Abstract: Objective　To quantitatively analyze the correlation between fat contents in liver and pancreas and glucose tolerance by magnetic resonance imaging (MRI). Methods　The objects taking related examination at People's Hospital of Huoju Development Zone from August 2020 to June 2022 were selected, and were divided into a normal glucose tolerance group (NGT group), an impaired glucose tolerance group (IGT group), and a type 2 diabetes mellitus group (T2DM group). There were 18 males and 12 females in the NGT group (healthy volunteers), and they were 50-73 (60.3±7.4) years old. There were 16 males and 14 females in the IGT group, and they were 52-70 (61.7±6.9) years old. There were 20 males and 10 females in the T2DM group, and they were 50-74 (61.9±7.2) years old. Four images of lipid phase, water phase, in-phase, and antiphase in each object were obtained by T1 VIBE Dixon sequence of 1.5T magnetic resonance. The measurement data of normal distribution were expressed as (x±s). The correlation between liver and pancreas fat contents and laboratory indicators were analyzed by Spearman rank correlation analysis. One-way ANOVA and Kruskal-Wallis H test were applied. The differences in the continuous variates between the 3 groups were evaluated according to the homogeneity of ANOVA. Results　The fat contents of liver and pancreas were correlated with glycosylated hemoglobin (hBA1c), high density lipoprotein (HDL), alanine aminotransferase (ALT), 2-hour postprandial blood glucose (P2BG), creatinine (CR), uric acid (UA), triglyceride (TG), and body mass index (BMI). The fat contents of liver and pancreas were (17.88±5.91)% and (12.15±4.75)% in the T2DM group, were (10.63±3.27)% and (8.35±3.54)% in the IGT group, and were (8.83±2.04)% and (5.42±1.31)% in the NGT group, with statistical differences (F=8.34 and 10.53; both P<0.05). Conclusion　MRI quantitative analysis for fat contents in liver and pancreas is sensitive to glucose tolerance, and can provide non-invasive examination means for clinical evaluation of of treatment effect.

Key words: Magnetic resonance imaging, Type 2 diabetes mellitus, Glucose tolerance test

摘要： 目的　磁共振成像（MRI）定量分析肝、胰脂肪含量与糖耐量的相关性。方法　于2020年8月至2022年6月期间在中山火炬开发区人民医院进行相关检查的对象中抽选符合条件的样本成为本次研究的对象，分为：糖耐量正常（NGT）组、糖耐量异常（IGT）组、2型糖尿病（T2DM）组。NGT组（健康志愿者）：男18例，女12例，年龄50～73（60.3±7.4）岁；IGT组：男16例，女14例，年龄52～70（61.7±6.9）岁；T2DM组：男20例，女10例，年龄50～74（61.9±7.2）岁。采用1.5T磁共振T1 VIBE Dixon序列获得受试者“脂相、水相，同相位、反相位”4幅图像。计量资料符合正态分布，使用平均值±标准差描述连续变量。采用Spearman秩相关检验对肝脏和胰腺脂肪含量与各实验室指标进行相关分析。采用单因素方差分析或Kruskal-Wallis H检验，根据方差检验的均一性评价3组组间连续变量的差异。结果　肝脏和胰腺脂肪含量与实验室检查糖化血红蛋白（hBA1c）、高密度脂蛋白（HDL）、丙氨酸氨基转移酶（ALT）、餐后２ｈ血糖（P2BG）、肌酐（CR）、尿酸（UA）、三酰甘油（TG）、体质量指数（BMI）指标相关。T2DM组肝、胰脂肪含量［（17.88±5.91）%、（12.15±4.75）%］均高于IGT组［（10.63±3.27）%、（8.35±3.54）%］与NGT组［（8.83±2.04）%、（5.42±1.31）%］，差异均有统计学意义（F=8.34、10.53，均P<0.05）。结论　磁共振对肝、胰脂肪含量的定量分析判断对糖耐量异常有较高敏感性，能够为临床提供无创性检查手段，辅助临床对治疗效果的评价。

关键词: 磁共振成像, 2型糖尿病, 糖耐量实验