JMIR Publications

ABSTRACT

Background:

Machine learning (ML) has become a vital part of medical imaging research. ML methods have evolved over the years from manual seeded inputs to automatic initialization. The advancements in the ML field have led to more intelligent and self-reliant computer-aided diagnosis (CAD) systems as the learning ability of ML methods has been constantly improving. More and more automated methods are appearing with powerful feature learning and representations. Recent advancements of ML with deeper and extensive representation approach commonly known as deep learning (DL) approaches have made a very significant impact on improving the diagnostics capabilities of the CAD systems.

Objective:

This review aims to survey both traditional ML and DL literature in particular application for breast cancer diagnosis. The review also provides a brief insight into some well-known DL networks

Methods:

In this paper, we present an overview of ML and DL techniques with particular application for breast cancer. Specifically, we search the PubMed, Google Scholar, Medline, ScienceDirect,Springer and Web of Science databases and retrieve the studies in DL for the past five years that have used multi-view mammogram datasets.

Results:

The analysis of traditional machine learning reveals the limited usage of the methods, whereas the DL methods have great potential for implementation in clinical analysis and improve the diagnostic capability of existing CAD systems.

Conclusions:

From the literature, it can be found that heterogeneous breast densities make the masses more challenging to detect and classify as compared to calcifications. The traditional ML methods present confined approaches either limited to particular density type or datasets. While the DL methods show promising improvements in breast cancer diagnosis. However, there are still issues of data scarcity and computational cost which has been overcome to a significant extent by applying data augmentation and improved computational power of DL algorithms.