Applied Machine Learning in Spiral Breast-CT: Can We Train a Deep Convolutional Neural Network for Automatic, Standardized and Observer Independent Classification of Breast Density?

Anna Landsmann,Jann Wieler,Andreas Boss,Patryk Hejduk,Cristina Rossi,Alexander Ciritsis,Karol Borkowski

doi:10.3390/diagnostics12010181

Abstract

The aim of this study was to investigate the potential of a machine learning algorithm to accurately classify parenchymal density in spiral breast-CT (BCT), using a deep convolutional neural network (dCNN). In this retrospectively designed study, 634 examinations of 317 patients were included. After image selection and preparation, 5589 images from 634 different BCT examinations were sorted by a four-level density scale, ranging from A to D, using ACR BI-RADS-like criteria. Subsequently four different dCNN models (differences in optimizer and spatial resolution) were trained (70% of data), validated (20%) and tested on a “real-world” dataset (10%). Moreover, dCNN accuracy was compared to a human readout. The overall performance of the model with lowest resolution of input data was highest, reaching an accuracy on the “real-world” dataset of 85.8%. The intra-class correlation of the dCNN and the two readers was almost perfect (0.92) and kappa values between both readers and the dCNN were substantial (0.71–0.76). Moreover, the diagnostic performance between the readers and the dCNN showed very good correspondence with an AUC of 0.89. Artificial Intelligence in the form of a dCNN can be used for standardized, observer-independent and reliable classification of parenchymal density in a BCT examination.

Highlights

Breast cancer (BC) constitutes the second most common cause of cancer death in women
The aim of this study was to investigate the potential of a machine learning algorithm to accurately classify parenchymal density in spiral breast-CT (BCT), using a deep convolutional neural network
We evaluated whether a deep convolutional neural network trained with 5589 images of 634 BCT examinations allows for accurate, objective and standardized breast CT density (BCTD) classification

Summary

Introduction

Breast cancer (BC) constitutes the second most common cause of cancer death in women. While not established everywhere, mammography screening is known to reduce mortality in breast cancer and it has been estimated that the combination of mammography screening with adjuvant therapy results in a relative reduction in mortality in breast cancer patients of 37% [2]. Risk factors for breast cancer are diverse. Hormonal influences such as estrogen replacement therapy and oral contraception are discussed as potential risk factors for breast cancer [3]. Epidemiological studies have shown that the risk for breast cancer in women with a higher density of breast tissue may increase two to six times [4]. In spite of the clinical importance of the density of breast tissue, it is difficult to determine in a clinical examination [5]

Objectives

Methods

Results