Speed of sound index for liver steatosis estimation: a reliability study in normal subjects.

PURPOSE Non-alcoholic fatty liver disease (NAFLD) is the most widespread type of chronic liver disease in the Western countries. Ultrasound (US) is widely used for NAFLD staging. The Resona 7 US system (Mindray Bio-Medical Electronics Co., Ltd.) includes an image optimization and speed of ultrasound-related feature, Sound Speed Index (SSI). SSI is applied in a region of interest (ROI) that could potentially aid in tissue characterization. The purpose of this study is to evaluate the reliability of SSI on various examination parameters on normal subjects. METHODS Twenty normal subjects were examined by two radiologists performing SSI measurements in the liver in different ROI depths and sizes. Intraclass correlation coefficient (ICC) was calculated to measure intra- and inter-observer variability and inter-ROI variability. RESULTS For all ROIs and both radiologists, the mean inter-observer ICC was 0.62 and the mean intraobserver ICC was 0.52 and 0.79. The mean SSI values for all ROIs and examiners were in the range 1528.79-1540.16 m/s. CONCLUSION The results indicate that SSI can lead to reliable measurements on normal subjects, independent of ROI size but dependent on ROI placement. More studies processing NAFLD patients, utilizing reference methods of liver fat quantification either for reliability or correlation with SSI, should be performed to further investigate the relevance of the SSI as a potential biomarker in clinical practice for liver steatosis grading.

Liver Ultrasound Attenuation: An Ultrasound Attenuation Index for Liver Steatosis Assessment.

OBJECTIVES: Nonalcoholic fatty liver disease (NAFLD) is the most widespread chronic liver disease type in the Western countries. Ultrasound (US) is used for NAFLD and hepatic steatosis (HS) grading. The most popular US method for NAFLD assessment is the hepatorenal index (HRI), but because of its limitations, other noninvasive methods have been developed. The Resona 7 US system has recently incorporated an US attenuation-related quantitative feature, liver ultrasound attenuation (LiSA), for HS estimation. The purpose of this study is to compare LiSA's and HRI's performance on NAFLD assessment. METHODS: A total of 159 NAFLD patients having a magnetic resonance imaging-proton density fat fraction (MRI-PDFF) examination were examined by 2 radiologists, who performed LiSA and HRI measurements in the liver. Correlation of LiSA's and HRI's measurements with MRI-PDFF values was calculated through Pearson correlation coefficient (PCC). To further investigate the performance of LiSA and HRI, optimum cutoffs, provided by the literature, were used to correspond HS grades to MRI-PDFF results. Moreover, a receiver operating characteristic (ROC) analysis on LiSA measurements and steatosis grades was performed. RESULTS: Magnetic resonance imaging-PDFF was better correlated with LiSA (PCC = 0.80) than HRI (PCC = 0.67). Receiver operating characteristic analysis showed better performance range for LiSA (77.8%-91.8%) than for HRI (72.8%-85.4%) on all HS grades for all studies used for corresponding MRI-PDFF values to HS grades. CONCLUSIONS: The results indicate that LiSA is more accurate than HRI in HS differentiation and can lead to more accurate grading of HS on NAFLD patients.

Deep learning networks on chronic liver disease assessment with fine-tuning of shear wave elastography image sequences.

Chronic liver disease (CLD) is currently one of the major causes of death worldwide. If not treated, it may lead to cirrhosis, hepatic carcinoma and death. Ultrasound (US) shear wave elastography (SWE) is a relatively new, popular, non-invasive technique among radiologists. Although many studies have been published validating the SWE technique either in a clinical setting, or by applying machine learning on SWE elastograms, minimal work has been done on comparing the performance of popular pre-trained deep learning networks on CLD assessment. Currently available literature reports suggest technical advancements on specific deep learning structures, with specific inputs and usually on a limited CLD fibrosis stage class group, with limited comparison on competitive deep learning schemes fed with different input types. The aim of the present study is to compare some popular deep learning pre-trained networks using temporally stable and full elastograms, with or without augmentation as well as propose suitable deep learning schemes for CLD diagnosis and progress assessment. 200 liver biopsy validated patients with CLD, underwent US SWE examination. Four images from the same liver area were saved to extract elastograms and processed to exclude areas that were temporally unstable. Then, full and temporally stable masked elastograms for each patient were separately fed into GoogLeNet, AlexNet, VGG16, ResNet50 and DenseNet201 with and without augmentation. The networks were tested for differentiation of CLD stages in seven classification schemes over 30 repetitions using liver biopsy as the reference. All networks achieved maximum mean accuracies ranging from 87.2%-97.4% and area under the receiver operating characteristic curves (AUCs) ranging from 0.979-0.990 while the radiologists had AUCs ranging from 0.800-0.870. ResNet50 and DenseNet201 had better average performance than the other networks. The use of the temporal stability mask led to improved performance on about 50% of inputs and network combinations while augmentation led to lower performance for all networks. These findings can provide potential networks with higher accuracy and better setting in the CLD diagnosis and progress assessment. A larger data set would help identify the best network and settings for CLD assessment in clinical practice.

Comparison of Sound Touch Elastography, Shear Wave Elastography and Vibration-Controlled Transient Elastography in Chronic Liver Disease Assessment using Liver Biopsy as the "Reference Standard".

Chronic liver disease (CLD) is currently a major cause of death. Ultrasound elastography (USE) is an imaging method that has been developed for CLD assessment. Our aim in the study described here was to evaluate and compare a new commercial variant of USE, sound touch elastography (STE), with already established USE methods, shear wave elastography (SWE) and vibration-controlled transient elastography (VCTE), using liver biopsy as the "reference standard." For our study, 139 consecutive patients underwent standard liver STE, SWE and VCTE examinations with the corresponding ultrasound devices. A receiver operator characteristic (ROC) curve analysis was performed on the stiffness values measured with each method. ROC analysis revealed, for SWE, STE and VCTE, areas under the ROC curve of 0.9397, 0.9224 and 0.9348 for fibrosis stage (F), F ≥ F1; 0.9481, 0.9346 and 0.9415 for F ≥ F2; 0.9623, 0.9591 and 0.9631 for F ≥ F3; and 0.9581, 0.9541 and 0.9632 for F = F4, respectively. In conclusion, STE performs similarly to SWE and VCTE in CLD stage differentiation.