. A patient receiving cytotoxic chemotherapy had normal baseline strain and left ventricular (LV) ejection fraction (EF) (left). Six months into therapy, the LVEF dropped by 6% but did not meet criteria for cardiotoxicity. However, the peak systolic global longitudinal strain (GLS) fell by 15.4% (a significant change based on the literature). Then, by 12 months there was a clinically significant fall in LVEF meeting the criteria for cardiotoxicity. See Online Videos 1, 2, and 3 for 4-chamber movie images demonstrating the changes in function. LVEF was calculated using the Biplane Simpson’s method. 6M = 6 months; 12M = 12 months.
Four-chamber movie images at baseline
Four-chamber movie images at baseline, showing normal LV global systolic function by 2D echo (LVEF were calculated using the Biplane Simpson’s method).
Four-chamber movie images at 6 months
Four-chamber movie images at 6 months, showing normal LV global systolic function by 2D echo (despite impaired strain).
Four-chamber movie images at 12 months
Four-chamber movie images at 12 months, showing reduced LV global systolic function by 2D echo.
Early Predictors of Cardiotoxicity
Detection of late subclinical consequences of cancer therapy
After chemotherapy regimens are completed, there are limited recommendations as to appropriate follow-up (Online Table A). However, specifically with anthracyclines, cardiotoxicity can be first detected several years after therapy (45,46). Hence, there has been a growing interest in detecting subclinical cardiotoxicity in survivors using myocardial deformation parameters with the hope of identifying high-risk patients and providing targeted therapy with cardioprotective medications to ultimately prevent further LV remodeling and progression to HF syndrome.
There are 9 published case-control studies that have used various myocardial deformation parameters to detect late subclinical cardiac injury, consisting of approximately 436 patients (median age 12.7 years, 30% to 100% women) (21,47–54), but none have provided data on prediction of subsequent cardiac events (Table 4). The only study in adult breast cancer survivors (49) showed a 7.7% reduction in GLS in patients compared with controls when imaged between 3.1 and 4.2 years post-therapy, with lower GLS values with adjuvant trastuzumab use. All other studies have been in survivors of various pediatric cancers treated with anthracyclines (21,47,48,50–54). The time between completion of therapy to cardiac imaging ranged from 8 months to 29.2 years. All studies have compared findings in patients with controls, with none comparing identified abnormalities to pre-therapy imaging. Therefore, it is unknown whether some of these patients had pre-therapy ventricular dysfunction. Two studies using TDI-based strain (48,52) have demonstrated a reduction in LS and longitudinal SR of the interventricular septum, LV lateral wall, and right ventricular free wall. Despite a difference in cumulative anthracycline doses between the studies (<300 mg/m2 vs. >350 mg/m2), both illustrated reductions in strain values, emphasizing that myocardial injury can occur at lower anthracycline doses as well. This variability in doses may also explain variations in the incidence of LV dysfunction of between 5% and 16%. Both studies have shown that anthracyclines can also affect right ventricular function, a concept that has not been adequately explored.
Summary of Studies That Have Used Advanced Myocardial Mechanics Parameters to Demonstrate Subclinical Myocardial Injury in Patients Who Previously Received Cancer Chemotherapy
The remaining 6 pediatric studies have used STE-based strain, but with significant heterogeneity with respect to the types of cancers, time of imaging, cumulative anthracycline dose, and the type of strain measurements. However, in anthracycline-treated survivors, a reduction was reported in most strain and SR parameters, ranging from 6.6% to 29.6% compared with controls (21,47,50,51,53,54). Possible reasons for this variation could include differences in follow-up duration, maximal dose of anthracyclines, and radiotherapy, and inclusion of patients with overt LV systolic dysfunction. There appears to be a discrepancy amongst studies with respect to the value of longitudinal deformation parameters in survivors, although radial and circumferential strain appear to be consistently abnormal. Similar to studies during therapy, the change in mechanics is regional, with the interventricular septum being the most consistently affected (47,51,53).
Rotational deformation parameters have also been assessed in survivors by the same group in 3 publications (21,53,54). Although there were differences in types of cancers, all of the included patients received similar anthracycline doses and were imaged at similar time points post-therapy. At a segmental level, the apical rather than basal rotational deformation appears to be consistently affected. Furthermore, a reduction in left ventricular peak torsion has been described (21). In layer-specific strain analysis, the changes in rotational parameters seem to vary across myocardial layers (53). With 3-dimensional (3D) echocardiography, global 3D systolic strain, twist, and torsion are reported to be reduced compared with controls (54).
Detection of myocardial injury from radiotherapy
There is limited literature on the detection of early myocardial changes from radiotherapy (RT) (Table 5), with data on approximately 232 patients (age 48 to 51 years, 40% to 100% of women). Two studies (55,56) in patients with breast cancer illustrated a relative fall in GLS of 9.8% to 10.2% and GLSR of 12.8% immediately after RT when compared with pre-therapy using TDI-based strain. The mean LV specific dose in these 2 studies ranged from 6.7 to 9.0 Gy. The strain drop was only seen in women with left-sided breast cancer (and not in those with right-sided cancer) and was only limited to the anterior LV myocardial segments, which received the highest radiation doses. Patients in both studies also received anthracycline and some received trastuzumab, making it difficult to differentiate the effect of RT from chemotherapy. This is important as the effects of RT and chemotherapy are likely additive (57). In patients with Hodgkin’s lymphoma treated with RT with or without doxorubicin 22 years previously, the reduction in STE-based GLS was highest in patients who had RT with doxorubicin (21%) and less in those who only had RT (14%), compared with controls. In 2 older studies (58,59), in patients with various cancers, a reduction in longitudinal systolic and diastolic strain was only present after 50 Gy of thoracic RT in patients not exposed to chemotherapy. However, the impact of radiation on measures of myocardial deformation has not been consistent with 3 other studies, which focused primarily on the toxicity of chemotherapy, not identifying an interaction between radiotherapy and strain (32,43,49). However, none of these latter studies provided data on radiation dose or the side of radiotherapy, both of which are important in the development of cardiac injury.
Summary of Studies That Used Advanced Myocardial Mechanics to Detect Early Myocardial Injury From Radiation Therapy
Discussion
There are several key messages in this review. Reductions in echocardiographic measures of myocardial deformation parameters are a sign of subclinical myocardial changes from cancer therapy and occur prior to any change in LVEF as assessed by conventional 2D echocardiography. Importantly, early reduction in myocardial deformation appears to forecast the development of subsequent cardiotoxicity, with STE measured GLS being the most consistent parameter. The thresholds of change in GLS to predict cardiotoxicity have ranged from 10% to 15% using STE. These thresholds generally have better negative predictive value than positive predictive value, probably reflecting the low prevalence of cardiotoxicity in the patients studied. Unfortunately, in survivors, although deformation parameters appear to detect subclinical myocardial changes, the value of these changes in predicting subsequent LV dysfunction or heart failure is unknown. Finally, RT also affects myocardial deformation, with changes occurring predominantly in those receiving therapy to the left chest and to myocardial segments receiving the highest radiation doses.
Cardiovascular complications of cancer therapy
Many of the chemotherapeutic agents in use today can have associated cardiovascular side effects, the most common of which are cardiomyopathy and HF (45,60). Amongst the various medications, the anthracycline class of drugs (e.g., doxorubicin and epirubicin) and the human epidermal growth factor receptor type 2 (HER 2) monoclonal antibody, trastuzumab, have been most commonly implicated and best studied. A recent meta-analysis of 55 published randomized controlled trials showed that the use of anthracycline-based versus nonanthracycline-based regimens were associated with a significantly increased risk of both clinical (odds ratio: 5.43) and subclinical (odds ratio: 6.25) cardiotoxicity (61). Despite this toxicity, anthracyclines remain the cornerstone of treatment in many malignancies, including lymphomas, leukemias, and sarcomas, and are still widely used in both advanced and early-stage breast cancer (60). Combined therapy generally increases the incidence of cardiotoxicity (62). This has been best demonstrated in women with HER 2–positive breast cancers treated with anthracycline followed by trastuzumab, in whom the incidence of cardiotoxicity has been reported to be as high as 41.9% in older women during long-term follow-up (46). Two types of cardiomyopathy have been defined to distinguish anthracycline-induced myocardial damage (type I) from trastuzumab-induced myocardial dysfunction (type II). Type I cardiomyopathy is related to the cumulative dose, is largely irreversible, and results from free radical formation and mitochondrial dysfunction ultimately leading to myofibrillar disarray and necrosis (63). In contrast, type II cardiomyopathy is not dose-related, may be reversible, and results in no apparent ultrastructural changes (63).
Detection of cardiotoxicity
The current recommendations for pre-treatment cardiac evaluation and monitoring of patients receiving cancer therapy are not specific and vary among the different guidelines by cardiovascular and oncology societies. A summary of the core recommendations from the major guidelines is summarized in Online Table A (60,64–67). The European Society for Medical Oncology has provided the most comprehensive recommendations for monitoring during and after chemotherapy, based on clinical risk factors and cumulative dose (64). The American Society of Echocardiography, in collaboration with the European Association of Cardiovascular Imaging have created an Expert Consensus Document on the evaluation of adult patients during and after cancer therapy that will soon be published. Although several imaging modalities such as cardiac magnetic resonance imaging or multigated acquisition scans can be employed in the evaluation of cardiotoxicity, the benefit of echocardiography comes from its versatility, lower cost, ability to assess more than ventricular function, and avoidance of repeated radiation exposure.
Diastolic function has also been explored as a marker of early cardiotoxicity in several studies, but the best diastolic parameter to follow is not clear. Also, no echocardiography studies to date have demonstrated that an early subclinical drop in LVEF or changes in diastolic parameters can predict subsequent cardiotoxicity. Furthermore, although controversial, several studies show that systolic strain changes occur prior to or in the absence of changes in traditional diastolic parameters (21,28,29,41,47,49,56,57). The strength of echo-measured myocardial deformation parameters therefore includes the ability to more readily detect regional abnormalities in LV function along with improved measurement reproducibility due to the semiautomated nature of the measurements, and the ability to forecast subsequent LV dysfunction. The reproducibility data for strain measurements presented in each study are summarized in Tables 1, 2, 4, and 5.
Tissue Doppler and STE-based strain have been used to detect early myocardial changes in patients receiving chemotherapy. With TDI, interventricular septal longitudinal SR appears to be most consistently reduced during therapy. However, the most clinically relevant data on predicting cardiotoxicity have been based on STE-based strain. Also, TDI-based strain analysis requires data acquisition for each myocardial segment with careful attention to frame rates as well as alignment of the walls with the Doppler beam (18). The measurements of strain and SR can be noisy, and significant expertise is required for proper interpretation (18). This makes clinical application more challenging as compared with STE-based strain, which can be obtained at lower frame rates using standard 2D images and has a more streamlined post-processing (18). In addition to easily measuring GLS from all 18 myocardial segments, STE allows measurement of radial and circumferential strain in multiple segments, as well as rotational parameters. Also, the reproducibility of STE-based strain analysis is superior to TDI-based analysis (18).
Normal ranges for GLS defined in a recent meta-analysis (mean GLS −19.7%; 95% confidence interval: −20.4% to −18.9%) (68) underpin the use of a normal cutoff exceeding −19%. However, because of baseline variability in strain values between different patients, within-patient change may be a more reliable parameter compared with a population-derived absolute cut-off value. The threshold for change in GLS to predict cardiotoxicity is not clear, although between 10% and 15% appears to have the best specificity. The observer variability of the GLS measurements based on the summarized studies is within the suggested threshold to predict cardiotoxicity.
Future directions
Much remains to be understood about the role of cardiovascular imaging in the identification and management of cardiotoxicity from cancer chemotherapy. Whether strain-based approaches could be reliably implemented in multiple centers, including nonacademic settings, needs to be studied. The ability of strain changes to predict subsequent cardiotoxicity needs to be examined in larger multicenter studies and in cancers other than breast cancer, where treatment with potentially cardiotoxic regimens is provided. Whether strain measurements are required at multiple time-points or a single selected time-point has to be determined. An approach that uses strain as the primary marker of cardiotoxicity to initiate cardioprotective therapy needs to be compared with a traditional LVEF-based approach. The long-term effect of strain changes that occur during therapy needs to be understood. The use of vendor-neutral methods to measure strain and their ability to predict cardiotoxicity also need to be explored for this technique to be more widely applied. Finally, the prognostic significance of strain abnormalities in survivors of cancer and those receiving radiation therapy has to be understood along with whether intervention would change the natural course of the cardiac disease.
Acknowledgments
The authors would like to thank Drs. Shizhen Liu and Juan Duero Posada for translating the Chinese and Spanish papers, respectively, that are included in this review; Melanie Anderson (an information specialist) for performing the literature search; and Dr. Tomoko Negishi for helping with Figure 4.
Appendix
Footnotes
Dr. Marwick has received a research grant from General Electric. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- 2D
- 2-dimensional
- 3D
- 3-dimensional
- echo
- echocardiography
- GCS
- global circumferential strain
- GLS
- global longitudinal strain
- GRS
- global radial strain
- HF
- heart failure
- LV
- left ventricular
- LVEF
- left ventricular ejection fraction
- RT
- radiotherapy
- SR
- strain rate
- STE
- speckle tracking echocardiography
- TDI
- tissue Doppler imaging
- Received December 19, 2013.
- Revision received January 24, 2014.
- Accepted January 28, 2014.
- American College of Cardiology Foundation
