Prognostic Impact of Global Longitudinal Strain and NT-proBNP on Early Development of Cardiotoxicity in Breast Cancer Patients Treated with Anthracycline-Based Chemotherapy

Background. The most important anthracycline side effect is cardiotoxicity, resulting in congestive heart failure (HF). Early detection of cardiac dysfunction and appropriate treatment can improve outcomes and reduce the progression of HF. The aim of our study was to evaluate changes in clinical data, echocardiographic parameters, and NT-proBNP, as well as their associations with early anthracycline-induced cardiotoxicity (AIC) in patients treated with anthracycline-based chemotherapy. Methods and Materials. Patients with breast cancer were prospectively assessed with echocardiography, as well as NT-proBNP testing at baseline, (T0), after two cycles (T1) and four cycles (T2) of chemotherapy. AIC was defined as a new decrease in the LVEF of 10 percentage points, to a value below the lower limit of normal. Results. We evaluated 85 patients aged 54.5 ± 9.3 years. After a cumulative dose of 237.9 mg/m2 of doxorubicin, 22 patients (25.9%) met the criteria of AIC after chemotherapy. Patients who subsequently progressed to cardiotoxicity had demonstrated a significantly larger impairment in LV systolic function compared to those who did not develop cardiotoxicity (LVEF: 54.0 ± 1.6% vs. 57.1 ± 1.4% at T1, p < 0.001, and 49.9 ± 2.1% vs. 55.8 ± 1.6% at T2, p < 0.001; GLS: −17.8 ± 0.4% vs. −19.3 ± 0.9% at T1, p < 0.001, and −16.5 ± 11.1% vs. −18.5 ± 0.9% at T2, p < 0.001, respectively). The levels of NT-proBNP increased significantly from 94.8 ± 43.8 ng/L to 154.1 ± 75.6 ng/L, p < 0.001. A relative decrease in GLS ≤ −18.0% (sensitivity: 72.73%; specificity: 92.06%; AUC, 0.94; p < 0.001) and a relative increase in NT-proBNP > 125 ng/L (sensitivity: 90.0%; specificity: 56.9%; AUC, 0.78; p < 0.001) from baseline to T1 predicted subsequent LV cardiotoxicity at T2. Conclusions. Decrease in GLS and elevation in NT-proBNP were significantly associated with AIC, and these could potentially be used to predict subsequent declines in LVEF with anthracycline-based chemotherapy.


Introduction
Breast cancer is one of the most commonly diagnosed cancers in women, and the incidence of cancer is increasing every year, with more than 1.5 million women worldwide developing breast cancer every year [1]. Since the 1990s, cancer-related mortality has steadily declined [2]. Anthracyclines have been in use since the 1960s. They are effective and widely used in breast cancer treatment, but they have many side effects [3]. The most important anthracycline adverse effect is cardiotoxicity, resulting in congestive heart failure (HF) [4]. Anthracycline-induced HF was first described in 1968 [5]. It is known that anthracycline-induced cardiotoxicity (AIC) may present with symptomatic or asymptomatic cardiotoxicity [6]. Early detection of cardiac dysfunction and appropriate treatment of these patients can improve outcomes and reduce the progression of HF.
Echocardiography is one of the most effective tests to assess cardiac function in cancer patients [6]. It is recommended to evaluate baseline left ventricular ejection fraction (LVEF) in all patients before cardiotoxic cancer treatment initiation and at completion of therapy [6]. A decrease in the LVEF of 10 percentage points, to a value below the lower limit of normal, is defined as cancer therapeutic-related cardiac dysfunction [6,7]. Furthermore, a relative percentage decrease in global longitudinal strain (GLS) of 15% from baseline is considered abnormal and a marker of early LV subclinical dysfunction. Consensus strongly supports a GLS-based follow-up of patients during and after cancer therapy [6,7].
The literature on the use of biomarkers (B-type natriuretic peptide (BNP)/N-terminal pro-brain natriuretic peptide (NT-proBNP)) for AIC risk stratification during cancer therapy still remains controversial. NT-proBNP is one of biomarkers for cardiovascular disease (CVD) risk stratification [8,9].
The purpose of our study was to evaluate the changes in clinical data, echocardiographic parameters, and NT-proBNP, as well as their associations with early cancer therapy-related LV dysfunction in breast cancer patients treated with anthracycline-based chemotherapy.

Study Population
The prospective study was performed during the period 2016-2018 at the Cardiology Department and Oncology and Hematology Department at the Hospital of the Lithuanian University of Health Sciences Kauno Klinikos. A total of 105 patients with breast cancer, who were treated with doxorubicin-based chemotherapy, were screened. Twenty patients were not included due to the exclusion criteria or later declined to participate. The final analytic cohort included 85 patients. A flowchart of patient selection is shown in Figure 1.  In all regimens, doxorubicin was given by 1 h intravenous infusion. In the beginning, trastuzumab was administered with taxanes, and then, it was given alone after doxorubicin standard treatment in an adjuvant setting for a period of 52 weeks. The present study did not show that the use of trastuzumab had statistically significant effect on LVEF and/or GLS decrease.
Two-dimensional echocardiography was performed according to the current recommendations [11]. The LVEF was assessed using Simpson's modified biplane method. LV long axis function was studied using the M-mode echocardiography and tissue Doppler imaging (TDI). The mitral annulus peak systolic velocities (S') and diastolic velocities (E') were measured in the basal segments of all six walls by TDI, and their averages (S' and E' mean) were calculated. The velocities of E-and A-waves were obtained from the transmitral pulsed-wave Doppler. Peak E-wave velocity and TDI peak early diastolic velocity ratio (E/E') were calculated. Using the M-mode approach, tricuspid annular plane systolic excursion (TAPSE) was measured. Peak tricuspid annulus systolic velocity (S' RV) by TDI was acquired, according to a standard protocol [11].
We performed an analysis of LV longitudinally, according to the recommendations in [12]. We defined AIC as a new decrease in the LVEF of 10 percentage points, to a value below the lower limit of normal (LVEF < 53%) [7].

Biochemical Markers
Study patients' blood samples (three milliliters of blood) were taken before chemotherapy (T0), after two cycles (T1), and after four cycles (T2) of doxorubicin-based chemotherapy. Samples were sent to the Department of Laboratory Medicine, at the Hospital of Lithuanian University of Health Sciences, Kauno Klinikos. Tests were performed with standard clinical practice methodology. Human NT-proBNP ELISA kits were used for the plasma NT-proBNP concentration measurements.

The Six-Minute Walking Test
The six-minute walking test (6MWT) was performed, according to the American Thoracic Society comprehensive guidelines [13], as an objective assessment of exercise capacity. The 6MWT was performed in a long straight hospital corridor, over a 30-m distance. The 6-min walk test was selected, and the distance in meters was recorded. Patients were provided with standardized instructions and encouragement. We measured S pO2 , HR responses, symptoms of dyspnea, objective fatigue, and walking distance.

Statistical Analysis
All analyses were performed using Statistical Package for the Social Sciences (SPSS), version 27.0 (Chicago, IL, USA). The categorical variables were shown as absolute numbers and percentages. The continuous variables were presented as means ± standard deviations. For continuous variables with normal distribution, Student's t test was used, and if data were not normally distributed, the Mann-Whitney rank-sum test and Wilcoxon test were used. The Chi-square (χ 2 ) or Fisher's exact test were used to compare proportions. Spearman's correlation coefficients were used if data were not normally distributed.
The ROC (Receiver Operating Characteristic) curve analysis method was used for determination of optimal values of NT-proBNP, 6MWT.
Logistic regression analyses were used to determine the association between cardiotoxicity and arterial hypertension, family history of CVS, NT-proBNP, and 6MWT via odds ratio (OR) with 95% confidence interval (CI). To estimate the adjusted odds ratio, a multivariate logistic regression analysis was performed. For all tests, the statistical significance level adopted was p ≤ 0.05.

Results
A total of 85 women with pathologically confirmed breast cancer were enrolled in this study. The study population age ranged from 33 to 75 years old (mean age: 54.5 ± 9.3 years). Patients' baseline clinical characteristics are presented in Table 1. None of the patients had symptoms or signs of heart failure during the study period. The most common CVD risk factors were AH (42.4%), family history of premature atherosclerotic CVD (27.1%), and dyslipidemia (27.1%). β-blockers and angiotensin-converting enzyme inhibitors/angiotensin receptor blockers were most commonly used in the study population. All the patients had a good exercise capacity before the treatment, and the average 6MWT was 569.6 ± 59.4 m (range 460.0-730.0). The baseline average value of NT-proBNP was normal 94.8 ± 43.8 ng/L (range 55.0-129.0) in the study population. Distribution of chemotherapy regimens is shown in Table 2. The majority of subjects (63.5% noncardioxicity group and 68.3% cardiotoxicity group) were treated with AC-paclitaxel. All regimens included similar cumulative doses of doxorubicin. Given a comparison of different regiments and medication contributing to the risks for AIC, we did not observe any differences between them. The median doxorubicin cumulative dose in cardiotoxicity group was 239.35 mg/m 2 (range 150.00-291.00), and noncardiotoxicity was 236.92 mg/m 2 (range 129.00-303.20), but this difference was not significant. The baseline 2DE and Doppler parameters (T0, pre-chemotherapy) are summarized in Table 3. All study subjects had normal LVEF (60.6 ± 1.8%, range 56-65) and GLS (−21.1 ± 0.5%, range 20.1-22.6) before treatment. Baseline M-mode and Doppler echocardiography analysis showed normal longitudinal LV systolic function, and LV diastolic function was also normal (Table 3). Table 3. Two-dimensional echocardiographic parameters before and after chemotherapy in patients with and without cardiotoxicity. LVEDD: left ventricle end-diastolic diameter; MAPSE: mitral annular plane systolic excursion; S': mitral annular peak systolic velocity, E': peak early diastolic transmitral flow velocity; A: peak late (atrial) diastolic transmitral flow velocity; E': peak mitral annular tissue velocity during early filling; E/E: peak early diastolic transmitral flow velocity to peak early mitral annular tissue velocity ratio; T0-before chemotherapy; T1-after two cycles of doxorubicin-based chemotherapy; T2-after four cycles of doxorubicin-based chemotherapy.

Variables
All patients underwent clinical evaluation, as well as 2DE and strain analysis by STE after two cycles (T1) and four cycles of doxorubicin-based chemotherapy (T2). A total of 22 patients (25.9%) met the criteria of LV cardiotoxicity after anthracycline-based chemotherapy. A decrease in LVEF, indicative of cardiotoxicity (from 61.8 ± 1.8% at baseline to 49.9 ± 2.1% at T2, p < 0.001), was more common three months after the end of the anthracycline therapy (n = 18, 81.8%) than that of early cardiotoxicity (n = 4, 18.2%). Table 1 shows clinical characteristics, and Table 3 summarizes echocardiographic parameters and their comparisons at baseline after two cycles and four cycles of doxorubicin-based chemotherapy in women who developed cardiotoxicity and those of women who remained without cardiotoxicity (n = 63, 74.1%).
There was no significant difference in baseline demographic and clinical characteristics between these two groups, aside from the fact that patients who developed cardiotoxicity had a higher prevalence of AH (68.2%), family history of premature atherosclerotic CVD (50%), and greater use of β-blockers (54.5%) ( Table 1).
The baseline LVEF and GLS were within normal limits in both groups. LV end diastolic diameter before chemotherapy was normal in both groups, but it was significantly larger in women who regressed to cardiotoxicity. M-mode and Doppler echocardiographic parameters did not differ between the two groups at baseline (Table 3).
Conventional parameters of longitudinal LV systolic function, such as mitral annulus plane systolic excursion (MAPSE) and mean peak mitral annulus systolic velocity by TDI (S'), decreased significantly from baseline in both groups (p < 0.001) during the study periods, but they were not significantly different between groups, except that MAPSE was lower at T1 follow-up (p < 0.05) in patients with cardiotoxicity than for those without cardiotoxicity ( Table 3).
The analysis of LV diastolic function showed that LV diastolic function mildly worsened in both groups during the follow-up periods, but it did not differ significantly between the groups after two cycles (T1) and four cycles (T2) of anthracycline-based chemotherapy ( Table 3).
In the univariate analysis, GLS, NT-proBNP, 6MWT, AH, and a family history of CVD were associated with subsequent LV cardiotoxicity p < 0.001 (Table 4), and, therefore, they were included in the multivariate logistic regression model. GLS and NT-proBNP were identified as independent predictors of an early development of cardiotoxicity in breast cancer patients treated with anthracycline-based chemotherapy. The area under the curve (AUC) of the ROC curves for GLS and NT-proBNP as independent predictors for an adverse outcome was 0.94, with p < 0.001 (95% CI: 0.868-0.980) and 0.78, as well as p < 0.001 (95% CI: 1.852-21.096), respectively. A relative decrease in GLS ≤ −18.0% (sensitivity: 72.73%; specificity: 92.06%) and a relative increase in NT-proBNP > 125 ng/L (sensitivity: 90.0%; specificity: 56.9%) were able to discriminate patients with or without LV cardiotoxicity ( Figure 4A,B).

Discussion
This study showed an early incidence of cancer therapy-related cardiotoxicity in breast cancer women treated with anthracycline-based chemotherapy before therapy with trastuzumab. Detection of early cardiotoxicity in this study was based on clinical assessment and echocardiography during the treatment with doxorubicin-based chemotherapy. In the present study, 22 of the 85 patients who received anthracycline-based chemotherapy exhibited an early LV dysfunction with no HF symptoms or signs during the monitoring period. The main findings of our study were: (1) asymptomatic LV systolic dysfunction occurred in many patients treated with anthracycline-based chemotherapy, particularly in patients with more common CV risk factors, and LV GLS was the parameter that was the earliest and most markedly impaired during chemotherapy; (2) LV GLS measured during doxorubicin-based chemotherapy was associated with the subsequent development of early cardiotoxicity based on LVEF criteria and may identify patients who will develop cardiotoxicity; (3) during anthracycline-based chemotherapy, the increase in serum NT-proBNP levels were more pronounced in patients with cardiotoxicity than without cardiotoxicity, and these were effective diagnostic indicators for predicting early LV cardiotoxicity; (4) exercise capacity in the 6MWT decreased in all patients during the study with chemotherapy, but the decrease was more pronounced in patients with than without cardiotoxicity and was associated with the development of LV cardiotoxicity.
Anthracyclines are a class of chemotherapies that include doxorubicin, widely used in the treatment of various solid and liquid tumors [14]. Doxorubicin-induced cardiotoxicity may occur through a variety of mechanisms, such as interaction with iron, activity of intracellular or intramitochondrial oxidizing enzymes, alteration of endothelin-1 expression in cardiomyocytes, and binding to topoisomerases [15][16][17][18][19][20]. AIC can manifest as asymptomatic LV dysfunction in up to 57% of treated patients [21,22], and symptomatic congestive HF can manifest in up to 16-20% of patients [14,23,24]. In clinical practice, however, these percentages may be even higher.
It is known that AIC is a dose-dependent and cumulative process at a total doxorubicin dose of 450 mg/m 2 [14]. Billingham et al.'s study results of endomyocardial biopsies showed histopathological changes with doses as low as 240 mg/m 2 , suggesting that subclinical cardiotoxicity may be present as early as the first dose [25]. The present study also found that, though the mean cumulative doxorubicin dose was 233.9 ± 29.6 mg/m 2 (ranging from 129.0 to 303.2) quite low, early subclinical LV systolic dysfunction was still observed in some patients in the study population, as is indicated in the literature [25].
Early subclinical cardiotoxicity diagnosis and treatment are determinant to improve prognosis. Therefore, early diagnosis of asymptomatic LV dysfunction is very important. Transthoracic echocardiography is the preferred imaging technique [26]. AIC diagnosis is based on a reduction in LVEF [6,7]. However, the LVEF measurement is a relatively insensitive tool for detecting cardiotoxicity at an early stage. This is mainly because significant changes in LVEF occur when a large amount of myocardial damage has taken place, and compensatory mechanisms are exhausted [27]. Deformation imaging by STE improves assessment of LV performance, yielding functional and prognostic information distinct from LVEF [28,29]. Kalam et al. published a large systematic meta-analysis (5721 adults) of GLS and LVEF with HF, acute myocardial infarction, and valvular and miscellaneous cardiac diseases, and they found that GLS was a more sensitive prognosticator of outcome compared to LVEF [26]. GLS has been reported to be more accurate with regard to LVEF in the detection of incipient and/or subclinical HF, for example, cancer therapy HF [28]. Consistent with previous studies [29][30][31], the present study showed that all patients treated with anthracycline-based chemotherapy had an early reduction in LV GLS, despite preserved LVEF, and it was more impaired after treatment in patients with cardiotoxicity. In patients with established cardiotoxicity, subclinical LV dysfunction was detected early-after two cycles of chemotherapy based on LV GLS change from baseline. Furthermore, LV GLS measured after two cycles of chemotherapy was able to independently predict subsequent cardiotoxicity. The ROC curve of LV GLS was plotted, and the AUC was 0.94 (95% CI: 0.868-0.980), and it was also concluded that the optimal GLS threshold was −18.3%, the diagnostic sensitivity of cardiotoxicity was 95.5%, and the specificity was 87.3%.
In recent years, natriuretic peptides have been widely used for the diagnosis of early doxorubicin-induced cardiotoxicity. BNP is a rapid and accurate indicator of HF, and it has a half-life of 15 to 20 min. NT-proBNP has a half-life of 60-120 min [32,33]. Because NT-proBNP has a long half-life and is stable, it can accumulate at higher concentrations, and therefore symptomatic and asymptomatic LV dysfunction can be diagnosed [33].
The combination of cardiac biomarkers and GLS is believed to increase diagnostic accuracy in early cardiotoxicity. The literature on the use of biomarkers for cardiotoxicity risk stratification before cancer therapy is still limited. It is recommended to measure cardiac serum biomarkers, such as cardiac troponin (cTn) I or T and BNP/NT-proBNP during cancer therapies, including anthracyclines and others [6,7]. Elevated NPs have the potential to identify early doxorubicin-induced cardiotoxicity because their rise can detect LV dysfunction earlier than echocardiography [34].
Feola et al. followed 53 breast cancer patients treated with anthracycline-based chemotherapy and found that patients had a reduction in LVEF of >10% at follow-up, with a baseline BNP of 55.5 ± 72.3 pg/mL, and those who did not had a baseline of 26.1 ± 21.4 pg/mL (p = 0.07 HR 0.96-1) (p = 0.07 HR 0.96-1) [9]. Other studies that investigated the association between BNP/NT-proBNP and cardiotoxicity showed different findings. Dodos et al. followed 100 patients who received anthracycline-based chemotherapy, and they did not find correlation between baseline NT-proBNP and LVEF reduction [35]. Sawaya et al. showed similar results-NT-proBNP did not predict cardiotoxicity [29].
The present study evaluated changes in serum NT-proBNP levels during anthracyclinebased chemotherapy and the association with cardiotoxicity in breast cancer patients. In our study, patients with progressive cardiotoxicity had higher baseline serum levels of NT-proBNP compared to patients without cardiotoxicity, suggesting that some of the study patients already had increased biological stress and myocardial strain [34] due to preexisting CV risk factors. In addition, increased concentrations of NT-proBNP were observed following anthracycline-based chemotherapy and higher NT-proBNP levels in the cardiotoxicity group compared to the non-cardiotoxic group, suggesting that increase was induced by the anthracycline-based treatment. The serial evaluation of NT-proBNP concentration in patients treated with doxorubicin-based chemotherapy showed a predictive value for the occurrence of cardiotoxicity. The area under the curve (AUC) was 0.78 (95% CI: 1.852-21.096), suggesting that the NT-proBNP level during the treatment with anthracycline had a role in diagnosing cardiotoxicity. ROC curve analysis revealed that the optimal NT-proBNP threshold was 125 ng/L. The diagnostic sensitivity of AIC was 90.0%, and the specificity was 56.9%.
6MWT is a widely available, easily performed, and well tolerated test for assessing the functional capacity of patients with HF. The benefit of 6MWT results in the diagnosis of HF has been studied in patients with HF [36], coronary artery disease [37], and cancer [38]. However, little is currently known about 6MWT and cardiotoxicity. One of our research objectives was to identify the role of 6MWT as an early predictor of cardiotoxicity in patients treated with anthracycline-based chemotherapy. However, although 6MET was associated with cardiotoxicity, it was not an independent predictor of cardiotoxicity in a multivariate analysis of a logistic regression model. This may have been due to the fact that none of the patients developed symptomatic LV dysfunction during the study.
One limitation of our study is that it is a relatively small sample. However, statistically significant correlation demonstrated that LV GLS and NT-proBNP may be used as independent predictors of doxorubicin-induced subclinical LV cardiotoxicity.
Another limitation of our study is that the chemotherapy regimen and the concomitant medication are nonhomogeneous. Most of our patients were treated with AC-P regimen, and all regimens used a similar cumulative dose of doxorubicin. It is known that some drugs used in breast cancer treatment regimens may contribute to cardiotoxicity, and we compared different regimens and did not find any difference in AIC [39]. Cardiotoxicity is considered the most important adverse reaction of doxorubicin-based chemotherapy in patients with breast cancer. Some drugs used in chemotherapy regimens have been shown to be associated with cardiac dysfunction. Therefore, our study results show that anthracycline-based chemotherapy, and not doxorubicin alone, may possibly contribute to AIC in breast cancer patients [40,41].

Conclusions
Despite their cardiotoxicity, anthracyclines are some of the most important drugs for breast cancer treatment. Therefore, while using these drugs, the risk factors associated with cardiotoxicity should be considered, and cardiac function should be accurately measured. Our study demonstrated that LV GLS and NT-pro BNP may be used as independent predictors of anthracycline-induced subclinical LV cardiotoxicity. Further studies are needed to establish more accurate baseline cut-off values for NT-proBNP and 6MWT in cancer patients, and they would be more suitable for the early diagnosis of cardiotoxicity.