End-to-End Convolutional Neural Network Framework for Breast Ultrasound Analysis Using Multiple Parametric Images Generated from Radiofrequency Signals

Abstract

Breast ultrasound (BUS) is an effective clinical modality for diagnosing breast abnormalities in women. Deep-learning techniques based on convolutional neural networks (CNN) have been widely used to analyze BUS images. However, the low quality of B-mode images owing to speckle noise and a lack of training datasets makes BUS analysis challenging in clinical applications. In this study, we proposed an end-to-end CNN framework for BUS analysis using multiple parametric images generated from radiofrequency (RF) signals. The entropy and phase images, which represent the microstructural and anatomical information, respectively, and the traditional B-mode images were used as parametric images in the time domain. In addition, the attenuation image, estimated from the frequency domain using RF signals, was used for the spectral features. Because one set of RF signals from one patient produced multiple images as CNN inputs, the proposed framework overcame the limitation of datasets in a broad sense of data augmentation while providing complementary information to compensate for the low quality of the B-mode images. The experimental results showed that the proposed architecture improved the classification accuracy and recall by 5.5% and 11.6%, respectively, compared with the traditional approach using only B-mode images. The proposed framework can be extended to various other parametric images in both the time and frequency domains using deep neural networks to improve its performance.

Deep learning algorithm using bispectrum analysis energy feature maps based on ultrasound radiofrequency signals to detect breast cancer

“…Kim et al. ( 18 ) employed CNN to process multiple parameter images generated from RF signals for benign and malignant breast tumor analysis, including grayscale, entropy, attenuation, and phase images. The highest accuracy and sensitivity were 83.00% and 92.24%, respectively.…”

“…It greatly reduces the number of network parameters and the computational complexity during learning. Weight sharing can be manifested in various ways, such as sharing convolution kernel weights or weights of the entire network module [18,19]. A well-designed weight sharing structure enhances network depth, efficiency, and a lightweight architecture.…”

“…Kim et al. ( 18 ) used CNN to process multiple parametric images generated from RF signals for breast benign and malignant classification. In their research, the highest classification accuracy of the network models based on entropy images, phase images, attenuation images, and ultrasound grayscale images were 82.00%, 74.50%, 74.50%, and 79.00%, respectively.…”

Improving breast cancer diagnosis by incorporating raw ultrasound parameters into machine learning

“…To address the limitations on utilizing B-mode images for AI input, several recent studies have employed radiofrequency (RF) data for convolutional neural network (CNN) based classification of malignant and benign breast masses. Kim et al ( 2022 ) generated parametric maps using RF data for the deep learning input, reporting a higher performance than B-mode input. Byra et al ( 2022 ) utilized RF data as deep learning input and achieved higher area under the curve (AUC) than B-mode for breast lesion classification and segmentation.…”