BI-RADS-Net: An Explainable Multitask Learning Approach for Cancer Diagnosis in Breast Ultrasound Images

TL;DR

BI-RADS-Net is an explainable deep learning model for breast cancer detection in ultrasound images that provides clinically relevant explanations and predicts malignancy likelihood, improving trust and interpretability in medical diagnosis.

Contribution

This paper introduces BI-RADS-Net, a novel multitask deep learning approach that combines cancer classification with explanation generation based on BI-RADS descriptors.

Findings

Improved accuracy on breast ultrasound dataset

Provides explanations using BI-RADS features

Predicts malignancy likelihood effectively

Abstract

In healthcare, it is essential to explain the decision-making process of machine learning models to establish the trustworthiness of clinicians. This paper introduces BI-RADS-Net, a novel explainable deep learning approach for cancer detection in breast ultrasound images. The proposed approach incorporates tasks for explaining and classifying breast tumors, by learning feature representations relevant to clinical diagnosis. Explanations of the predictions (benign or malignant) are provided in terms of morphological features that are used by clinicians for diagnosis and reporting in medical practice. The employed features include the BI-RADS descriptors of shape, orientation, margin, echo pattern, and posterior features. Additionally, our approach predicts the likelihood of malignancy of the findings, which relates to the BI-RADS assessment category reported by clinicians. Experimental validation on a dataset consisting of 1,192 images indicates improved model accuracy, supported by explanations in clinical terms using the BI-RADS lexicon.