Multifeature analysis of an ultrasound quantitative diagnostic index for classifying nonalcoholic fatty liver disease

Multifeature analysis of an ultrasound quantitative diagnostic index for classifying nonalcoholic fatty liver disease
复制标题

DOI:
10.1038/srep35083
复制
发表时间:
2016-10-13
期刊:
影响因子:
4.6
通讯作者:
Yeh, Chih-Kuang
Yeh, Chih-Kuang
中科院分区:
综合性期刊3区
文献类型:
--
作者:
Liao, Yin-Yin;Yang, Kuen-Cheh;Yeh, Chih-Kuang

文献摘要

被引文献

相似文献

非酒精性脂肪性肝病(NAFLD)是一种与代谢综合征相关的慢性肝病。本研究应用了基于超声成像中纹理、后向散射和衰减特征的综合分析,旨在评估NAFLD的严重程度。从394例临床病例中获得的超声射频数据进行了分析,以提取三个纹理特征(自相关,总和平均值和总和方差),信噪比(SNR),和中心频率下移的斜率(CFDS斜率)。纹理,SNR和CFDS斜率相结合,产生一个定量诊断指数(QDI),范围从0到6。我们使用训练数据训练QDI,然后将其应用于测试数据以评估其效用。在训练数据中,NAFLD和重度NAFLD的受试者工作特征曲线下面积(AUC)分别为0.81和0.84。在测试数据中,NAFLD和重度NAFLD的AUC分别为0.73和0.81。QDI能够区分重度NAFLD和正常肝脏与轻度NAFLD,并且与代谢因素显著相关。本研究探讨了使用QDI提供关于肝组织不同物理特征的信息以提高NAFLD分级能力的潜力。
Nonalcoholic fatty liver disease (NAFLD) is a chronic liver disease related to metabolic syndrome. This study applied an integrated analysis based on texture, backscattering, and attenuation features in ultrasound imaging with the aim of assessing the severity of NAFLD. Ultrasound radiofrequency data obtained from 394 clinical cases were analyzed to extract three texture features (autocorrelation, sum average, and sum variance), the signal-to-noise ratio (SNR), and the slope of the center-frequency downshift (CFDS slope). The texture, SNR, and CFDS slope were combined to produce a quantitative diagnostic index (QDI) that ranged from 0 to 6. We trained the QDI using training data and then applied it to test data to assess its utility. In training data, the areas (AUCs) under the receiver operating characteristic curves for NAFLD and severe NAFLD were 0.81 and 0.84, respectively. In test data, the AUCs were 0.73 and 0.81 for NAFLD and severe NAFLD, respectively. The QDI was able to distinguish severe NAFLD and a normal liver from mild NAFLD, and it was significantly correlated with metabolic factors. This study explored the potential of using the QDI to supply information on different physical characteristics of liver tissues for advancing the ability to grade NAFLD.