A Priori Prediction of Breast Cancer Response to Neoadjuvant Chemotherapy using Quantitative Ultrasound, Texture Derivative and Molecular Subtype

Abstract

Abstract The purpose of this study was to investigate the performances of the tumor response prediction prior to neoadjuvant chemotherapy based on quantitative ultrasound, tumour core-margin, texture derivative analyses, and molecular parameters in a large cohort of patients (n = 208) with locally advanced breast cancer and combined them to best determine tumour responses with machine learning approach. Two multi-features response prediction algorithms using a k-nearest neighbour and support vector machine were developed with leave-one out and hold-out cross-validation methods to evaluate the performance of the response prediction models. In a leave-one out approach, the quantitative ultrasound-texture analysis based model attained a good classification performance with 80% of accuracy and AUC of 0.83. Including molecular subtype in the model improve the performance to 83% of accuracy and 0.87 of AUC. Due to limited number of sample in the training process, a model developed with a hold-out approach exhibited slightly higher bias error in classification performance. The most relevant features selected in predicting the response groups are core-to-margin, texture derivative, and molecular subtype. These results imply that that tumour-margin, baseline texture-derivative analysis methods combined with molecular subtype can be potentially used for the prediction of ultimate treatment response in patients prior to neoadjuvant chemotherapy.