EUVIMEDEUVIMED entdecken
Crossref

Differentiating advanced from non-advanced hepatic fibrosis: a hybrid deep learning-radiomics model leveraging synthetic contrast-enhanced CT from CycleGAN

Jian He, Yue Pan, Menghua Yang, Na Hu, Fangyan Li, Lin Zhao, Yangping Xiao, Jifu Long, Pinggui Lei

Abdominal Radiology · 04.06.2026 · 51 (12) · 6484-6495

DOI: 10.1007/s00261-026-05563-5 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Abdominal Radiology, online erschienen am 04.06.2026. Verfasst von Jian He und 8 weiteren Autor:innen. Beteiligte Einrichtung: Department of Radiology, Affiliated Hospital of Guizhou Medical University, Guiyang, China. Hauptthemen laut MEDLINE: Liver Cirrhosis, Deep Learning, Tomography, X-Ray Computed, Contrast Media. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 44 Literaturangaben. Gefördert von: Guizhou Provincial Key Laboratory for Digestive System Diseases.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

OBJECTIVES: This study aimed to develop and validate a hybrid deep learning-radiomics model that leveraged Cycle-consistent generative adversarial networks (CycleGAN)-synthesized contrast-enhanced computed tomography (CE-CT) images to differentiate advanced from non-advanced hepatic fibrosis. METHODS: This retrospective study included 410 patients with biopsy-confirmed hepatic fibrosis (2017-2024). A trained CycleGAN model was used to generate synthetic three-phase CE-CT images from the corresponding non-contrast computed tomography (NC-CT) data. Each group of images was randomly split 6:4 ratio into training and test sets. After region of interest (ROI) segmentation, handcrafted radiomic (HCR) features were extracted. Concurrently, eight end-to-end deep learning (DL) models were trained; DL features were extracted from the best-performing model. Feature selection was performed using Spearman's rank correlation and the least absolute shrinkage and selection operator (LASSO). Six machine learning classifiers were developed for each feature type (HCR, DL, and late-fused DL features) using the final selected feature set. The performance of models was assessed by the area under the receiver operating characteristic curve (AUC), accuracy, calibration curves, decision curve analysis (DCA) and the DeLong test. RESULTS: Models utilizing synthetic CE-CT images outperformed those based on NC-CT. DL feature-based models surpassed HCR-based models. A late-fusion hybrid model integrating DL features further improved performance, achieving an AUC of 0.880 (95% CI: 0.819-0.942). CONCLUSIONS: The model based on synthetic CE-CT images demonstrated excellent diagnostic performance. Moreover, the hybrid model combining both real NC-CT and synthetic CE-CT images further improved diagnostic performance. The hybrid model can serve as a non-invasive diagnostic method for differentiating advanced from non-advanced hepatic fibrosis.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Liver CirrhosisDeep LearningTomography, X-Ray ComputedContrast MediaRadiographic Image Interpretation, Computer-AssistedHumansRadiomicsGenerative Adversarial NetworksRetrospective StudiesDiagnosis, DifferentialMaleFemaleMiddle AgedGenerative Artificial Intelligence

Stichwörter der Autor:innen

Computed tomographyCycle-consistent generative adversarial networksDeep learningHepatic fibrosisRadiomics

Substanzen

Contrast Media

Autor:innen und Institutionen

  • Department of Radiology, Affiliated Hospital of Guizhou Medical University, Guiyang, China.
  • Department of Radiology, Affiliated Hospital of Guizhou Medical University, Guiyang, China. pingguilei@foxmail.com.
  • Guizhou Provincial Key Laboratory for Digestive System Diseases, Guiyang, China. pingguilei@foxmail.com.

Förderung

  • Guizhou Provincial Key Laboratory for Digestive System Diseases

Zitieren

APA

He, J., Pan, Y., Yang, M., Hu, N., Li, F., Zhao, L., Xiao, Y., Long, J., & Lei, P. (2026). Differentiating advanced from non-advanced hepatic fibrosis: a hybrid deep learning-radiomics model leveraging synthetic contrast-enhanced CT from CycleGAN. Abdominal Radiology, 51(12), 6484-6495. https://doi.org/10.1007/s00261-026-05563-5

Vancouver

He J, Pan Y, Yang M, Hu N, Li F, Zhao L, et al. Differentiating advanced from non-advanced hepatic fibrosis: a hybrid deep learning-radiomics model leveraging synthetic contrast-enhanced CT from CycleGAN. Abdominal Radiology. 2026;51(12):6484-6495. doi:10.1007/s00261-026-05563-5

BibTeX
@article{He2026,
  author = {He, Jian and Pan, Yue and Yang, Menghua and Hu, Na and Li, Fangyan and Zhao, Lin and Xiao, Yangping and Long, Jifu and Lei, Pinggui},
  title = {Differentiating advanced from non-advanced hepatic fibrosis: a hybrid deep learning-radiomics model leveraging synthetic contrast-enhanced CT from CycleGAN},
  journal = {Abdominal Radiology},
  year = {2026},
  volume = {51},
  number = {12},
  pages = {6484--6495},
  doi = {10.1007/s00261-026-05563-5}
}
Kostenlos starten →