Radiomics of Gadoxetate Disodium-Enhanced MRI in Glypican 3-Positive Hepatocellular Carcinoma Identification and Prognosis Evaluation

Abstract

Background: To investigate the impact of preoperative gadoxetate disodium (Gd-EOB-DTPA) MRI-based radiomics on predicting glypican 3 (GPC3) and recurrence-free survival (RFS) of hepatocellular carcinoma (HCC). Methods: Between January 2014 and October 2018, 259 patients with solitary HCC ≤5 cm who underwent liver resection and preoperative Gd-EOB-DTPA MRI were recruited. Draw support from 5-fold cross-validation, the best of 15 radiomics models obtained by combining five feature selection strategies and three classifiers was integrated with independent clinicoradiologic predictors into the comprehensive nomogram. Findings: GPC3 was an independent risk factor for postoperative recrudescence in patients with solitary HCC below 5 cm. Alpha-fetoprotein >20 ng/mL, homogenous T2-weighted imaging and hypointensity on hepatobiliary phase were independently vulnerable to GPC3 in the clinicoradiologic model. With ten features selected by support vector machines-recursive feature elimination, logistic regression-based classifier achieved the best performance among 15 radiomics models. After 5-fold cross-validation, our comprehensive nomogram acquired better average area under receiver operating characteristic curves with 95% confidence intervals (training and validation cohorts: 0.931[0.889-0.972] Vs. 0.943[0.880-0.998]) than the clinicoradiologic algorithm (0.738[0.651-0.824] Vs. 0.739[0.563-0.915]) and the optimal radiomics model (0.943[0.910-0.978] Vs. 0.931[0.860-0.989]). Net reclassification indexes further demonstrated the superiority of GPC3 nomogram over the clinicoradiologic and radiomics algorithms(46.54%, P 77.9 to 48.2 months, P=0.044), which was analogue to that of the histological GPC3-positive phenotype (from >73.9 to 43.2 months, P<0.001). Interpretation: Preoperative radiomics-based nomogram of Gd-EOB-DTPA MRI satisfactorily distinguished GPC3 status and outcomes of patients suffered from solitary HCC within 5 cm. Funding: This research was supported by National Natural Science Foundation of China (No.91859107), Zhongshan Hospital, Fudan University (No. 2018ZSLC22), Shanghai Municipal Key Clinical Specialty (No.W2019-018), and Clinical Research Plan of SHDC (No.SHDC2020CR1029B). The funding agency played no role in this study. Declaration of Interest: The authors declare no potential conflict of interest. Ethical Approval: This retrospective study was approved and granted an exemption from written informed consent by the ethics committee of our hospital (No. B2021-113).