Assessment of an MR Elastography‐Based Nomogram as a Potential Imaging Biomarker for Predicting Microvascular Invasion of Hepatocellular Carcinoma

Background Microvascular invasion (MVI) is a well‐established poor prognostic factor for hepatocellular carcinoma (HCC). Preoperative prediction of MVI is important for both therapeutic and prognostic purposes, but noninvasive methods are lacking. Purpose To develop an MR elastography (MRE)‐based nomogram for the preoperative prediction of MVI in HCC. Study Type Prospective. Subjects A total of 111 patients with surgically resected single HCC (52 MVI‐positive and 59 MVI‐negative), randomly allocated to training and validation cohorts (7:3 ratio). Field Strength/Sequence 2D‐MRE and conventional sequences (T1‐weighted in‐phase and opposed phase gradient echo, T2‐weighted fast spin echo, diffusion‐weighted single‐shot spin echo echo‐planar, and dynamic contrast‐enhanced T1‐weighted gradient echo) at 3.0 T. Assessment MRE‐stiffness and conventional qualitative and quantitative MRI features were evaluated and compared between MVI‐positive and MVI‐negative HCCs. Statistical Tests Univariable and multivariable logistic regression analyses were applied to identify potential predictors for MVI, and a nomogram was constructed according to the predictive model. Receiver operating characteristic (ROC) curve analysis was performed to evaluate the diagnostic performance. Harrell's C‐index evaluated the discrimination performance of the nomogram, calibration curves analyzed its diagnostic performance and decision curve analysis determined its clinical usefulness. A P value <0.05 was considered statistically significant. Results Tumor stiffness >6.284 kPa (odds ratio [OR] = 24.38) and the presence of arterial peritumoral enhancement (OR = 6.36) were independent variables associated with MVI. The areas under the ROC curves for tumor stiffness were 0.81 (95% confidence interval [CI]: 0.70, 0.89) and 0.77 (95% CI: 0.60, 0.90) in the training and validation cohorts, respectively. When both predictive variables were integrated, the best nomogram performance was achieved with C‐indices of 0.88 (95% CI: 0.78, 0.94) and 0.87 (95% CI: 0.71, 0.96) in the two cohorts, fitting well in calibration curves. The decision curve exhibited optimal net benefit with a wide range of threshold probabilities for the nomogram. Data Conclusion An MRE‐based nomogram may be a potential noninvasive imaging biomarker for predicting MVI of HCC preoperatively. Evidence Level 2. Technical Efficacy Stage 2.