[Study of magnetic resonance imaging based on liver imaging reporting and data system for evaluating phosphatidylinositol proteoglycan-3 expression in hepatocellular carcinoma].

Abstract

Objective: To clarify the diagnostic value of magnetic resonance imaging based on liver imaging reporting and data system (LI-RADS) for phosphatidylinositol proteoglycan-3 (GPC3) expression in hepatocellular carcinoma (HCC). Methods: Clinical and pathological data of 95 HCC cases with positive GPC3 expression (+) and 40 HCC cases with negative GPC3 expression (-) were retrospectively analyzed, and their MRI image features based on the 2018 version of LI-RADS were compared. Multivariate logistic regression analysis was used to determine the main predictors of GPC3 expression. Receiver operating characteristic curve was used further to determine the diagnostic efficacy of combined clinical imaging model to predict GPC3 expression. Enumeration data were compared with χ2 test or Fisher's exact test. Measurement data were compared using independent samples t-test or Mann-Whitney U-test. Results: There were statistically significant differences between HCC in GPC3 (+) and GPC3(-) group at alpha-fetoprotein (AFP) levels (χ2=31.814, P<0.000 1), and MRI features: capsular enhacement (χ2=4.108, P=0.043), halo type enhancement (χ2=4.847, P=0.028), and lesion apparent dispersion coefficient (ADC) (t=2.552, P=0.011 8). Multivariate regression analysis showed that AFP>20 μg/L (OR=9.358, P<0.000 1) and ADC≤1.404×10-3 mm2/s (OR=1.003, P=0.017) were independent predictors for GPC3 expression in HCC. The combined model and the area under the curve value for the diagnosis of GPC3(+) in HCC was 0.810, and its diagnostic sensitivity and specificity were 76.8% and 77.5%, respectively. Conclusion: AFP>20 μg/L and ADC≤1.404×10-3 mm2/s may indicate the expression of GPC3 in HCC, and the combination of the two diagnostic indicators can provide a simple and effective non-invasive diagnostic method for clinical practice.