DISCUSSION
In our study of 96 long-term survivors with diffuse large B-cell lymphoma treated with anthracyclines, we examined the role of left and right ventricular parameters before the start of anthracycline and after the completion chemotherapy in identifying the late occurrence of CAEs (cardiac death, symptomatic heart failure, arrhythmia, Subclinical Cardiac dysfunction). Although LVGLS had proven prognostic value, RVEF provided additional and independent information and added incremental value to a model including LVGLS. We found that by using a model combined LVGLS and RVEF could increase the sensitivity for predicting the late CAE. To our knowledge, this is the first study to evaluate the role of combining RV and LV function among patients with lymphoma for predicting late CAE in the context of R-CHOP regimen. Previous published studies mainly focused on the role of LVGLS in cardiovascular toxicity[20-22]. This study has the following novel findings: 1). We reported the proportion of patients with CAE caused by R-CHOP treatment; 2). We applied simultaneous 3-D LV and 3-D RV assessment 3). RVEF decreased during the chemotherapy was a strong predictor of CAE, and 4) combining LVGLS and RVEF may increase the sensitivity for estimating cardiovascular prognosis after chemotherapy.
The majority of CAE reported in our cohort were arrhythmia (8) including and heart failure (5) with an occurrence of 8.3% and 5.2%, respectively within a median of 6 years of follow-up. The pooled proportion of patients developing heart failure (5.2%) in our study is in line with previous studies reporting on the development of symptomatic heart failure, with a 4-5% incidence at a cumulative doxorubicin dose of 400mg/m2[23, 24]. In our study, patients received concomitant treatment with cyclophosphamide, which complicated the interpretation of the association of doxorubicin and left ventricular dysfunction. A study that explored the association between cyclophosphamide doses and left ventricular dysfunction, found patients whose doses greater than 1.55 g/m2 were associated with a low incidence of cardiotoxicity[25]. Little is known about right ventricular (RV) function in patients with lymphoma treated with R-CHOP, a small sample study found survivors had markedly increased risk (26%) of having impaired RV function (based on ejection fraction, and compared with normative values) [13]. The proportion of patients with impaired RV function (13%) in our data is quite high, and in many conditions, impaired RV function is associated with an increased risk of heart failure.
LVEF was one of the most important parameters for the assessment of cardiac function and for predicting CAE. However, many studies have proved that LVEF remained stable and within normal limits during the whole course of chemotherapy[20, 26, 27]. Anthracycline-induced cardiomyopathy can be a regional or diffuse pattern so the global LV function could remain normal by compensatory mechanism of healthy cardiomyocyte in the early stage[28]. Similarly, LVEF was not chosen in stepwise Cox proportional hazard analysis in our study. GLS is a sensitive measure of cardiac function and cardiac injury. Many studies have proved that patients with a reduction in GLS despite a preserved EF are at risk for cardiac injury and cardiac dysfunction [29, 30]. A previous study found GLS was associated with cardiac events while LVEF was within the normal range in patients treated with anthracyclines. In our data, LVGLS was also associated with the late CAE, and the threshold of GLS allowed us to identify patients who developed CAE with 66.7% sensitivity and 82.1% specificity. The use of GLS may allow the identification of a group of patients at risk of subsequent CAE and help to guide preventive medicine.
The assessment of RV function has become increasingly popular and been recognized as important to predict cardiac toxicity [31]. RV functional assessment with 2-dimensional and Doppler echocardiography has shortcomings due to the complex geometry of the right ventricle. Nagata et al validated the accuracy of RV volumes and RVEF measurements by 3D TTE against CMR. Although the manual editing of RV endocardial surface was required by 3D TTE, the 3D TTE was more widely adopted and convenient compared with CMR measurements. RV function by 3D TTE has emerging roles in future prognosis on chemotherapy patients [15]. RVEF<40% was a significant independent predictor of mortality after adjustments for age and LVEF. Gulati et al found that RVEF <45% was a powerful predictor of adverse cardiac events [32]. Some studies focused on the consistent impairment of RV function among in survivors of childhood cancer after reaching adulthood[33, 34]. In the current study, we determined the prognostic value of 3D RVEF in R-CHOP treated patients without severe cardiac disease. We demonstrated that 3D RVEF is significantly associated with late CAE. Using variables selected by a stepwise forward procedure in Cox proportional hazard analysis, we established an incremental value of 3DRVEF over the LVGLS for predicting CAE. Our results clearly showed that 3DRVEF is an independent predictor of CAE.
Current guidelines on follow-up after chemotherapy focus on left ventricular function[9]. Many studies have started to explore and validate the important role of impaired RV function in prognosis in a wide range of diseases[10, 11, 35]. We feel that impaired RV function is a strong predictor of CAE in cancer survivors and a combination of LV and RV function parameters during chemotherapy can facilitate the prevention and treatment.
Finally, we want to emphasize the importance of the 3-D echo image quality for the meaningful analysis of the cardiovascular adverse event.