CT-Based Radiomics Features Combined with AFP for Predicting Vessels Encapsulating Tumor Clusters and Prognosis of Hepatocellular Carcinoma.

Abstract

Objective: This study aims to develop a CT-based radiomics nomogram for preoperative prediction of vessels encapsulating tumor clusters (VETC) and patient prognosis in hepatocellular carcinoma (HCC).

Patients and methods: The retrospective, single-center study included 231 (77 VETC+ and 154 VETC-) HCC patients who underwent preoperative CT scan, and were randomly divided into training and validation cohorts at a 7:3 ratio. Radiomics features were extracted from CT images during the plain, arterial and venous phases. These features were then selected using the Least Absolute Shrinkage and Selection Operator (LASSO). Predictive factors were chosen through univariate and multivariate logistic regression. A prognostic nomogram integrating clinical factor and radiomics features was developed and validated. The model's predictive accuracy was systematically evaluated using the area under the receiver operating characteristic curve (AUC), while calibration curves assessed agreement between predicted and observed outcomes. To quantify clinical utility, decision curve analysis (DCA) was implemented. Furthermore, the model's prognostic performance for postoperative disease-free survival (DFS) was examined through Kaplan-Meier analysis.

Results: The nomogram integrating four radiomics features and alpha-fetoprotein (AFP) exhibited robust predictive performance, with AUC values of 0.782 (95% confidence interval [CI]: 0.708-0.856) in the training cohort and 0.755 (95% CI: 0.628-0.882) in the validation cohort. Calibration curves demonstrated excellent agreement between predicted and observed outcomes in both cohorts. DCA revealed significant clinical utility of the nomogram. Additionally, the model-stratified VETC+ HCC patients showed significantly worse DFS compared to VETC- counterparts (log-rank p = 0.035).

Conclusion: The CT-based radiomics nomogram, integrating radiomics features and AFP, provides a non-invasive and reliable tool for predicting VETC and stratifying prognosis in HCC patients.