A Light Gradient Boosting Machine-Enabled Early Prediction of Cardiotoxicity for Breast Cancer Patients

Abstract

<h3>Purpose/Objective(s)</h3> Cardiac adverse events such as decreased ejection fraction and heart failure are the major toxicities in breast cancer patients receiving chemoradiotherapy, which significantly compromises patients' survival and quality of life. However, to predict the increased risk of severe cardiotoxicities could be challenging due to the variations of current methods used to evaluate the cardiotoxicity symptoms, individual susceptibility and early markers of injury. A test that facilitates identification of patients who are likely to develop severe cardiotoxicities could have a personalized therapeutic implication for appropriate personalized treatment and/or early intervention in the care of breast cancer patients. In this study, we developed, for the first time to our knowledge, a Light Gradient Boosting Machine (LightGBM)-enabled predictive model to integrate patients' chart from electronic medical records (EMRs) for early prediction of severe cardiotoxicities. <h3>Materials/Methods</h3> A total of 179 breast cancer patients were included in this study. The patients were randomly partitioned to the training set and the validation set for LightGBM predictive model development and validation. The training features extracted from each patient include age, cancer stage, tumor size, tumor location, medical history, chemotherapy drugs, targeted therapy drugs, hormone therapy drugs, distant metastases, surgery, radiation therapy position, radiotherapy dose, electrocardiograph (ECG) signal score and left ventricular ejection fraction (LVEF) value before chemoradiotherapy. The utility of the LightGBM predictive model constructed in predicting severe cardiotoxicities was evaluated by the ROC analysis in the validation set. <h3>Results</h3> Among the 179 breast cancer patients, severe cardiotoxicities were found in the patients with older age, higher LVEF value before radiotherapy, later tumor stage, abnormal ECG signals with bradycardia, tachycardia, T wave and Q wave abnormalities. The patients who received chemotherapy or targeted therapy using epicorubicin, doxorubicin, evacizumab and pertuzumab were more vulnerable to severe cardiotoxicities than others. The AUC value of the LightGBM predictive model achieved in the validation set was 0.82, higher than the clinical significance threshold of 0.70. The feature important analysis showed that age, the LVEF value before chemoradiotherapy, cancer position, targeted therapy, tumor stage and hormone therapy were the most valuable risk predicting factors. <h3>Conclusion</h3> The LightGBM framework proposed herein affords a means to use EMR data to individualize the prediction of severe cardiotoxicities at point of care of patients with breast cancer receiving chemoradiotherapy, which can facilitate the identification of patients for whom early intervention are warranted before the therapy, thus potentially improving the utility of chemoradiotherapy for breast cancer from a precision treatment perspective.