Accuracy evaluation of quantitative diagnosis method of liver fibrosis based on multi-Rayleigh model using optimal combination of input moments. (3rd June 2020)

• Record Type: Journal Article
• Authors: Zhang, Chuang; Hirata, Shinnosuke; Hachiya, Hiroyuki
• Is Part Of: Japanese journal of applied physics. Volume 59:Number SK(2020)
• Publication Date: 2020-06-03
• Subjects: ultrasound image -- multi-Rayleigh model -- liver fibrosis; Physics -- Periodicals; 621.05
• DOI: 10.35848/1347-4065/ab9352
• Languages: English
• ISSNs: 0021-4922

Abstract

Abstract: We have been developing a quantitative estimation method of liver fibrosis using the multi-Rayleigh model to express a probability density function of the echo envelope. In the estimation, a combination of moments of the echo amplitude is used as input to estimate model parameters. To realize a robust quantification of liver fibrosis stage, we searched for the optimal combination of moments in the complex space. By using the estimated results in the multi-Rayleigh model parameters' space, we found the optimal combination of moments in real and imaginary spaces that can improve the estimation accuracy of multi-Rayleigh model parameters. From the results of the accuracy analysis, it was found that it is difficult to estimate the mixture ratio for the early stage of fibrosis even if the optimal combination of moments is used. However, estimation of the variance ratio is accurate enough to quantify fibrosis and can provide useful information. It became clear that the quantification method using the multi-Rayleigh model is effective.