Abstract:The early detection of liver fibrosis is of significant importance. Deep learning analysis of ultrasound backscattered radiofrequency (RF) signals is emerging for tissue characterization as the RF signals carry abundant information related to tissue microstructures. However, the existing methods only used the time-domain information of the RF signals for liver fibrosis assessment, and the liver region of interest (ROI) is outlined manually. In this study, we proposed an approach for liver fibrosis assessment using deep learning models on ultrasound RF signals. The proposed method consisted of two-dimensional (2D) convolutional neural networks (CNNs) for automatic liver ROI segmentation from reconstructed B-mode ultrasound images and one-dimensional (1D) CNNs for liver fibrosis stage classification based on the frequency spectra (amplitude, phase, and power) of the segmented ROI signals. The Fourier transform was used to obtain the three kinds of frequency spectra. Two classical 2D CNNs were employed for liver ROI segmentation: U-Net and Attention U-Net. ROI spectrum signals were normalized and augmented using a sliding window technique. Ultrasound RF signals collected (with a 3-MHz transducer) from 613 participants (Group A) were included for liver ROI segmentation and those from 237 participants (Group B) for liver fibrosis stage classification, with a liver biopsy as the reference standard (Fibrosis stage: F0 = 27, F1 = 49, F2 = 51, F3 = 49, F4 = 61). In the test set of Group A, U-Net and Attention U-Net yielded Dice similarity coefficients of 95.05% and 94.68%, respectively. In the test set of Group B, the 1D CNN performed the best when using ROI phase spectrum signals to evaluate liver fibrosis stages ≥F1 (area under the receive operating characteristic curve, AUC: 0.957; accuracy: 89.19%; sensitivity: 85.17%; specificity: 93.75%), ≥F2 (AUC: 0.808; accuracy: 83.34%; sensitivity: 87.50%; specificity: 78.57%), and ≥F4 (AUC: 0.876; accuracy: 85.71%; sensitivity: 77.78%; specificity: 94.12%), and when using the power spectrum signals to evaluate ≥F3 (AUC: 0.729; accuracy: 77.14%; sensitivity: 77.27%; specificity: 76.92%). The experimental results demonstrated the feasibility of both the 2D and 1D CNNs in liver parenchyma detection and liver fibrosis characterization. The proposed methods have provided a new strategy for liver fibrosis assessment based on ultrasound RF signals, especially for early fibrosis detection. The findings of this study shed light on deep learning analysis of ultrasound RF signals in the frequency domain with automatic ROI segmentation.

Computer-aided diagnosis of fatty liver based on ultrasonic images]

Computer Aided Diagnosis of Fatty Liver Ultrasonic Images Based on Support Vector Machine

Computerized Features for LI-RADS Based Computer-Aided Diagnosis of Liver Lesions

Quantitative analysis and identification of liver B-scan ultrasonic image based on BP neural network

Fatty Liver Diagnosed by B-mode Ultrasonography Based on the Complexity Analysis

B-Scan Ultrasound Image Feature Extraction for Quantitative Grading of Fatty Liver Severity

A Computer-Aided Diagnosis Scheme for Detection of Fatty Liver in Vivo Based on Ultrasound Kurtosis Imaging.

Feature Analysis and Automatic Classification of B-mode Ultrasound Images of Fatty Liver

Deep Learning With Ultrasound Images Enhance the Diagnosis of Nonalcoholic Fatty Liver

SD Rats’ Fatty Liver Tissue Classification Based on Ultrasound Radiofrequency Signal

Intelligent Diagnosis of Non-alcoholic Fatty Liver Disease Based on Multi-modal Ultrasound Features

Fatty liver classification via risk controlled neural networks trained on grouped ultrasound image data

The Effects of TIS and MI on the Texture Features in Ultrasonic Fatty Liver Images

Recognition of fatty liver using hybrid neural network

Computer-aided Cirrhosis Diagnosis Via Automatic Liver Capsule Extraction and Combined Geometry-Texture Features.

Liver Fibrosis Identification Based on Ultrasound Images Captured under Varied Imaging Protocols

Method of edge energy template for detection of ultrasonic based fatty liver

Ultrasound‐Based Machine Learning Approach for Detection of Nonalcoholic Fatty Liver Disease

Application of Fuzzy Enhancement in the Diagnosis of Liver Cancer from Ultrasound Images

Ultrasonic Assessment of Liver Fibrosis Using One-Dimensional Convolutional Neural Networks Based on Frequency Spectra of Radiofrequency Signals with Deep Learning Segmentation of Liver Regions in B-Mode Images: A Feasibility Study

Classification Method of Ultrasonic Images of Fatty Livers Based on Threshold Segmentation