New Method for Evaluating Fatty Liver in Children by Ultrasound Backscattering Statistical
Parameter Imaging
DANG Yingnan1, GAO Ruiyang1, BIN Guangyu1, WU Shuicai1, CUI Boxiang2, ZHOU Zhuhuang1
1. Beijing International Science and Technology Cooperation Base for Intelligent Physiological Measurement and Clinical
Transformation, Faculty of Environment and Life, Beijing University of Technology, Beijing 100124, China;
2. College of Medicine, Chang Gung University, Taoyuan Taiwan 333323, China
Abstract：Objective Ultrasonic homodyne K distribution is a statistical model of backscattering envelope with the most physical
parameters. To explore the feasibility of ultrasonic homodyne K imaging in evaluating children’s fatty liver. Methods An improved
method of homodyne K distribution parameter estimation based on neural network was proposed to train the estimation model
with eigenvalue, and parallel computation was introduced into backscattering statistical parameter imaging. The backscattering
signals of children were collected with a 3.5 MHz probe. The improved backscattering statistical parameter imaging was applied
to the evaluation of fatty liver in children, and performance of the Nakagami imaging m parameter and the homodyne K imaging
k parameter and log10(α) parameter in diagnosing children’s fatty liver was analyzed. Three classifications were performed: G0
vs. G1-G3, G0-G1 vs. G2-G3, G0–G2 vs. G3. Results A total of 393 cases of fatty liver ultrasound backscattering signals were
obtained, of which 165 cases were normal G0, 33 cases were mild G1, 78 cases were moderate G2 and 117 cases were severe G3.
The average running time of the parallel computing method for ultrasonic Nakagami imaging and ultrasonic homodyne K imaging
was significantly shorter than that of the sliding window method, and the computing speed of the parallel computing method was
about 3 times that of the sliding window method. The parameters of log10(α) obtained the highest area under the curve, specificity,
and accuracy in all three binary classifications. Conclusion In this study, an ultrasound backscattering statistical parameter imaging
method based on parallel computing is designed to improve the imaging speed. Ultrasound backscattering statistical parameter
imaging can be used for quantitative evaluation of children’s fatty liver, especially for early detection, and the research results can
provide a new idea and a new method for quantitative ultrasound evaluation of children’s fatty liver.
党英楠1，高瑞阳1，宾光宇1，吴水才1，崔博翔2，周著黄1. 超声背散射统计参数成像评价儿童脂肪肝的新方法[J]. 中国医疗设备, 2023, 38(9): 17-24.
DANG Yingnan1, GAO Ruiyang1, BIN Guangyu1, WU Shuicai1, CUI Boxiang2, ZHOU Zhuhuang1. New Method for Evaluating Fatty Liver in Children by Ultrasound Backscattering Statistical
Parameter Imaging. China Medical Devices, 2023, 38(9): 17-24.
