A Novel Deep Learning System for Breast Lesion Risk Stratification in Ultrasound Images

Abstract

This paper presents a novel deep learning system to classify breast lesions in ultrasound images into benign and malignant and into Breast Imaging Reporting and Data System (BI-RADS) six categories simultaneously. A multitask soft label generating architecture is proposed to improve the classification performance, in which task-correlated labels are obtained from a dual-task teacher network and utilized to guide the training of a student model. In student model, a consistency supervision mechanism is embedded to constrain that a prediction of BI-RADS is consistent with the predicted pathology result. Moreover, a cross-class loss function that penalizes different degrees of misclassified items with different weights is introduced to make the prediction of BI-RADS closer to the annotation. Experiments on our private and two public datasets show that the proposed system outperforms current state-of-the-art methods, demonstrating the great potential of our method in clinical diagnosis.