Fatty liver evaluation with double-Nakagami model under low-resolution conditions

Convolutional neural network classification of ultrasound parametric images based on echo-envelope statistics for the quantitative diagnosis of liver steatosis

PURPOSE

Early detection and quantitative evaluation of liver steatosis are crucial. Therefore, this study investigated a method for classifying ultrasound images to fatty liver grades based on echo-envelope statistics (ES) and convolutional neural network (CNN) analyses.

METHODS

Three fatty liver grades, i.e., normal, mild, and moderate-to-severe, were defined using the thresholds of the magnetic resonance imaging-derived proton density fat fraction (MRI-PDFF). There were 10 cases of each grade, totaling 30 cases. To visualize the texture information affected by the deposition of fat droplets within the liver, the maps of first- and fourth-order moments and the heat maps formed from both moments were employed as parametric images derived from the ES. Several dozen to hundreds of regions of interest (ROIs) were extracted from the liver region in each parametric image. A total of 7680 ROIs were utilized for the transfer learning of a pretrained VGG-16 and classified using the transfer-learned VGG-16.

RESULTS

The classification accuracies of the ROIs in all types of the parametric images were approximately 46%. The fatty liver grade for each case was determined by hard voting on the classified ROIs within the case. In the case of the fourth-order moment maps, the classification accuracy of the cases through hard voting mostly increased to approximately 63%.

CONCLUSIONS

The formation of parametric images derived from the ES and the CNN classification of the parametric images were proposed for the quantitative diagnosis of liver steatosis. In more than 60% of the cases, the fatty liver grade could be estimated solely using ultrasound images.

Ultrasound normalized cumulative residual entropy imaging: Theory, methodology, and application

BACKGROUND AND OBJECTIVE

Ultrasound information entropy imaging is an emerging quantitative ultrasound technique for characterizing local tissue scatterer concentrations and arrangements. However, the commonly used ultrasound Shannon entropy imaging based on histogram-derived discrete probability estimation suffers from the drawbacks of histogram settings dependence and unknown estimator performance. In this paper, we introduced the information-theoretic cumulative residual entropy (CRE) defined in a continuous distribution of cumulative distribution functions as a new entropy measure of ultrasound backscatter envelope uncertainty or complexity, and proposed ultrasound CRE imaging for tissue characterization.

METHODS

We theoretically analyzed the CRE for Rayleigh and Nakagami distributions and proposed a normalized CRE for characterizing scatterer distribution patterns. We proposed a method based on an empirical cumulative distribution function estimator and a trapezoidal numerical integration for estimating the normalized CRE from ultrasound backscatter envelope signals. We presented an ultrasound normalized CRE imaging scheme based on the normalized CRE estimator and the parallel computation technique. We also conducted theoretical analysis of the differential entropy which is an extension of the Shannon entropy to a continuous distribution, and introduced a method for ultrasound differential entropy estimation and imaging. Monte-Carlo simulation experiments were performed to evaluate the estimation accuracy of the normalized CRE and differential entropy estimators. Phantom simulation and clinical experiments were conducted to evaluate the performance of the proposed normalized CRE imaging in characterizing scatterer concentrations and hepatic steatosis (n = 204), respectively.

RESULTS

The theoretical normalized CRE for the Rayleigh distribution was π/4, corresponding to the case where there were ≥10 randomly distributed scatterers within the resolution cell of an ultrasound transducer. The theoretical normalized CRE for the Nakagami distribution decreased as the Nakagami parameter m increased, corresponding to that the ultrasound backscattered statistics varied from pre-Rayleigh to Rayleigh and to post-Rayleigh distributions. Monte-Carlo simulation experiments showed that the proposed normalized CRE and differential entropy estimators can produce a satisfying estimation accuracy even when the size of the test samples is small. Phantom simulation experiments showed that the proposed normalized CRE and differential entropy imaging can characterize scatterer concentrations. Clinical experiments showed that the proposed ultrasound normalized CRE imaging is capable to quantitatively characterize hepatic steatosis, outperforming ultrasound differential entropy imaging and being comparable to ultrasound Shannon entropy and Nakagami imaging.

CONCLUSION

This study sheds light on the theory and methodology of ultrasound normalized CRE. The proposed ultrasound normalized CRE can serve as a new, flexible quantitative ultrasound envelope statistics parameter. The proposed ultrasound normalized CRE imaging may find applications in quantified characterization of biological tissues. Our code will be made available publicly at https://github.com/zhouzhuhuang.

Clinical performance of ultrasonic backscatter parametric and nonparametric statistics in detecting early hepatic steatosis

Diagnosis of early hepatic steatosis would allow timely intervention. B-mode ultrasound imaging was in question for detecting early steatosis, especially with a variety of concomitant parenchymal disease. This study aimed to use the surgical specimen as a reference standard to elucidate the clinical performance of ultrasonic echogenicity and backscatter parametric and nonparametric statistics in real-world scenarios. Ultrasound radio-frequency (RF) signals of right liver lobe and patient data were collected preoperatively. Surgical specimen was then used to histologically determine staging of steatosis. A backscatter nonparametric statistic (h), a known backscatter parametric statistic, i.e., the Nakagami parameter (m), and a quantitative echo intensity (env) were calculated. Among the 236 patients included in the study, 93 were grade 0 (<5% fat) and 143 were with steatosis. All the env, m and h statistics had shown significant discriminatory power of steatosis grades (AUC = 0.643-0.907 with p-value < 0.001). Mann-Whitney U tests, however, revealed that only the backscatter statistics m and h were significantly different between the groups of grades 0 and 1 steatosis. The two-way ANOVA showed a significant confounding effect of the elevated ALT on env (p-value = 0.028), but no effect on m or h. Additionally, the severe fibrosis was found to be a significant covariate for m and h. Ultrasonic signals acquired from different scanners were found linearly comparable.

Appropriate indicator of modeling error for threshold-based model selection in statistics-based ultrasound tissue characterization

Analysis of the envelope statistics of ultrasound echo signals contributes to quantitative tissue characterization in medical ultrasound. Many probability distribution model functions have been studied, and the model function that should be used for tissue characterization depends on the type of disease, even in the same organ. Thus, an appropriate model selection is important for an accurate diagnosis. In this study, we aimed to select a model using threshold processing for modeling errors instead of a simple selection by minimizing the modeling error. For this purpose, we compared several indicators of modeling errors using random number simulations, ultrasonic simulation, and phantom experiment. The results validated that the Mahalanobis distance of moments is an appropriate indicator because it enables the use of a constant threshold value, regardless of the type of model function and data length.

Relationship between transmission/reception conditions of high-frequency plane wave compounding and evaluation accuracy of extended amplitude envelope statistics

The double-Nakagami (DN) model provides a method for analyzing the amplitude envelope statistics of quantitative ultrasound (QUS). In this study, the relationship between the sound field characteristics and the robustness of QUS evaluation was evaluated using five HF linear array probes and tissue-mimicking phantoms. Compound plane-wave imaging (CPWI) was used to acquire echo data. Five phantoms containing two types of scatterers were used to mimic fatty liver tissue. After clarifying the relationship between the sound field characteristics of the probes and QUS parameters, DN QUS parameters in 10 rat livers with different lipidification were evaluated using one HF linear array probe. For both phantom and in situ liver analyses, correlations between fat content and multiple QUS parameters were confirmed, suggesting that the combination of CPWI using a HF linear array probe with the DN model is a robust method for quantifying fatty liver and has potential clinical diagnostic applications.

Verification of effect of interference between multiple scatterers on the evaluation of backscattering coefficient

Backscatter coefficient analysis methods for biological tissues have been clinically applied but assume a homogeneous scattering medium. In addition, there are few examples of actual measurement studies in the HF band, and the consistency with theory has not been sufficiently confirmed. In this paper, the effect of correlations among scatterer positions on backscattering was investigated by performing experiments on inhomogeneous media having two types of scattering source with different structural and acoustic properties. In the echo data of phantoms containing two types of scatterer acquired by multiple sensors, the power and frequency dependence of the backscatter coefficient were different from theoretical calculations due to the interference effects of each scatterer. The effect of interference between the two types of scatterer was confirmed to be particularly strong for echoes acquired by the sensor at high intensity and HF, or for a higher number density of strong scatterers.

Modified multi-Rayleigh model-based statistical analysis of ultrasound envelope for quantification of liver steatosis and fibrosis

PURPOSE

Quantitative diagnosis of the degree of fibrosis progression is currently a focus of attention for fatty liver in nonalcoholic steatohepatitis (NASH). However, previous studies have focused on either lipid droplets or fibrotic tissue, and few have reported the evaluation of both in patients whose livers contain adipose and fibrous features. Our aim was to evaluate fibrosis tissue and lipid droplets in the liver.

METHODS

We used an analytical method combining the multi-Rayleigh (MRA) model and a healthy liver structure filter (HLSF) as a technique for statistical analysis of the amplitude envelope to estimate fat and fibrotic volumes in clinical datasets with different degrees of fat and fibrosis progression.

RESULTS

Fat mass was estimated based on the non-MRA fraction corresponding to the signal characteristics of aggregated lipid droplets. Non-MRA fraction has a positive correlation with fat mass and is effective for detecting moderate and severe fatty livers. Progression of fibrosis was estimated using MRA parameters in combination with the HLSF. The proposed method was used to extract non-healthy areas with characteristics of fibrotic tissue. Fibrosis in early fatty liver suggested the possibility of evaluation. On the other hand, fat was identified as a factor that reduced the accuracy of estimating fibrosis progression in moderate and severe fatty livers.

CONCLUSION

The proposed method was used to simultaneously evaluate fat mass and fibrosis progression in early fatty liver, suggesting the possibility of quantitative evaluation for discriminating between lipid droplets and fibrous tissue in the early fatty liver.

Quantification of limitations in statistical analysis of ultrasound echo envelope amplitudes

Ultrasound echo envelope statistics have been widely studied for quantitative tissue characterization. In ultrasound measurements, the size of the region-of-interest (ROI) is limited by several factors, such as the locality of the tissue characteristics and the depth dependence of the acoustic field of the ultrasound beam. In this case, the evaluated echo envelope statistics vary even when the envelope amplitudes follow the same population without any noise. In this study, the statistical variance of the moments caused by this finite number of samples was quantified based on the central limit theorem and the law of error propagation. The proposed principles were validated by random number simulation and used to quantify the statistical variance of Nakagami parameter estimation. Finally, the effective number of independent samples in an ultrasonic measurement was quantified based on the relationship between the ROI size and the ultrasound spatial resolution.

Determination of thermal and mechanical indices of an ultrasound probe with an inclined beam-axis in accordance with the new JIS (or IEC Standards)

According to the new JIS (or IEC Standards), it is necessary to measure the bounded-square output power (P 1×1) for the determination of the soft tissue thermal index (TIS) for both scanning and non-scanning modes of an ultrasonic diagnostic equipment. However, there is no report about how to measure P 1x1 for an ultrasound probe with an inclined beam-axis (UPIB). We designed an angle-corrected 1 cm × 1 cm absorbing mask in accordance with the spirit of IEC 62359 Ed. 2.0 and established a reasonable process for the measurement of TIS and mechanical index (MI) of a paste type probe (PTP), which is a specially designed UPIB for examination of the blood flow in a carotid artery more conveniently and efficiently. Both the TIS and MI of the PTP were much less than 1.0 at all selectable presettings of the blood flow monitor, the FURUHATA device, suggesting it is as safe as a common ultrasonic Doppler diagnostic device.

Improved robustness of multi-component analysis in amplitude envelope statistics using plane waves

We compared the evaluation accuracy of amplitude envelope statistics under the transmission and reception conditions of compounded plane wave imaging (CPWI) and focused beam imaging (FBI). In a basic study using a homogeneous phantom, we found that the amplitude gradient in the depth direction and the point spread function in the lateral direction spread in the FBI reduced the accuracy of evaluation in amplitude envelope statistics. On the other hand, CPWI showed a more stable evaluation than FBI because of the elimination of sound field characteristics. In CPWI, the multi-Rayleigh model discriminated signals from two types of scatterer with high accuracy in the evaluation using phantoms mimicking fatty liver. It was confirmed that the combination of CPWI and the multi-Rayleigh model is effective for detecting early fatty liver disease. The results show that CPWI is effective for improving the robustness of amplitude envelope statistics.

Backscatter properties of two-layer phantoms using a high-frequency ultrasound annular array

In a previous study, an annular-array transducer was employed to characterize homogeneous scattering phantoms and excised rat livers using backscatter envelope statistics and frequency domain analysis. A sound field correction method was also applied to take into account the average attenuation of the entire scattering medium. Here, we further generalized the evaluation of backscatter coefficient (BSC) using the annular array in order to study skin tissues with a complicated structure. In layered phantoms composed of two types of media with different scattering characteristics, the BSC was evaluated by the usual attenuation correction method, which revealed an expected large difference from the predicted BSC. In order to improve the BSC estimate, a correction method that applied the attenuation of each layer as a reference combined with a method that corrects based on the attenuation of the analysis position were applied. It was found that the method using the average attenuation of each layer is the most effective. This correction method is well adapted to the extended depth of field provided by an annular array.

Finite element simulation of transmission and reflection of acoustic waves in the ultrasonic transducer

In scanning acoustic microscopy (SAM), the image quality depends on several factors such as noise level, resolution, and interaction of the waves with sample boundaries. The theoretical equations for the reflection coefficient and transmission coefficient are suitable for plane boundaries but fail for curved/rough boundaries. We presented a finite element method-based modeling for the loss coefficients in SAM. A focused and unfocused lens with a flat object, furthermore a focused lens with a curved object was selected for loss coefficients calculation. The loss calculation in terms of energy for defining the acoustic reflection and transmission losses and its dependence on the radius of curvature of the test object has also been presented.

Basic concept and clinical applications of quantitative ultrasound (QUS) technologies

In the field of clinical ultrasound, the full digitalization of diagnostic equipment in the 2000s enabled the technological development of quantitative ultrasound (QUS), followed by multiple diagnostic technologies that have been put into practical use in recent years. In QUS, tissue characteristics are quantified and parameters are calculated by analyzing the radiofrequency (RF) echo signals returning to the transducer. However, the physical properties (and pathological level structure) of the biological tissues responsible for the imaging features and QUS parameters have not been sufficiently verified as there are various conditions for observing living tissue with ultrasound and inevitable discrepancies between theoretical and actual measurements. A major issue of QUS in clinical application is that the evaluation results depend on the acquisition conditions of the RF echo signal as the source of the image information, and also vary according to the model of the diagnostic device. In this paper, typical examples of QUS techniques for evaluating attenuation, speed of sound, amplitude envelope characteristics, and backscatter coefficient in living tissues are introduced. Exemplary basic research and clinical applications related to these technologies, and initiatives currently being undertaken to establish the QUS method as a true tissue characterization technology, are also discussed.