Prediction of Pathological Complete Response after Neoadjuvant Chemotherapy for breast cancer using ensemble machine learning

Abstract

Neoadjuvant Chemotherapy is administered intravenously during the treatment of breast cancer. Prior to surgery, doctors recommend chemotherapy to shrink the large size of an invasive tumor. This research work proposes a Deux Machine Learning framework, implementing a double ensemble of Machine Learning algorithms for building an optimized and efficient solution to predict a complete pathological response of patients after Neoadjuvant Chemotherapy. Unlike algorithms which focus on the accuracy of prediction, the performance of the Deux Machine Learning framework is measured using a multi criteria decision-making technique known as weighted simple additive weighting (WSAW). The WSAW comprehensive performance score is calculated by considering ten evaluation metrics, namely, Accuracy, Mean Absolute Error, Root Mean Square Error, TP Rate, FP Rate, Precision, Recall, F-Measure, MCC, and ROC. The results are validated using the k-fold cross validation technique, achieving an accuracy of 99.08%. When the performance of the proposed framework is compared with the performance of state-of-the-art classifiers such as SVM and random forest, the results are quite promising. With the growing trend of the applications of Artificial Intelligence in Cancer research, Machine Learning has an important future in prognostication and decision-making. Keywords: Machine learning, Prediction, Neoadjuvant chemotherapy, Breast cancer, Pathological response