A total of 79 female patients admitted to Qilu Hospital of Shandong University were enrolled in this study and were divided into three groups according to their treatment plans. Patients in G1 group who received anthracycline-free and non-targeted regimens [Docetaxel (T) 75 mg/m2 on day 1 and Carboplatin (Cb) AUC 6 on day 1, 21 days for a cycle, a total of six cycles]; G2 included patients treated with anthracycline-containing regimens without targeted therapy [Doxorubicin (A) 60 mg/m2 on day 1 or Epirubicin (E) 60–100 mg/m2 on day 1, Cyclophosphamide (C) 600 mg/m2 on day 1, 21 days for a cycle, a total of four cycles; Docetaxel (T) 100 mg/m2 iv on day 1, 21 days as a cycle, a total of four cycles]; G3 included patients treated with anthracyclines combined with targeted drugs [Doxorubicin (A) 50–60 mg/m2 on day 1 or Epirubicin (E) 80–100 mg/m2 on day 1, Cyclophosphamide (C) 600 mg/m2 on day 1, 21 days for a cycle, a total of four cycles; followed by Docetaxel (T) 100 mg/m2 on the first day, 21 days as a cycle, a total of four cycles. At the same time, Trastuzumab (H) was given at the first dose of 8 mg/kg, followed by 6 mg/kg for a period of 21 days, for a total of four cycles. Combined with or without Pertuzumab, first dose of 840 mg and then 420 mg for each period of 21 days, for a period of four cycles]. All patients underwent echocardiography before chemotherapy (T0) and after the end of chemotherapy cycle 2 (T2), cycle 4 (T4), cycle 6 (T6), and cycle 8 (T8).

The patients who met the following inclusion criteria were enrolled: 1) those aged less than 70 years old; 2) patients diagnosed with breast cancer by histology or cytopathology; 3) patients set to receive adjuvant chemotherapy combined with trastuzumab combined with or without pertuzumab; 4) those with no previous history of chemotherapy or radiotherapy. The major exclusion criteria included poor echocardiography image quality that did not allow complete analysis of the ultrasonic data; patients with heart failure, history of coronary heart disease, acute coronary syndrome or myocarditis, myocardial infarction, severe valvular disease, cardiomyopathy and other serious heart disease; and patients with persistent atrial fibrillation and severe arrhythmia. The study protocol was approved by the Ethics Committee of Shandong University Qilu Hospital (Ethics approval number: 2018163), and written informed consent was obtained from all participants.

Data on age, body mass index, current or previous smoking, diabetes mellitus, coronary artery disease, systolic blood pressure (SBP), diastolic blood pressure (DBP), and administration of antihypertensive drugs and statins were collected at enrollment. Serum levels of fasting blood glucose (GLU), serum triglycerides, total cholesterol, low-density lipoprotein cholesterol (LDL-C) and high-density lipoprotein cholesterol (HDL-C) were measured in the clinical laboratory departments of Qilu hospitals.

All subjects were examined by a GE Vivid E95 or GE Vivid E9 color Doppler ultrasound with an M5S-D probe (probe frequency 1.4–4.6 Hz) and a 4V-D probe (probe frequency 1.5–4.0 Hz). Ultrasound images were collected, and bilateral brachial artery blood pressure was measured. In addition to standard echocardiography, the study included 3–5 cardiac cycles at frame rates of 50–70 frames per second on apical sections of four-, three- and two-chamber views that were stored digitally in original format for analysis. All strain analyses were performed using semiautomatic speckle tracking technology (EchoPAC203, GE Medical System, Milwaukee, Wisconsin) using the entire left ventricular model (the three apical views). Parts that were inadequately tracked were excluded. Three-dimensional full volume left ventricular acquisition using a maximum volume ratio matrix array sensor was attempted in all patients. The biplanar Simpson method was used to measure LVEF, while the measurement of GLS and myocardial work was based on automated function imaging (AFI). When the myocardial tracking was not accurate, the myocardium could be manually adjusted.

CTRCD (Zamorano et al., 2016) is defined as a reduction in LVEF of at least 5% from baseline to <53% (Negishi and Negishi, 2018) in absolute value, accompanied by symptoms or signs of heart failure, or a reduction in LVEF of at least 10% to an absolute value <53% without symptoms or signs of heart failure.

Statistical analysis involved the use of SPSS version 26.0 (SPSS, Chicago, IL) and GraphPad Prism 9.0 (GraphPad, United States). All measurement data were tested for normality and homogeneity of variance. Measurement data meeting the normal distribution were measured as the mean ± SD ( X ¯ ± s). Measurement data that did not meet the normal distribution were expressed as the median (interquartile range) [M (Q1∼Q3)], and the data between groups were compared using the Mann-Whitney U test. Paired t tests were used to compare the data from the same patient before and after chemotherapy. Survival analysis was performed by the Kaplan-Meier method, and the log-rank test was used for univariate analysis between groups. Multivariate survival analysis was performed by a Cox proportional risk regression model. p < 0.05 indicated that the difference was statistically significant.

A total of 206 breast cancer patients were screened from November 2018–January 2021. Among them, 79 were newly treated breast cancer patients who met the inclusion criteria and completed multicycle follow-up. All were female patients aged 22–66 years old, with a median age of 48 years old. There were 20 cases in the G1 group, 38 cases in the G2 group, and 21 cases in the G3 group. Among them, seven patients had hypertension, three patients had coronary heart disease, and three patients had diabetes. Eight patients were treated with dextroimide, and 26 patients were treated with other drugs (salvia miltiorrhiza polyphenol, cyclic adenosine glucamine, and safflower yellow). There were no significant differences in age, weight, height, BMI, body surface area (BSA), blood pressure, blood glucose or blood lipids among the different treatment groups (p > 0.05) (Table 1).

Compared with the G1 group, GLS in T4 and LVEF, GWI and GWE in T6 were significantly decreased in the G2 and G3 groups (p < 0.05). With the progression of chemotherapy, the values of LVEF, GLS, GWI and GWE gradually decreased; GLS significantly decreased at T2; and LVEF, GWI and GWE significantly decreased at T4 (Figure 1).

The 3D-STE parameters of the research object are shown in Table 2. Compared with the basic condition before chemotherapy, 3D-GLS, 3D-GCS, 3D-GAS and 3D-GRS were significantly decreased (p < 0.05). The decrease in 3D-GLS and 3D-GCS in G2 and G3 was more obvious than that in G1, the changes were statistically significant in T6, and 3D-GAS and 3D-GRS were significantly changed in T8 (Figure 2).

LVEF decreased in the course of chemotherapy, and the degree of decrease was G3 > G2 > G1. GLS decreased in the course of chemotherapy, and the degree of decrease was G3 > G2 > G1. GWI showed a downward trend in the course of chemotherapy, the change in GWI in the G1 group was not obvious, and the change range was G3 > G2. GWI showed a downward trend during chemotherapy, with no significant change in the G1 group, a similar decrease in the G2 group and G3 group at T4, and G3 > G2 after G6. Note: LVEF, left ventricular ejection fraction; GLS, left ventricular overall longitudinal strain; GWI, global myocardial work index; GWE, myocardial work efficiency; T0, baseline before chemotherapy; T2, after the second cycle of chemotherapy; T4, after the fourth cycle of chemotherapy; T6, after the sixth cycle of chemotherapy; T8, after the eighth cycle of chemotherapy. *p < 0.01 compared with G1.

3D-GLS, 3D-GCS, 3D-GAS and 3D-GRS all showed a downward trend during chemotherapy. The changes in 3D-GLS, 3D-GAS and 3D-GRS in the G3 group were significantly higher than those in the G1 and G2 groups (p < 0.05). The degree of decrease in 3D-GCS was G3 > G2 > G1. Note: 3D-GLS, left ventricular three-dimensional longitudinal peak strain; 3D-GCS, left ventricular three-dimensional peripheral peak strain; 3D-GAS, left ventricular three-dimensional area peak strain; 3D-GRS, left ventricular three-dimensional radial peak strain. *p < 0.01 compared with G1.

During the follow-up, a total of nine patients developed CTRCD (see Supplementary Figure S1 for an example of image analysis). In the G1 group, one patient developed CTRCD at T8, and in the G2 group, one patient developed CTRCD at T2, one patient at T6, and one patient at T8. In the G3 group, one patient developed CTRCD at T2, two patients at T6, and two patients at T8.

The difference in clinical data between the event group (n = 9) and the nonevent group (n = 70) is shown in Table 3. There were no significant differences in age, systolic blood pressure, diastolic blood pressure, heart rate, blood glucose, HDL-C or TC between the two groups (p > 0.05), while BMI, BSA, LDL-C and TG in the event group were significantly increased (p < 0.05) (Table 3). L2v is defined as the rate of change of LVEF at T2, and G2v (GLS) compared with baseline, I2v (GWI), E2v (GWE), S2v (3D-GLS), C2v (3D-GCS), and A2v (3D-GAS) are similarly defined.

Age, diastolic blood pressure, heart rate, E2v and A2v were not correlated with the occurrence of CTRCD. Systolic blood pressure, BMI, BSA, serum glucose levels, serum LDL-C levels, serum HDL-C levels, serum TC levels, serum TG levels, L2v, G2v, I2v, S2v and C2v were significantly correlated with the occurrence of end events (p < 0.05) (Supplementary Table S1).

The BMI level and G2v and C2v as parameters for predicting the occurrence of CTRCD in breast cancer patients after tumor treatment showed good predictive value, with AUCs of 0.922 (95% CI: 0.858–0.987), 0.776 (95% CI: 0.600–0.952) and 0.703 (95% CI: 0.488–0.918), respectively. At the 95th percentile risk threshold, the sensitivities were 88.9, 55.6, and 66.7%, and the specificities were 88.6, 94.3, and 81.4%, respectively. Among them, BMI is the optimal parameter. When BMI is higher than 27.03, the probability of CTRCD is 1.301 times that with a BMI less than 27.03. (Table 4, Figure 3).

The ROC curves of BMI, G2v and C2v for predicting the occurrence of cardiotoxicity are shown. The area under the curve of BMI (0.922) was better than that of G2v (0.776) and C2v (0.703), and its predictive value for the occurrence of cardiotoxicity was the highest.