Hepatic fat quantification of magnetic resonance imaging whole-liver segmentation for assessing the severity of nonalcoholic fatty liver disease: comparison with a region of interest sampling method

作者：Qinhe Zhang, Ying Zhao, Shifeng Tian, Luhan Xie, Lihua Chen, Anliang Chen, Nan Wang, Qingwei Song, Haonan Zhang, Lizhi Xie, Zhiwei Shen, Ailian Liu · 发表于：Quantitative Imaging in Medicine and Surgery · 年份：2021 · DOI：10.21037/qims-20-989 · 被引用次数：35 · 研究领域：Liver Disease Diagnosis and Treatment、Cardiovascular Disease and Adiposity、MRI in cancer diagnosis

BACKGROUND: Accurate and early assessment of the hepatic fat content is crucial for patients with nonalcoholic fatty liver disease (NAFLD). For years, magnetic resonance imaging (MRI) has been considered the optimal noninvasive method for the assessment of fat accumulation. To avoid time-consuming manual placement of multiple regions of interest (ROI), the use of whole-liver segmentation has been proposed to measure liver fat, mainly for heterogeneous fat deposition. However, it remains uncertain whether the hepatic mean fat fraction (FF) obtained by whole-liver segmentation with the inclusion of intrahepatic vasculature is consistent with the traditional ROI sampling method. In this study, we assessed the accuracy of hepatic mean FF obtained by whole-liver segmentation in patients of NAFLD with different severities using the ROI sampling method as a reference standard. METHODS: Hepatic FFs were measured by whole-liver segmentation and the ROI sampling method (reference standard) using MRI scanning with the iterative decomposition of water and fat with echo an asymmetry at least-square estimation-iron quantification (IDEAL-IQ) sequence. SPSS version 25.0 software was used to analyze the correlation and consistency of data between the two methods. RESULTS: There was a strong correlation in hepatic FF between whole-liver segmentation and the ROI sampling method in healthy, mild, and moderate steatosis patients (r = 0.943, 0.990, and 0.961, respectively). Bland-Altman analysis s...