Acoustic structure quantification (ASQ) is a recently developed technique widely used for detecting liver fibrosis. Ultrasound Nakagami parametric imaging based on the Nakagami distribution has been widely used to model echo amplitude distribution for tissue characterization. We explored the feasibility of using ultrasound Nakagami imaging as a model-based ASQ technique for assessing liver fibrosis. Standard ultrasound examinations were performed on 19 healthy volunteers and 91 patients with chronic hepatitis B and C (n = 110). Liver biopsy and ultrasound Nakagami imaging analysis were conducted to compare the METAVIR score and Nakagami parameter. The diagnostic value of ultrasound Nakagami imaging was evaluated using receiver operating characteristic (ROC) curves. The Nakagami parameter obtained through ultrasound Nakagami imaging decreased with an increase in the METAVIR score (p < 0.0001), representing an increase in the extent of pre-Rayleigh statistics for echo amplitude distribution. The area under the ROC curve (AUROC) was 0.88 for the diagnosis of any degree of fibrosis (≥F1), whereas it was 0.84, 0.69, and 0.67 for ≥F2, ≥F3, and ≥F4, respectively. Ultrasound Nakagami imaging is a model-based ASQ technique that can be beneficial for the clinical diagnosis of early liver fibrosis.

Accurate and generalizable quantitative scoring of liver steatosis from ultrasound images via scalable deep learning

Entropy is a quantitative measure of signal uncertainty and has been widely applied to ultrasound tissue characterization. Ultrasound assessment of hepatic steatosis typically involves a backscattered statistical analysis of signals based on information entropy. Deep learning extracts features for classification without any physical assumptions or considerations in acoustics. In this study, we assessed clinical values of information entropy and deep learning in the grading of hepatic steatosis. A total of 205 participants underwent ultrasound examinations. The image raw data were used for Shannon entropy imaging and for training and testing by the pretrained VGG-16 model, which has been employed for medical data analysis. The entropy imaging and VGG-16 model predictions were compared with histological examinations. The diagnostic performances in grading hepatic steatosis were evaluated using receiver operating characteristic (ROC) curve analysis and the DeLong test. The areas under the ROC curves when using the VGG-16 model to grade mild, moderate, and severe hepatic steatosis were 0.71, 0.75, and 0.88, respectively; those for entropy imaging were 0.68, 0.85, and 0.9, respectively. Ultrasound entropy, which varies with fatty infiltration in the liver, outperformed VGG-16 in identifying participants with moderate or severe hepatic steatosis (p < 0.05). The results indicated that physics-based information entropy for backscattering statistics analysis can be recommended for ultrasound diagnosis of hepatic steatosis, providing not only improved performance in grading but also clinical interpretations of hepatic steatosis.

Ultrasound imaging is a first-line assessment tool for hepatic steatosis. Properties of tissue microstructures correlate with the statistical distribution of ultrasound backscattered signals, which can be described by the Nakagami distribution (a widely adopted approximation of backscattered statistics). The double Nakagami distribution (DND) model, which combines two Nakagami distributions, was recently proposed for using high-frequency ultrasound to analyze backscattered statistics corresponding to lipid droplets in the fat-infiltrated liver. This study evaluated the clinical feasibility of the DND model in ultrasound parametric imaging of hepatic steatosis by conducting clinical experiments using low-frequency ultrasound dedicated to general abdominal examinations. A total of 204 patients were recruited, and ultrasound image raw data were acquired using a 3.5 MHz array transducer for DND parametric imaging using the sliding window technique. The DND parameters were compared with hepatic steatosis grades identified histologically. A receiver operating characteristic (ROC) curve analysis was used to evaluate the diagnostic performance. The results indicated that DND parametric imaging constructed using a sliding window with the side length of five times the pulse length of the transducer provided stable and reliable DND parameter estimations and visualized changes in the backscattered statistics caused by hepatic steatosis. The DND parameter increased with the hepatic steatosis grade. The areas under the ROC curve for identifying hepatic steatosis were 0.76 (≥mild), 0.81 (≥moderate), and 0.82 (≥severe). When using low-frequency ultrasound, DND imaging allows the clinical detection of hepatic steatosis and reflects information associated with lipid droplets in the fat-infiltrated liver.

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.

To explore factors associated with persistent hepatitis B virus (HBV) infection in a cohort of hepatocellular carcinoma (HCC)-affected families and then investigate factors that correlate with individual viral load among hepatitis B surface antigen (HBsAg)-positive relatives. We evaluated non-genetic factors associated with HBV replication in relatives of patients with HCC. Relatives of 355 HCC cases were interviewed using a structured questionnaire. Demographics, relationship to index case, HBsAg status of mothers and index cases were evaluated for association with the HBV persistent infection or viral load by generalized estimating equation analysis. Among 729 relatives enrolled, parent generation (P = 0.0076), index generation (P = 0.0044), mothers positive for HBsAg (P = 0.0007), and HBsAg-positive index cases (P = 5.98 × 10^-8) were associated with persistent HBV infection. Factors associated with HBV viral load were evaluated among 303 HBsAg-positive relatives. Parent generation (P = 0.0359) and sex (P = 0.0007) were independent factors associated with HBV viral load. The intra-family HBV viral load was evaluated in families clustered with HBsAg-positive siblings. An intra-family trend of similar HBV viral load was found for 27 of 46 (58.7%) families. Male offspring of HBsAg-positive mothers (P = 0.024) and older siblings were associated with high viral load. Sex and generation play important roles on HBV viral load. Maternal birth age and nutritional changes could be the reasons of viral load difference between generations.