| Background: the quantitative CT(QCT) was currently used in the measuring of bone mineral density(BMD) in human bones and was used to diagnose osteoporosis. Its measuring result is quantitative diagnosis. Therefore, based on its theory, the QCT can quantitatively calculate the content of the specific substance. The author supposed that it could also do quantitative measurement on liver fat content.Objective: This study was designed to evaluate the repeatability of measurement of fatty liver in different regions of interest(ROI) by QCT.Material and Method: a retrospective analysis was made on 31 cases that had received bone mineral density QCT measurement for various reasons from Nov.2012 to Dec.2013, male 19 cases, female 12 cases, age 22-39 years old. The patients were asked about the disease history and were found that they had no positive feature. Besides, no abnormal biochemical parameters was found and no disease related to liver metabolism, such as fatty liver, alcoholic liver disease and hepatocellular carcinoma, etc., was found.Method: all patients received bone mineral density QCT plain scan. Five-sample solid modelling of Toshiba 64-row CT cushioning Mindways Company was adopted to conduct spiral scanning on all patients. The patients laid on the back, with feet first; scanning technology parameters: 120 Kv, 125 m As, the height of bed 730 mm, screw pitch: 0.985, DFOV: 400-500 mm. All the images were rebuilt as thin layer of 1mm. The rebuilt images were transferred to 3D spine software in QCT Pro of Mindways Company to measure. Because the collected cases of this time were to receive bone mineral density, and the scan area of all cases were not strictly according to midsection or whole abdomen range, so the livers, on the diaphragmatic dome of some patients, were not scanned wholly. At the same time, because the anatomic variation of left lobe of liver was relatively large, so the left lobe of liver of all cases would not be tested in this study and the right lobe of liver would only be scanned partly. The method of selecting the interest region of right lobe of liver: try to select the largest layer of right lobe of scanned liver, avoiding motion artifacts, avoiding volume effect artifacts resulted from biliary ducts, vessels and fossa for gallbladder in the liver. The areas of ROI were separately 100mm2, 200mm2 and 300mm2, and the thickness of ROI was 9mm. Exported the measured results. The valuable data in this study was BMD value and slope value. Percentage value of liver fat content(beta) would be calculated according to the relative computational formula provided by Mindways Company. The formula is as follows: 1 Observed BMD=BMD/FUC 2 Expected liver HU=1047.9+(slope+0.2174)*4.842-999.6 3 Expected Fat HU=942.9+(slope+0.2174)*(-31.42)-999.6 4 Expected liver BMD=Expected liver HU/slope 5 Expected Fat BMD=Expected Fat HU/slope 6 Beta=(Expected Liver BMD-Observed BMD) /(Expected Liver BMD – Expected Fat BMD) This study invited two highly educated physicians to do the measurement independently according to the listed method after training. All the cases were tested every other day, three times in total. Took the mean value from the beta value calculated from every case. Using SPSS 22 software to conduct variance analysis. Compared the difference of beta values under different ROI area of the same measurer. Adopted Bonferroni method to conduct pairwise comparison. Adopted Bland-Altman method to conduct beta value repeatability assessment under different ROI area. Calculated difference value 95% confidence interval(%). The consistency evaluation of two measures adopted Bland-Altman method and interclass correlation coefficient(ICC), ICC>0.75 indicating good consistency. P<0.05 was thought to have statistical significance.Result:1 the mean value of beta value under ROI 100 mm2 tested by first physician was 0.0386±0.275, the mean value of beta value under ROI 200 mm2 was 0.0361±0.266, the mean value of beta value under ROI 300 mm2 was 0.0353±0.275; the mean value of beta value under ROI 100 mm2 tested by second physician was 0.0390±0.299, the mean value of beta value under ROI 200 mm2 was 0.0367±0.307, the mean value of beta value under ROI 300 mm2 was 0.0353±0.307.2Conducted variance analysis on the beta values under different ROI area of the same measurer. Pairwise comparison of the testing results of two measurers, and the significance was larger than 0.05, that is, it had no statistical significance. This indicated that there was no difference between the testing results of under different ROI area by two physicians.3Adopted Bland-Altman method to conduct beta value repeatability assessment under different ROI area. The consistency of 100-200mm2, 100-300mm2 and 200-300mm2 of first physician were separately-0.014%-0.019%,-0.013%-0.020% and-0.018%-0.019%; the consistency of 100-200mm2, 100-300mm2 and 200-300mm2 of second physician were separately-0.018%-0.022%,-0.013%-0.021% and-0.014%-0.017%.4The beta value repeatability testing under the same ROI area by different measurer adopted Bland-Altman method. The consistency of ROI 100mm2 was-0.030%-0.030 and the interclass correlation coefficient was 0.858; 200mm2 consistency-0.026%-0.025%, correlation coefficient 0.9; 300mm2 consistency-0.024%-0.024%, correlation coefficient 0.91. All the correlation coefficients were larger than 0.75, that is, the testing results under the same ROI area had consistency.Conclusion: The measuring results of different ROI area by the same measurer and the measuring results under the same ROI area by different measures, they both had good repeatability and consistency. Therefore, this study results indicated that selecting different interest region(ROI) when measuring liver fat content by QCT would not affect the final results. This study had laid a root for further researching QCT quantitative measuring liver fat content. |
PDF Full Text Request