| Objective:To investigate the accuracy of ultra-short echo time(UTE)MR sequence quantitatively measuring severe liver iron accumulation and assess the influence of fat on the liver iron quantification.Materials and Method:Sixty-three male New Zealand rabbits were divided into two groups,30 in the iron overload group and 33 in the iron overload combined with fatty liver group.In the iron overload group,iron dextran(100mg Fe/ml)was injected intravenously in the ear to construct different grades of iron overload models.The iron overload combined with fatty liver group was fed high-fat diet for 2-8 weeks to establish different grades of fatty liver models.In the process of fatty liver modeling,different doses of iron were injected to establish different grades of iron overload models.All MR examinations were performed on a clinical 3T scanner(MR750W,GE Discovery 750W)using a 16-channel soft surface coil.UTE imaging was performed using a three-dimensional(3D)gradient echo sequence with eight TEs of 0.03,0.08,0.13,0.23,0.43,0.73,1.03 and 2.03 ms were respectively applied in UTE imaging and scanned.T2*Map imaging was performed using a gradient echo sequence with twelve TEs of 1.2,2.5,3.8,5.2,6.5,7.8,9.1,10.4,11.7,13.0,14.3and 15.7ms respectively.Calculation of UTE-R2* value: Two radiologists engaged in liver imaging use Radi Ant DICOM Viewer software to trace the area of interest,and then use the GAME model in Matlab software to calculate R2* value.Calculation of GRER2* value: Two radiologists engaged in liver imaging trace the region of interest on the R2* map generated by the Function Tool software package,and then directly acquire the R2* value.After scanning of all animal models,the liver tissues were obtained,and the liver iron concentration(LIC)was measured by inductively-coupled plasma spectrometer(ICP).Obtain liver tissues for conventional HE staining,use computer quantitative auxiliary software Image J to determine fat fraction,and use Brunt grading standard for fat grading.All statistical analyses were performed using SPSS and Graph Pad Prism software.The correlation between R2* and LIC was evaluated by Spearman rank correlation analysis.In order to evaluate the influence of fat on iron determination,a linear regression equation R2* value and LIC was established for the iron overload group or the iron overload combined with fatty liver group,respectively.The covariance analysis was used to compare slopes of the regression equation for two groups.Receiver operating characteristic(ROC)analysis was performed in all 49 rabbits to differentiate four iron levels on the basis of prespecified LIC thresholds(1.8,3.2,7.0,and 15.0 mg of iron per gram of dry tissue)that are widely used in iron chelation therapy.P<0.05 was considered statistical significance.Results:Among 63 New Zealand rabbits,49 rabbits survived(77.8%)and underwent MR scan.With the increase of the injected iron dose,the magnetic resonance image signal intensity of the iron overload rabbit model showed a downward trend.The range of LIC measured by inductively coupled plasma spectrometer is 0.9-27.4 mg Fe/g,with an average value of 10.25±7.47 mg Fe/g.the range of fat fraction measured by pathology is:1.43-68.46%,with an average value of 12.83±15.12 %.As the iron dose increased,the liver iron concentration showed a consistent increase.The T2*MAP sequence can only accurately quantify mild to moderate iron overload,and the GRE-R2* value has a significant positive correlation with liver iron concentration(r=0.920,P<0.0001).The UTE sequence can accurately quantify the liver iron concentration involved in this experiment,and the UTE-R2* value is significantly positively correlated with the liver iron concentration(r=0.947,P<0.0001).The correlation coefficients of R2* and LIC regression equations of the iron overload group and the iron overload combined with fatty liver group were r=0.953 and r=0.940,respectively.Covariance analysis showed that there was no significant difference in the slope of the equation between the iron overload group and the iron overload combined with fatty liver group(P>0.05),indicating that fat has no effect on iron quantification.ROC analysis shows that the best cut-off values of R2* for different grades of iron deposits of 1.8,3.2,7.0 and 15.0 mg/g are 623,678,1400 and 2050 Hz,respectively,and the areas under the curve are 0.919,0.968,0.992 and 0.998 respectively.Conclusion:The liver iron deposition coexisting with fat model can be constructed by injecting different amounts of iron dextran during high-fat feeding.The R2* value based on the multi-echo gradient sequence T2*MAP can only accurately quantify the mild to moderate iron overload model.It failed to quantify the heavy iron deposition.The R2* value based on UTE sequence can accurately quantify liver iron,especially for heavy iron deposition.At the same time,the fat-suppressing UTE sequence can eliminate the influence of fat on liver iron quantification.It can accurately quantify liver iron when coexisting with fat.It has a good graded diagnostic performance for different grades of iron overload. |
PDF Full Text Request