Preoperatively Identify the Microvascular Invasion of Hepatocellular Carcinoma with the Restricted Spectrum Imaging

Rationale And Objectives: To noninvasively and preoperatively identify the microvascular invasion (MVI) of hepatocellular carcinoma (HCC) with the restricted spectrum imaging (RSI).

Materials And Methods: 62 patients were included into this prospective study and underwent the RSI examination with a 3.0-T scanner. Mono-exponential diffusion-weighted imaging-derived apparent diffusion coefficient (ADC) and RSI-derived metrics including f (fraction of restricted diffusion), f (fraction of hindered diffusion), f (fraction of free diffusion), and ff (the multiply of f and f) were calculated. Univariate and multivariate logistic regression were used to select the independent risk factors. Nomogram-based model was constructed with the selected indexes. Receiver operative characteristics analysis and calibration curve were used to evaluate the diagnostic accuracy.

Results: MVI-positive HCC showed significantly higher f and lower ADC values (ADC: 1.549 ± 0.228 ×10 vs 1.365 ± 0.239 ×10 mm/s, P = .003; f: 0.1633 ± 0.0341 vs 0.2221 ± 0.0491, P < .001). Tumor size and f were selected as independent risk factors for MVI. The nomogram-based model was then constructed with tumor size and f. Nomogram-based model (area under ROC curve [AUC]= 0.856) yielded the best diagnostic accuracy followed by f (AUC=0.842) and ADC (AUC=0.708). The AUC of both the f and nomogram model were significantly higher than that of ADC.

Conclusion: RSI-derived metrics can be utilized to noninvasively and efficiently identify the MVI of HCC. Considering the importance of MVI as a significant prognostic factor for HCC, the utilization of RSI has the potential to assist in prognostic prediction and clinical management.