CT-based radiomics combined with clinical features for non-invasive differentiation of hepatitis B and schistosomiasis cirrhosis: a multicenter study
10 affiliations
- 1Department of Nuclear Medicine, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, No.1277 Jiefang Avenue, Wuhan 430022 Hubei, P. R. China
- 2Department of Radiology, Chinese PLA General Hospital of Central Theater Command, Wuhan, China
- 3Department of Radiology, The People’s Hospital of Honghu, Honghu, China
- 4Hubei Provincial Clinical Medical Research Center for Medical Imaging, Wuhan, Hubei China
- 5Department of Radiology, The First Affiliated Hospital of USTC (Anhui Provincial Hospital), Hefei, China
- 6Key Laboratory of Biological Targeted Therapy, the Ministry of Education, Wuhan, China
- 7Department of Radiology, Honghu Fourth People’s Hospital, Honghu, China
- 8Regenerative Medicine Research Center, West China Hospital, Chengdu, China
- 9Department of Ultrasound, The People’s Hospital of Honghu, Honghu, China
- 10Department of Radiology, Union Hospital, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, China
Abstract
Background: Accurate etiological differentiation between hepatitis B virus (HBV)-related cirrhosis and schistosomiasis-related cirrhosis remains clinically important in endemic regions because these conditions differ substantially in their underlying pathogenesis, clinical management, and prognosis. A non-invasive imaging approach may provide complementary information for etiological assessment. To develop a CT-based clinical-radiomics model for the non-invasive differentiation of HBV-related and schistosomiasis-related cirrhosis and to evaluate its diagnostic performance in a multicenter cohort.
Methods: This retrospective multicenter study included 143 patients with HBV-related or schistosomiasis-related cirrhosis from three institutions between August 2019 and December 2021. Radiomics features were extracted from non-contrast CT images, and a radiomics signature (Rad-score) was established using minimum redundancy maximum relevance (mRMR) and least absolute shrinkage and selection operator (LASSO) regression. Clinical and integrated clinical-radiomics models were subsequently developed using multivariable logistic regression. Model performance was evaluated using receiver operating characteristic analysis, calibration curves, and decision curve analysis (DCA).
Results: The Rad-score was significantly lower in patients with HBV-related cirrhosis than in those with schistosomiasis-related cirrhosis (both P < 0.001 in the training and internal validation cohorts). The integrated clinical-radiomics model demonstrated excellent diagnostic performance, with area under the receiver operating characteristic curve (AUC) values of 0.990 in the training cohort, 0.920 in the internal validation cohort, and 0.782 in the external validation cohort. Compared with the clinical model alone, both the radiomics and integrated models demonstrated more consistent performance across the internal and external validation cohorts, while the integrated model showed the greatest overall clinical utility on DCA.
Conclusion: The proposed CT-based clinical-radiomics model provides a promising non-invasive approach for differentiating HBV-related and schistosomiasis-related cirrhosis. As a complementary imaging tool, it may assist etiological assessment in endemic regions when interpreted alongside routine clinical and laboratory evaluation. Further prospective multicenter studies are warranted to validate its clinical utility.
Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1186/s12880-026-02693-6.
Related
- Global, regional, and national burden of chronic hepatitis C including cirrhosis from 1990 to 2021, alongside projections for the next 15 years.
- Sofosbuvir/Velpatasvir in Chronic Hepatitis C, Compensated and Decompensated Cirrhosis, and Retreatment Settings
- Ultrasound-based deep learning radiomics with explainable machine learning for predicting benign and malignant thyroid nodules in lymphocytic thyroiditis: a multicenter study
- Diagnostic performance of liver biochemical and hemostatic markers for assessing cirrhosis severity in chronic hepatitis B
- Mortality trends in alcohol-related liver disease and chronic hepatitis and fibrosis or cirrhosis not coded as alcohol-related: a population-based analysis of US CDC WONDER data, 1999–2020
BibTeX
@article{tong2026ctbased,
title = {{CT-based radiomics combined with clinical features for non-invasive differentiation of hepatitis B and schistosomiasis cirrhosis: a multicenter study}},
author = {Junjie Tong and Xiaojun Xu and Fan Hu and Zhengquan Wang and Ling Ma and Jieming Li and Xin Li and Zhong Yang and Xiaotian Xia},
journal = {BMC Medical Imaging},
year = {2026},
month = aug,
doi = {10.1186/s12880-026-02693-6},
url = {https://doi.org/10.1186/s12880-026-02693-6},
note = {CC BY-NC-ND 4.0},
}
RIS
TY - JOUR
TI - CT-based radiomics combined with clinical features for non-invasive differentiation of hepatitis B and schistosomiasis cirrhosis: a multicenter study
AU - Junjie Tong
AU - Xiaojun Xu
AU - Fan Hu
AU - Zhengquan Wang
AU - Ling Ma
AU - Jieming Li
AU - Xin Li
AU - Zhong Yang
AU - Xiaotian Xia
PY - 2026
DA - 2026/08/31
JO - BMC Medical Imaging
DO - 10.1186/s12880-026-02693-6
UR - https://doi.org/10.1186/s12880-026-02693-6
AB - Background: Accurate etiological differentiation between hepatitis B virus (HBV)-related cirrhosis and schistosomiasis-related cirrhosis remains clinically important in endemic regions because these conditions differ substantially in their underlying pathogenesis, clinical management, and prognosis. A non-invasive imaging approach may provide complementary information for etiological assessment. To develop a CT-based clinical-radiomics model for the non-invasive differentiation of HBV-related and schistosomiasis-related cirrhosis and to evaluate its diagnostic performance in a multicenter cohort. Methods: This retrospective multicenter study included 143 patients with HBV-related or schistosomiasis-related cirrhosis from three institutions between August 2019 and December 2021. Radiomics features were extracted from non-contrast CT images, and a radiomics signature (Rad-score) was established using minimum redundancy maximum relevance (mRMR) and least absolute shrinkage and selection operator (LASSO) regression. Clinical and integrated clinical-radiomics models were subsequently developed using multivariable logistic regression. Model performance was evaluated using receiver operating characteristic analysis, calibration curves, and decision curve analysis (DCA). Results: The Rad-score was significantly lower in patients with HBV-related cirrhosis than in those with schistosomiasis-related cirrhosis (both P < 0.001 in the training and internal validation cohorts). The integrated clinical-radiomics model demonstrated excellent diagnostic performance, with area under the receiver operating characteristic curve (AUC) values of 0.990 in the training cohort, 0.920 in the internal validation cohort, and 0.782 in the external validation cohort. Compared with the clinical model alone, both the radiomics and integrated models demonstrated more consistent performance across the internal and external validation cohorts, while the integrated model showed the greatest overall clinical utility on DCA. Conclusion: The proposed CT-based clinical-radiomics model provides a promising non-invasive approach for differentiating HBV-related and schistosomiasis-related cirrhosis. As a complementary imaging tool, it may assist etiological assessment in endemic regions when interpreted alongside routine clinical and laboratory evaluation. Further prospective multicenter studies are warranted to validate its clinical utility. Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1186/s12880-026-02693-6.
SN - 1471-2342
ER -