Evolutionary pruning of transfer learned deep convolutional neural network for breast cancer diagnosis in digital breast tomosynthesis

This study reviews the technique of convolutional neural network (CNN) applied in a specific field of mammographic breast cancer diagnosis (MBCD). It aims to provide several clues on how to use CNN for related tasks. MBCD is a long-standing problem, and massive computer-aided diagnosis models have been proposed. The models of CNN-based MBCD can be broadly categorized into three groups. One is to design shallow or to modify existing models to decrease the time cost as well as the number of instances for training; another is to make the best use of a pretrained CNN by transfer learning and fine-tuning; the third is to take advantage of CNN models for feature extraction, and the differentiation of malignant lesions from benign ones is fulfilled by using machine learning classifiers. This study enrolls peer-reviewed journal publications and presents technical details and pros and cons of each model. Furthermore, the findings, challenges and limitations are summarized and some clues on the future work are also given. Conclusively, CNN-based MBCD is at its early stage, and there is still a long way ahead in achieving the ultimate goal of using deep learning tools to facilitate clinical practice. This review benefits scientific researchers, industrial engineers, and those who are devoted to intelligent cancer diagnosis.

Download Full-text

Comparison Between Feature-Based and Convolutional Neural Network–Based Computer-Aided Diagnosis for Breast Cancer Classification in Digital Breast Tomosynthesis

Journal of Biotechnology Research ◽

10.32861/jbr.51.1.18 ◽

2019 ◽

pp. 1-18

Author(s):

Siwa Chan ◽

Jinn-Yi Yeh

Keyword(s):

Breast Cancer ◽

Neural Network ◽

Convolutional Neural Network ◽

Digital Breast Tomosynthesis ◽

Cancer Classification ◽

Breast Tomosynthesis ◽

Breast Cancer Classification ◽

Computer Aided ◽

Feature Based ◽

Aided Diagnosis

Digital breast tomosynthesis (DBT) is a promising new technique for breast cancer diagnosis. DBT has the potential to overcome the tissue superimposition problems that occur on traditional mammograms for tumor detection. However, DBT generates numerous images, thereby creating a heavy workload for radiologists. Therefore, constructing an automatic computer-aided diagnosis (CAD) system for DBT image analysis is necessary. This study compared feature-based CAD and convolutional neural network (CNN)-based CAD for breast cancer classification from DBT images. The research methods included image preprocessing, candidate tumor identification, three-dimensional feature generation, classification, image cropping, augmentation, CNN model design, and deep learning. The precision rates (standard deviation) of the LeNet-based CNN CAD and the feature-based CAD for breast cancer classification were 89.84 (0.013) and 84.46 (0.082), respectively. The T value was -4.091 and the P value was 0.00 < 0.05, which indicate that the LeNet-based CNN CAD significantly outperform the feature-based CAD. However, there is no significantly differences between the LeNet-based CNN CAD and the feature-based CAD on other criteria. The results can be applied to clinical medicine and assist radiologists in breast cancer identification.

Download Full-text