Bazzocchi A, Diano D, Albisinni U, Marchesini G, Battista G, Guglielmi G
Department of Specialized, Diagnostic, and Experimental Medicine, University of Bologna, Sant'Orsola Malpighi Hospital, Bologna, Italy.
Br J Radiol. 2014 Sep;87(1041):20140232. doi: 10.1259/bjr.20140232. Epub 2014 Jun 12.
To investigate the predictive value for hepatic steatosis of a new software for the quantification of visceral fat by dual-energy X-ray absorptiometry (DXA) and to design new regions of interest (ROIs).
Adult volunteers were prospectively screened for hepatic steatosis by ultrasonography to obtain a well-balanced population according to the presence/absence of the disease. 90 adult patients without steatosis and 90 with steatosis (mild, 53.3%; moderate, 37.7%; and severe, 10.0%) were recruited. On the same day, all subjects were submitted to blood testing and to anthropometric and whole-body DXA for body composition evaluation. A new software for android visceral fat assessment was employed, and six new "liver-suited" ROIs as well as two modified android ROIs were designed. Their association with steatosis grade was tested by correlation analysis.
Fat mass (FM) of the new ROIs showed the highest correlation coefficients with steatosis grade (ρ = 0.610-0.619; p < 0.001), which was also confirmed by multivariate analysis. On the whole population, the new ROIs maintained the highest predictive role for liver steatosis, with areas under the receiver operating characteristic curve up to 0.820 ± 0.032. Inter- and intra-operator agreement for the new ROIs was excellent (k = 0.915-1.000 and k = 0.927-1.000).
New ROIs could be designed, standardized and implemented in DXA whole-body scan to provide more specific and predictive values of hepatic lipid content.
This is the first study to investigate the predictive value for hepatic steatosis of visceral and regional FM assessed on the hepatic site by DXA in comparison with ultrasonography, anthropometry and surrogate markers derived by previously validated algorithms (fatty liver index).
研究一种通过双能X线吸收法（DXA）定量内脏脂肪的新软件对肝脂肪变性的预测价值，并设计新的感兴趣区域（ROIs）。
通过超声对成年志愿者进行前瞻性肝脂肪变性筛查，以根据疾病的有无获得一个均衡的人群。招募了90例无脂肪变性的成年患者和90例有脂肪变性的成年患者（轻度，53.3%；中度，37.7%；重度，10.0%）。在同一天，所有受试者均接受血液检测、人体测量和全身DXA以评估身体成分。采用一种用于评估腹部内脏脂肪的新软件，并设计了六个新的“适合肝脏”的ROIs以及两个改良的腹部ROIs。通过相关性分析测试它们与脂肪变性分级的关联。
新ROIs的脂肪量（FM）与脂肪变性分级显示出最高的相关系数（ρ = 0.610 - 0.619；p < 0.001），多变量分析也证实了这一点。在整个人群中，新ROIs对肝脂肪变性保持着最高的预测作用，受试者工作特征曲线下面积高达0.820 ± 0.032。新ROIs的操作者间和操作者内一致性极佳（k = 0.915 - 1.000和k = 0.927 - 1.000）。
可以在DXA全身扫描中设计、标准化并实施新的ROIs，以提供更具特异性的肝脂质含量预测值。
这是第一项研究通过DXA在肝脏部位评估的内脏和局部FM对肝脂肪变性的预测价值，并与超声、人体测量学以及先前经验证算法（脂肪肝指数）得出的替代标志物进行比较的研究。
