R2*and Fat Fraction Value Of IDEAL-IQ For Predicting Pathological Grade Of Hepatocellular Carcinoma

Objective Explore the possibility that R2*value and fat fraction of IDEAL-IQ sequence in evaluating the pathological grade of hepatocellular carcinoma(HCC).Methods Totally 20 patients(22 lesions)with HCC confirmed by pathology and patients without any treatment before MR examination underwent 3T MR examination(included IDEAL-IQ sequence),the imaging date were analyzed retrospectively.The IDEAL-IQ image were processed using the software provided by the manufacturer to create water,fat,R2*,and fat fraction map.Using AW4.5 workstation,two radiologists placed ROIs on the lesions essence ingredient and the liver parenchyma with no obvious lesion by enhancement(away from the lesion of 3 cm above),and avoid the intrahepatic vessels and bile ducts.Double-blind method is used to measure the ROIs of R2*value and fat fraction.HCC pathological grading is according to Edmondson-Steiner classification method into 3 groups:3 well,9 moderately and 10 poorly differentiated.According to the pathology or MR imaging to determine the background of liver cirrhosis.Using SPSS 21.0 statistical analysis software for statistical analysis.The consistencies of the data between the two observers were tested.Difference between HCC and liver background in terms of R2*and fat fraction were evaluated by independent sample t test.The correlation between R2*value,fat fraction of HCC,and pathological grades were analyzed.The R2*value and fat fraction of different grades were compared between each groups.The ROC curves were used to evaluate the R2*value and fat fraction of different pathological grades.Results The data consistency of two observers was good.R2*value of HCC and background liver was statistically significant different(p<0.05).R2*values of the HCC lesion are lower than their liver background.In HCC lesions and liver fat fraction has no statistical significance in the different background groups.The R2*value was moderate correlated with the pathological grade of HCC(p<0.05,rs?0.453).The value of R2*was increased with poorer pathological grade of HCCs.There were significant difference between three pathologic groups of R2*value and fat fraction.The area under the curve(AUC)of fat fraction value for differentiating highly differentiated HCC was 0.865,with the cut-off value of fat fraction of 6.05%,the sensitivity was 66.7%,the specificity was 100%.Conclusion Fat fraction value can be used as a MR quantitative index for determining the well-differentiated HCC.The R2*value was moderate correlated with the pathological grade of HCC.