This study was performed to prospectively develop and validate a radiomics nomogram for predicting postoperative early recurrence (≤1 year) of hepatocellular carcinoma (HCC) using whole-lesion radiomics features on preoperative gadoxetic acid-enhanced magnetic resonance (MR) images. In total, 155 patients (training cohort: n = 108; validation cohort: n = 47) with surgically confirmed HCC were enrolled in this IRB-approved prospective study. Three-dimensional whole-lesion regions of interest were manually delineated along the tumour margins on multi-sequence MR images. Radiomics features were generated and selected to build a radiomics score using the least absolute shrinkage and selection operator (LASSO) method. Clinical characteristics and qualitative imaging features were identified by two independent radiologists and combined to establish a clinical-radiological nomogram. A radiomics nomogram comprising the radiomics score and clinical-radiological risk factors was constructed based on multivariable logistic regression analysis. Diagnostic performance and clinical usefulness were measured by receiver operation characteristic (ROC) and decision curves. In total, 14 radiomics features were selected to construct the radiomics score. For the clinical-radiological nomogram, the alpha-fetoprotein (AFP) level, gross vascular invasion and non-smooth tumour margin were included. The radiomics nomogram integrating the radiomics score with clinical-radiological risk factors showed better discriminative performance (AUC = 0.844, 95%CI, 0.769 to 0.919) than the clinical-radiological nomogram (AUC = 0.796, 95%CI, 0.712 to 0.881; P = 0.045), with increased clinical usefulness confirmed using a decision curve analysis. Incorporating multiple predictive factors, the radiomics nomogram demonstrated great potential in the preoperative prediction of early HCC recurrence after surgery.

Objective: The aim of this study was to assess the potential role of fractal analysis derived from Gd-EOB-DTPA-enhanced MRI in predicting vessels that encapsulate tumor clusters (VETC) in patients with hepatocellular carcinoma (HCC). Methods: This retrospective study included 505 patients with HCC who underwent Gd-EOB-DTPA-enhanced MRI before surgical resection at two medical centers (training set: 253 patients, internal test set: 108 patients, external test set: 144 patients). The fractal dimension (FD) and lacunarity were extracted from the hepatobiliary phase of the tumor using box-counting algorithms. Additionally, conventional imaging features were evaluated. Univariate and multivariate logistic regression analyses were conducted in the training set to identify independent predictors for VETC, and a nomogram was created to visualize the final predictive model. The performance of these models was tested in the internal and external test sets. Recurrence-free survival (RFS) and overall survival (OS) were analyzed using the Kaplan–Meier method along with the log-rank test. Results: VETC-positive HCC exhibited higher FD and lacunarity than VETC-negative HCC ( P < 0 .001). The FD-lacunarity model achieved an area under receiver operating characteristics curve (AUC) of 0.78 (95% confidence interval [CI]: 0.70–0.87) in the internal test set and 0.79 (95%CI: 0.70–0.86) in the external test set. Multivariate logistic regression analysis identified serum alpha-fetoprotein, tumor size, intratumor artery, FD, and lacunarity as independent predictors for VETC, which were used for constructing the hybrid model. A clinical model was established using AFP, tumor size, and intratumor artery alone. The diagnostic performance of the hybrid model was significantly surpassed that of the clinical-radiological model when fractal parameters were incorporated, with AUCs increasing from 0.72 to 0.80 in the internal test set and from 0.65 to 0.84 in the external test set (all P < 0.05). Patients predicted by the hybrid model to have VETC-positive HCC exhibited significantly shorter RFS and OS compared to those predicted to have VETC-negative HCC ( P < 0.05). Conclusion: Fractal analysis based on Gd-EOB-DTPA-enhanced MRI enabled the quantitative characterization of VETC status by fractal dimension and lacunarity. The hybrid model may assist in estimating VETC and stratifying prognosis in patients with HCC.

Objectives. To predict the regenerative rate of liver in patients with HCCs after right hepatectomy using texture analysis on preoperative CT combined with clinical features. Materials and Methods. 88 patients with 90 HCCs who underwent right hepatectomy were retrospectively included. The future remnant liver was semiautomatically segmented, and the volume of future remnant liver on preoperative CT (LV_pre) and the volume of remnant liver on following-up CT (LV_fu) were measured. We calculated the regeneration index (RI) by the following equation: (LV_fuLV_pre)/LV_pre)×100 (). The support vector machine recursive method was used for the feature selection. The Naive Bayes classifier was used to predict liver RI, and 5-fold cross-validation was performed to adjust the parameters. Sensitivity, specificity, and accuracy were calculated to evaluate the diagnostic efficiency of the model. Results. The mean RI was 142.99±92.17. Of all clinical parameters and texture features, the AST, ALB, PT-INR, Perc.10, and S(5, 5)Correlat were found to be statistically significant with RI. The diagnostic sensitivity, specificity, and accuracy of the model in the training group were 0.902, 0.634, and 0.768, and the AUC value of the obtained model was 0.841. In the test group, the sensitivity, specificity, and accuracy of the model were 1.0, 0.429, and 0.778, respectively, and the AUC value was 0.844. Conclusion. The use of texture analysis on preoperative CT combined with clinical features can be helpful in predicting the liver regeneration rate in patients with HCCs after right hepatectomy.

Objectives: This study aimed to assess the potential role of 3-dimensional (3D) fractal dimension (FD) derived from contrast-enhanced CT images in predicting microvascular invasion (MVI) in patients with hepatocellular carcinoma (HCC). Materials and methods: This retrospective study included 655 patients with surgically confirmed HCC from two medical centers (training set: 406 patients; internal test set: 170 patients; external test set: 79 patients). Box-counting algorithms were used to compute 3D FD values from portal venous phase images. Univariable and multivariable logistic regression analyses identified independent predictors. The model’s area under the curve (AUC) was calculated. Recurrence-free survival (RFS) and overall survival (OS) were evaluated using the Kaplan–Meier method. Results: Patients with MVI-positive HCC demonstrated significantly higher FD values compared to those with MVI-negative HCC ( p < 0.01). The FD achieved AUCs of 0.786 (95% CI: 0.713–0.849) in the internal test set and 0.776 (95% CI: 0.669–0.874) in the external test set. A combined model incorporating alpha-fetoprotein, tumor size, tumor number, and FD showed superior diagnostic performance for MVI prediction compared to the clinical model, with AUCs of 0.795 (95% CI: 0.720–0.860) vs 0.752 (95% CI: 0.670–0.825) in the internal test set, and 0.826 (95% CI: 0.721–0.915) vs 0.739 (95% CI: 0.613–0.849) in the external test set. Patients stratified as high-risk MVI exhibited significantly worse RFS and OS outcomes compared to low-risk MVI patients ( p < 0.05). Conclusion: The 3D FD values differed significantly between MVI-positive and MVI-negative HCC patients. Integrating FD into the clinical model enhanced MVI prediction accuracy and may help identify patients at high risk. Key Points: Question The predictive value of three-dimensional (3D) fractal dimension (FD) derived from contrast-enhanced CT images for identifying MVI-positive HCC remains unclear. Findings Quantitative indicators derived from fractal analysis were able to predict MVI. The developed model demonstrated improved performance when incorporating fractal dimension. Clinical relevance Fractal analysis based on contrast-enhanced CT is a feasible approach for evaluating MVI and provides additional clinical value for prognostic assessment. It may serve as a reference for preoperative MVI estimation and assist clinicians in executing more tailored therapies. Graphical