Prediction of Early Recurrence After Thermal Ablation for Hepatocellular Carcinoma Using Contrast-Enhanced CT and Habitat Analysis
10.3969/j.issn.1005-5185.2025.09.006
- VernacularTitle:基于增强CT联合生境分析预测肝细胞癌热消融术后早期复发
- Author:
Yanfen ZHAO
1
;
Zhu LIU
;
Xiaoqin WEI
;
Yong DU
Author Information
1. 川北医学院附属医院放射科,四川 南充 637000
- Publication Type:Journal Article
- Keywords:
Carcinoma,hepatocellular;
Tomography,X-ray computed;
Thermal ablation;
Recurrence;
Habitat imaging;
Nomograms;
Machine learning;
Forecasting
- From:
Chinese Journal of Medical Imaging
2025;33(9):929-935,947
- CountryChina
- Language:Chinese
-
Abstract:
Purpose To develop a nomogram based on contrast-enhanced CT and habitat analysis for predicting early recurrence after thermal ablation in hepatocellular carcinoma,enabling risk stratification and personalized patient management.Materials and Methods This retrospective study included 107 patients with hepatocellular carcinoma treated with thermal ablation from Affiliated Hospital of North Sichuan Medical College and Nanchong Central Hospital from January 2016 to March 2022.Tumor lesions were manually segmented on preoperative contrast-enhanced CT images.Habitat encoding was performed on volumes of interest using K-means clustering.Radiomic features were extracted from each phase and habitat subregion.Machine learning algorithms were used to construct phase-specific models,with performance compared to select the optimal model.A combined model integrating optimal radiomic features and independent clinical risk factors was developed and evaluated.Results The Adaboost algorithm yielded the optimal model for the arterial phase,Naive Bayes for the venous phase,and MLP for the combined arterial-venous phase.The combined model demonstrated superior performance,achieving concordance indices of 0.711(training cohort)and 0.709(validation cohort)for predicting early recurrence.Significant differences in recurrence-free survival were observed between high-risk group and low-risk group(log-rank P<0.05).Conclusion Habitat imaging derived from contrast-enhanced CT effectively and noninvasively assesses recurrence-free survival after thermal ablation for hepatocellular carcinoma,demonstrating potential for guiding clinical treatment and decision-making.
