| Purpose: To evaluate the imaging features of Gd-EOB-DTPA enhanced magnetic resonance imaging,T1 Mapping and Gd-EOB-DTPA multi-phase dynamic contrast-enhanced(MDCE-MRI)perfusion parameters in diagnosing microvascular invasion of hepatocellular carcinoma.Methods: According to inclusion and exclusion criteria,72 patients underwented magnetic resonance imaging(MRI)plain scan and Gd-EOB-DTPA general dynamic contrast-enhanced MRI were retrospectively included.36 patients underwented T1 mapping imaging and multi-phase dynamic contrast-enhanced MRI scan were prospectively included.We studied 108 lesions in a total of 108 patients.Two investigators independently analyzed the imaging features of liver lesions and compared them with pathology to evaluate the morphological features of MVI images.The MDCE-MRI images were converted to the NIFTI format required for analysis by MRIcron,and the respiratory motion correction was performed using MATALAB R2018b(imported statistical parameter mapping package SPM 12),and finally PMI0.4 was imported to calculate the perfusion parameters of HCC.The independent sample t test(normal distribution)or the Mann-Whitney U test(skewed distribution)test assessed the the difference of T1 reduction ratio and the perfusion parameters between the presence and absence of MVI.For the differential quantitative indicators,pathological as the gold standard,the area under the receiver operating characteristic curve analyzed the optimal quantitative index,the cutoff value and calculated the sensitivity,specificity,positive predictive value,negative predictive value and Youden index to diagnose MVI of hepatocellular carcinoma.Results: 1.HCC diameter equal to or greater than 3cm,hepatobiliary phase with low or slightly lower signal around the tumor,tumor with incomplete capsule or no complete capsule,non-smooth tumor edge(P<0.05)in Gd-EOB-DTPA enhanced MRI morphological indexs were risk factors for MVI.Diagnosis of MVI sensitivity,specificity was 73.5%,67.6%,61.8%,75.7%,88.2%,37.8%,91.2%,39.2%.Thepositive predictive value and negative predictive value was 0.51,0.85,0.54,0.81,0.37,0.88,0.41,0.91,respectively.The Youden index was 0.411,0.375,0.260,0.304.2.The T1 reduction ratio was not statistically significant between the MVI and MVI-free groups(P>0.05).3.Quantitative data perfusion parameters VPF,AFF,Ve was statistically different(P<0.05)between the presence and absence of MVI,the area under the receiver operating characteristic curve of diagnostic MVI was 0.735(95%CI: 0.545-0.924),0.726(95%CI: 0.530-0.923),0.749(95%CI: 0.554-0.944).Ve had the highest diagnostic efficiency compared with other perfusion parameters.When the cutoff value was 13.25mL/100 mL,the sensitivity and specificity was 77.8% and 74.1%,respectively.The positive predictive value and negative predictive value was 0.50,0.91,and the Youden index was 0.52.4.The Yoden index(0.52)for Ve diagnosis MVI was greater than the Yoden index for diagnosis of MVI for morphological indicators(0.41,0.38,0.26,0.30).Conclusion: 1.Tumor diameter equal to or greater than 3cm,hepatobiliary stage low or slightly lower signal around the tumor,tumor with incomplete or no complete capsule,non-smooth tumor edge in Gd-EOB-DTPA enhanced MRI morphological indexs had a certain diagnostic value.Tumor diameter equal to or greater than 3cm had a highest clinical diagnostic value among morphological indexs.2.The MDCE-MRI perfusion parameters VPF,AFF and Ve had a certain diagnostic value for MVI.Ve had the highest diagnostic value in perfusion parameters and was higher than the imaging morphological indexs. |