Objective: To evaluate the accuracy of proton magnetic resonance spectroscopy (1H MRS) in quantitative measurement of intra-hepatic triglyceride (TG) stores in fatty liver at the levels of animal model and human body.
Methods: 40 male Wistar rats were randomly divided into 2 equal groups: fatty liver model group, fed with high fat diet for 16 weeks, and control group, fed with standard diet. Every 4 weeks 5 rats from each group underwent 1H MRS to examine the intra-hepatic triglyceride contents and then were sacrificed. The liver tissues were stained with Sudan III] to detect the fat content. Eleven hospitalized patients with fatty liver suspected or confirmed by CT and with indications of liver puncture underwent 1H MRS and then liver puncture within 7 days. Linear regression was conducted to analyze the correlation between the results of 1H MRS and liver pathology.
Results: Intrahepatic lipid was detected in all the rat models, some rats of the control group, and the patients with fatty live. The lipid peak area and the ratio of the CH2 lipid peak area relative to the water peak area increased along with the duration of high fat diet feeding. In the rat models there was a close positive correlation between the hepatic TG content measured in vivo by 1H MRS and pathological measurement of liver fat content [r = 0.899; P <0001; slope 14.45; gamma-intercept -0.51]. The similar correlation existed in the patients too [r =0.814; P <0001; slope 33.72; gamma-intercept 27.04]. When the data of the rats and patients were analyzed together, the percentage of liver fat was highly positively correlated with the ratio of area under fat peak to the area under water peak (r = 0.878; P <0001; slope 56.33; gamma-intercept -0.68).
Conclusion: A relative quantitative method to measure intrahepatic lipid content by 1H MRS, a safe, non-invasive method, is primarily established at the animal level and has been verified in humans with a good correlation with the pathological examination.