Methods
The study population is part of the Israel Cardio-Oncology Registry (ICOR) – a prospective registry enrolling all patients evaluated in the cardio-oncology clinic at Tel Aviv Sourasky Medical Center. All patients signed an informed consent at the first visit in the clinic and are then followed prospectively. The registry was approved by the local ethics committee (Identifier: 0228-16-TLV) and is registered in clinicaltrials.gov (Identifier: NCT02818517).
This cohort evaluated patients diagnosed with breast cancer planned for ANT based therapy. All patients performed at least 2 echocardiographic evaluations, including GLS; at baseline before chemotherapy (T1) and during Doxorubicin therapy (T2) for the assessment of correlation between Dst and the standard diastolic parameters. A 3rd echocardiography exam (T3), assessed after the completion of Doxorubicin therapy, within 3 months, was evaluated for the development of GLS relative reduction between T1 and T3.
The exclusion criteria included LVEF<53% at T1 and significant GLS relative reduction at T2≥15%.
A complete past medical history, cardiac risk factors and medical treatment were noted.
Diastolic strain was evaluated by the time of lengthening (Dst) (ms) as shown in Fig. 1. The change in Dst was assessed between T1 to T2. A clinically significant reduction in GLS was considered as a relative reduction of ≥15% from T1 echo to T3, adhered to the standard benchmark set by previous studies [15].
Three standard apical views (4-chamber, 2-chamber, and apical long-axis) were recorded using a General Electric system, model Vivid S70 echocardiogram and were performed by the same vendor, technician and interpreting cardiologist. Routine Left ventricle (LV) echocardiographic parameters included LV diameters, and LVEF [16]. Early trans-mitral flow velocity (E), late atrial contraction (A) velocity, deceleration time (DT) and early diastolic mitral annular velocity (medial and lateral e’) were measured in the apical 4-chamber view to provide an estimate of LV diastolic function [17]. The peak E/e’ ratio was calculated (septal, lateral and average mitral E/e’ ratio). Left atrium volume index (LAVI) was calculated using the biplane area length method at end-systole [16]. Images were acquired using high frame rate (>50 frames/s) [18], and thereafter stored digitally for offline analysis. GLS was measured using STE software and tracking within an approximately 5 mm wide region of interest. An end-systolic frame was used to initialize LV boundaries which were then automatically tracked throughout the cardiac cycle. Manual corrections were performed to optimize boundary tracking as needed. Optimization of images for endocardial visualization through adjustment of gain, compress, and time-gain compensation controls was done prior to acquisition. Ds were evaluated by measuring the time of lengthening (ms) of the myocardium during diastole, from the point of aortic valve closure (AVC) to the early peak of the curve for each segment (Fig. 1). Dst was assessed in three apical views (2, 3 and 4 chamber), with 6 segments measured per each view, with a total of 18 segments per each exam.
Assessment of the relationship between accepted echocardiographic diastolic parameters and the speckle strain derived parameters was done using repeated measures mixed linear regression models with e’ or E/e’ as the dependent variable, strain measurements as fixed independent variables, and patient ID as the random variable. To assess the predictive ability of diastolic strain parameters on significant GLS reduction, individual logistic regression models were built with significant GLS reduction as the dependent variable and relative decrease in each diastolic strain parameter between T2 and echo T1 as independent variables. Using the results of the above models the best predictor of significant GLS reduction was used in a multivariate model to assess its ability to independently predict significant GLS reduction. First, a multivariate model was built with covariates including relative GLS reduction between T2 and T1, baseline cardiac risk factors, cardiotoxic chemotherapy used and cardioprotective medication used. The above primary model was then narrowed using a stepwise forward and backwards Akaike information criterion (AIC) based method in order to select the best predictive model which has lowest AIC. To further illustrate the diagnostic predictive power of Diastolic strain alone or in combination with the other model covariates receptor-operator (ROC) curves were built and AUC with 95% CI and Youden indexes were calculated. Comparison between AUC of ROC curves was done using the DeLong & DeLong method. To detect whether adding Diastolic strain data contributed to the multivariate model predictive ability, net classification index (NRI) was calculated for a logistic model with and without the added variable, 95% confidence interval for NRI (and its positive and negative components) was calculated using a bootstrapping method. Continuous variables are shown as mean±SD, while discrete variable as n(%). Results were considered significant when p<0.05. As this is a primary proof of concept investigation, all assessments were considered hypothesis generating and were not corrected for multiple comparisons. All calculations were done using R version 3.5.0, R Foundation for Statistical Computing, Vienna, Austria.