The Value Of Gd-EOB-DTPA Enhanced MRI Combined With T1 Mapping In Predicting Microvascular Invasion And Ki-67 High Expression In Hepatocellular Carcinoma

Objective To evaluate the value of gadolinium ethoxybenzyl diethylenetriamine pentaacetic acid（Gd-EOB-DTPA）enhanced MRI combined with T1 mapping in predicting microvascular invasion（MVI）and Ki-67 high expression in hepatocellular carcinoma（HCC）.Methods A retrospective analysis of 813 patients with highly suspected HCC who underwent Gd-EOB-DTPA enhanced MRI examination in the Affiliated Hospital of Qingdao University from March 2019 to June 2021.According to the inclusion and exclusion criteria of this study,92 patients with HCC（a total of 92 lesions）were included.According to the presence or absence of MVI,the patients were divided into microvascular invasion（MVI+）group and non-microvascular invasion（MVI-）group,According to the expression level of Ki-67,the patients were divided into Ki-67 high expression group（Ki-67>20%）and Ki-67 low expression group（Ki-67≤20%）.MRI plain scan,Gd-EOB-DTPA enhanced scan and T1mapping sequence scan were performed in all patients before operation.The included imaging qualitative parameters were tumor location,tumor margin,pseudocapsule,peritumoral enhancement in arterial phase,peritumoral hypointensity in the hepatobiliary phase;quantitative parameters include tumor diameter,plain scan T1 value(T1_pre),hepatobiliary phase T1 value(T1_HBP),T1 decrease value（ΔT1）,T1 value reduction rate（ΔT1%）,the ADC value.The information collected was subjected to a one-way analysis,the measurement data were tested by independent sample t test or Mann-Whitney U test,and the counting data were accurately tested byχ² test or Fisher（exact probability method）.The statistically significant variables were included in the binary Logistic regression model for analysis,the independent risk factors affecting MVI and Ki-67 high expression were obtained.The receiver operating characteristic（ROC）curve was drawn to evaluate the diagnostic efficacy of independent risk factors,and developed a nomogram prediction model,and evaluate the prediction efficiency of this model for HCC MVI and Ki-67 high expression.P<0.05 means the difference was statistically significant.Results There were significant differences between MVI+group（n=36）and MVI-group（n=56）in tumor margin,pseudocapsule,peritumoral enhancement in arterial phase and peritumoral hypointensity in the hepatobiliary phase（χ²=26.408,12.752,8.762,21.814,P<0.05）.There were significant differences in diameter,T1_pre,T1_HBP and ADC values（t/Z=-4.628,-3.278,-2.775,3.044,P<0.05）.Binary Logistic regression analysis showed that tumor margin unevenness and smoothness（OR=6.480,95%CI=2.064～20.351,P=0.001）,peritumoral hypointensity in the hepatobiliary phase（OR=6.368,95%CI=1.850～21.918,P=0.003）and lower ADC value（OR=0.030,95%CI=0.003～0.366,P=0.006）were independent risk factors for MVI in HCC.ROC curve analysis showed that the AUC of tumor margin,peritumoral hypointensity in the hepatobiliary phase and ADC value in predicting MVI in HCC were 0.773（95%CI=0.671～0.874,P<0.001）,0.734（95%CI=0.623～0.845,P<0.001）,0.670（95%CI=0.555～0.784,P=0.006）.The AUC of the nomogram model constructed by combining tumor margin,peritumoral hypointensity in the hepatobiliary phase and ADC values to predict the occurrence of MVI in HCC was 0.870（95%CI=0.793～0.947,P<0.001）,with a sensitivity of 86.1%and specificity of 76.8%,and this model has a high degree of calibration,suggesting that it can be used to predict the occurrence of MVI in HCC with some accuracy.There were significant differences between Ki-67 high expression group（37 cases）and Ki-67 low expression group（55 cases）in tumor margin,pseudocapsule,peritumoral enhancement in arterial phase and peritumoral hypointensity in the hepatobiliary phase（χ²=10.323,8.272,5.440,9.894,P<0.05）.There were significant differences in diameter,T1_pre and T1_HBP（t/Z=-2.634,-6.856,-7.143,P<0.05）.Binary Logistic regression analysis showed larger T1_pre（OR=1.007,95%CI=1.002～1.011,P=0.004）,larger T1_HBP（OR=1.009,95%CI=1.003～1.016,P=0.005）and larger diameter（OR=1.020,95%CI=0.999～1.042P=0.058）were independent risk factors for Ki-67 high expression.ROC curve analysis showed that the AUC T1_pre,T1_HBP and diameter to predict Ki-67 high expression in HCC were0.860（95%CI=0.783～0.938,P<0.001）,0.839（95%CI=0.758～0.920,P<0.001）and 0.662（95%CI=0.551～0.773,P=0.008）.The nomogram model constructed by combining T1_pre,T1_HBP and diameter predicted Ki-67 high expression with an AUC of 0.916（95%CI=0.860～0.973,P<0.001）,a sensitivity of 89.2%and a specificity of 85.5%,and this model has a high degree of calibration,suggesting that it is used to predict the occurrence of Ki-67 high expression in HCC with high accuracy.Conclusion In Gd-EOB-DTPA enhanced MRI combined with T1 mapping sequences,the nomogram model constructed by combining tumor margin,peritumoral hypointensity in the hepatobiliary phase and ADC values could effectively predict the occurrence of HCC MVI,and the nomogram model constructed by combining T1_pre,T1_HBP and diameter had high accuracy in predicting the occurrence of Ki-67 high expression in HCC.