Machine Learning-Based CEMRI Radiomics Integrating LI-RADS Features Achieves Optimal Evaluation of Hepatocellular Carcinoma Differentiation

CEMRI-Based Quantification of Intratumoral Heterogeneity for Predicting Aggressive Characteristics of Hepatocellular Carcinoma Using Habitat Analysis: Comparison and Combination of Deep Learning

RATIONALE AND OBJECTIVES

To explore both an intratumoral heterogeneity (ITH) model based on habitat analysis and a deep learning (DL) model based on contrast-enhanced magnetic resonance imaging (CEMRI) and validate its efficiency for predicting microvascular invasion (MVI) and pathological differentiation in hepatocellular carcinoma (HCC).

METHODS

CEMRI images were retrospectively obtained from 277 HCCs in 265 patients. Habitat analysis and DL features were extracted from the CEMRI images and selected with the least absolute shrinkage and selection operator approach to develop ITH and DL models, respectively, and these robust features were then integrated to design a fusion model for predicting MVI and poorly differentiated HCC (pHCC). The predictive value of the three models was assessed using the area under the receiver operating characteristic curve (AUC).

RESULTS

The training and validation sets comprised 221 HCCs and 56 HCCs, respectively. The ITH and DL models presented AUC values of (0.90 vs. 0.87) for predicting MVI in the training set, with AUC values of 0.86 and 0.83 in the validation set. The AUC values of the ITH model to predict pHCC were 0.90 and 0.86 in the two sets, respectively; they were 0.84 and 0.80 for the DL model. The fusion model yielded the best performance for predicting MVI and pHCC in the training set (AUC=0.95, 0.90) and in the validation set (AUC=0.89, 0.87), respectively.

CONCLUSION

A fusion model integrating ITH and DL features derived from CEMRI images can serve as an excellent imaging biomarker for predicting aggressive characteristics in HCC.

Machine learning-based model for predicting tumor recurrence after interventional therapy in HBV-related hepatocellular carcinoma patients with low preoperative platelet-albumin-bilirubin score

Introduction

This study aimed to develop a prognostic nomogram for predicting the recurrence-free survival (RFS) of hepatitis B virus (HBV)-related hepatocellular carcinoma (HCC) patients with low preoperative platelet-albumin-bilirubin (PALBI) scores after transarterial chemoembolization (TACE) combined with local ablation treatment.

Methods

We gathered clinical data from 632 HBV-related HCC patients who received the combination treatment at Beijing You'an Hospital, affiliated with Capital Medical University, from January 2014 to January 2020. The patients were divided into two groups based on their PALBI scores: low PALBI group (n=247) and high PALBI group (n=385). The low PALBI group was then divided into two cohorts: training cohort (n=172) and validation cohort (n=75). We utilized eXtreme Gradient Boosting (XGBoost), random survival forest (RSF), and multivariate Cox analysis to pinpoint the risk factors for RFS. Then, we developed a nomogram based on the screened factors and assessed its risk stratification capabilities and predictive performance.

Results

The study finally identified age, aspartate aminotransferase (AST), and prothrombin time activity (PTA) as key predictors. The three variables were included to develop the nomogram for predicting the 1-, 3-, and 5-year RFS of HCC patients. We confirmed the nomogram's ability to effectively discern high and low risk patients, as evidenced by Kaplan-Meier curves. We further corroborated the excellent discrimination, consistency, and clinical utility of the nomogram through assessments using the C-index, area under the curve (AUC), calibration curve, and decision curve analysis (DCA).

Conclusion

Our study successfully constructed a robust nomogram, effectively predicting 1-, 3-, and 5-year RFS for HBV-related HCC patients with low preoperative PALBI scores after TACE combined with local ablation therapy.

Multiparametric MRI-based whole-liver radiomics for predicting early-stage liver fibrosis in rabbits

OBJECTIVES

To develop and validate a whole-liver radiomic model using multiparametric MRI for predicting early-stage liver fibrosis (LF) in rabbits.

METHODS

A total of 134 rabbits (early-stage LF, n = 91; advanced-stage LF, n = 43) who underwent liver magnetic resonance elastography (MRE), hepatobiliary phase, dynamic contrast enhanced (DCE), intravoxel incoherent motion (IVIM), diffusion kurtosis imaging, and T2* scanning were enrolled and randomly allocated to either the training or validation cohort. Whole-liver radiomic features were extracted and selected to develop a radiomic model and generate quantitative Rad-scores. Then, multivariable logistic regression was utilized to determine the Rad-scores associated with early-stage LF, and effective features were integrated to establish a combined model. The predictive performance was assessed by the area under the curve (AUC).

RESULTS

The MRE model achieved superior AUCs of 0.95 in the training cohort and 0.86 in the validation cohort, followed by the DCE-MRI model (0.93 and 0.82), while the IVIM model had lower AUC values of 0.91 and 0.82, respectively. The Rad-scores of MRE, DCE-MRI and IVIM were identified as independent predictors associated with early-stage LF. The combined model demonstrated AUC values of 0.96 and 0.88 for predicting early-stage LF in the training and validation cohorts, respectively.

CONCLUSIONS

Our study highlights the remarkable performance of a multiparametric MRI-based radiomic model for the individualized diagnosis of early-stage LF.

ADVANCES IN KNOWLEDGE

This is the first study to develop a combined model by integrating multiparametric radiomic features to improve the accuracy of LF staging.