A gadoxetic acid–enhanced MRI–based multivariable model using LI-RADS v2018 and other imaging features for preoperative prediction of macrotrabecular–massive hepatocellular carcinoma. Issue 153 (August 2022)

• Record Type: Journal Article
• Authors: Liang, Yingying; Xu, Fan; Wang, Zihua; Tan, Caihong; Zhang, Nianru; Wei, Xinhua; Jiang, Xinqing; Wu, Hongzhen
• Is Part Of: European journal of radiology. Issue 153(2022)
• Publication Date: 2022-08
• Subjects: ACR American College of Radiology -- AFP alpha-fetoprotein -- ALT alanine aminotransferase -- AP arterial phase -- APHE arterial phase hyperenhancement -- DWI diffusion-weighted imaging -- EOB-MRI gadoxetic acid–enhanced magnetic resonance imaging -- HBP hepatobiliary phase -- LI-RADS Liver Imaging Reporting and Data System -- MTM-HCC macrotrabecular-massive hepatocellular carcinoma -- PA plasma albumin -- PT prothrombin time -- PVP portal venous phase -- ROC receiver operating characteristic curves -- STB serum total bilirubin -- TTP total plasma protein -- TP transitional phase; Hepatocellular carcinoma -- Magnetic resonance imaging -- Gadoxetic acid -- Diagnosis; Medical radiology -- Periodicals; Radiology -- Periodicals; Radiologie médicale -- Périodiques; Medical radiology; Periodicals; 616.075705
• DOI: 10.1016/j.ejrad.2022.110356
• Languages: English
• ISSNs: 0720-048X

Abstract

Highlights: Absence of enhancing "capsule", absence of blood products in mass and presence of ascites was useful for predicting MTM-HCC. The EOB-MRI-based model yielded a sensitivity, specificity, and AUC of 35.90% (21.20, 52.80), 94.44% (84.60, 98.80), and 0.731 (0.629, 0.818). Model A achieved a comparable AUC than model D (0.73 vs. 0.70, p = 0.036), but a higher AUC than model B (0.73 vs. 0.67, p = 0.048) and model C (0.73 vs. 0.65, p = 0.005). Abstract: Purpose: To identify imaging features of macrotrabecular-massive hepatocellular carcinoma (MTM-HCC) using LI-RADS v2018 and other imaging features and to develop a gadoxetic acid–enhanced MRI (EOB-MRI)-based model for pretreatment prediction of MTM-HCC. Materials and methods: A total of 93 patients with pathologically proven HCC (39 MTM-HCC and 54 non-MTM-HCC) were retrospectively evaluated with EOB-MRI at 3 T. Imaging analysis according to LI-RADS v2018 was evaluated by two readers. Univariate and multivariate analyses were performed to determine independent predictors for MTM-HCC. Different logistic regression models were built based on MRI features, including model A (enhancing capsule, blood products in mass and ascites), model B (enhancing capsule and ascites), model C (blood products in mass and ascites), and model D (blood products in mass and enhancing capsule). Diagnostic performance was assessed by receiver operating characteristic (ROC) curves. Results: After multivariate analysis, absence of enhancing capsule (odds ratio = 0.102, p = 0.010), absence of blood products in mass (odds ratio = 0.073, p = 0.030), and with ascites (odds ratio = 55.677, p = 0.028) were identified as independent differential factors for the presence of MTM-HCC. Model A yielded a sensitivity, specificity, and AUC of 35.90% (21.20, 52.80), 94.44% (84.60, 98.80), and 0.731 (0.629, 0.818). Model A achieved a comparable AUC than model D (0.731 vs. 0.699, p = 0.333), but a higher AUC than model B (0.731 vs. 0.644, p = 0.048) and model C (0.731 vs. 0.650, p = 0.005). Conclusion: The EOB-MRI-based model is promising for noninvasively predicting MTM-HCC and may assist clinicians in pretreatment decisions.