Evaluation and Prognostication of Gd-EOB-DTPA MRI and CT in Patients With Macrotrabecular-Massive Hepatocellular Carcinoma.

BACKGROUND: Macrotrabecular-massive hepatocellular carcinoma (MTM-HCC) is highly aggressive. Comparing the diagnosis ability of CT and gadoxetate disodium (Gd-EOB-DTPA) MRI for MTM-HCC are lacking. PURPOSE: To compare the performance of Gd-EOB-DTPA MRI and CT for differentiating MTM-HCC from non-MTM-HCC, and determine the prognostic indicator. STUDY TYPE: Retrospective. SUBJECTS: Post-surgery HCC patients, divided into the training (N = 272) and external validation (N = 44) cohorts. FIELD STRENGTH/SEQUENCE: 3.0 T, T1-weighted imaging, in-opp phase, and T1-weighted volumetric interpolated breath-hold examination/liver acquisition with volume acceleration; enhanced CT. ASSESSMENT: Three radiologists evaluated clinical characteristics (sex, age, liver disease, liver function, blood routine, alpha-fetoprotein [AFP] and prothrombin time international normalization ratio [PT-INR]) and imaging features (tumor length, intratumor fat, hemorrhage, arterial phase peritumoral enhancement, intratumor necrosis or ischemia, capsule, and peritumoral hepatobiliary phase [HBP] hypointensity). Compared the performance of CT and MRI for diagnosing MTM-HCC. Follow-up occurred every 3-6 months, and nomogram demonstrated the probability of MTM-HCC. STATISTICAL TESTS: Fisher test, t-test or Wilcoxon rank-sum test, area under the curve (AUC), 95% confidence interval (CI), multivariable logistic regression, Kaplan-Meier curve, and Cox proportional hazards. Significance level: P < 0.05. RESULTS: Gd-EOB-DTPA MRI (AUC: 0.793; 95% CI, 0.740-0.839) outperformed CT (AUC: 0.747; 95% CI, 0.691-0.797) in the training cohort. The nomogram, incorporating AFP, PT-INR, and MRI features (non-intratumor fat, incomplete capsule, intratumor necrosis or ischemia, and peritumoral HBP hypointensity) demonstrated powerful performance for diagnosing MTM-HCC with an AUC of 0.826 (95% CI, 0.631-1.000) in the external validation cohort. Median follow-up was 347 days (interquartile range [IQR], 606 days) for the training cohort and 222 days (IQR, 441 days) for external validation cohort. Intratumor necrosis or ischemia was an independent indicator for poor prognosis. DATA CONCLUSION: Gd-EOB-DTPA MRI might assist in preoperative diagnosis of MTM-HCC, and intratumor necrosis or ischemia was associated with poor prognosis. EVIDENCE LEVEL: 4 TECHNICAL EFFICACY: Stage 2.