EUVIMEDEUVIMED entdecken
Crossref

Deep transfer learning radiomics combined with explainable machine learning for preoperative thymoma risk prediction based on CT

Shujian Wu, Lifang Fan, Yimin Wu, Jingya Xu, Yong Guo, Hu Zhang, Zhengyuan Xu

European Journal of Radiology · 2025

DOI: 10.1016/j.ejrad.2025.112266 ↗

Kurz gefasst

Fachartikel in der Zeitschrift European Journal of Radiology, 2025. Verfasst von Shujian Wu und 6 weiteren Autor:innen. Beteiligte Einrichtung: Department of Radiology, The First Affiliated Hospital of Wannan Medical College, Wuhu, Anhui 241001, China. Hauptthemen laut MEDLINE: Thymoma, Tomography, X-Ray Computed, Thymus Neoplasms, Deep Learning. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 35 Literaturangaben. Gefördert von: Wannan Medical College, Scientific Research Foundation of Education Department of Anhui Province of China.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

OBJECTIVE: To develop and validate a computerized tomography (CT)‑based deep transfer learning radiomics model combined with explainable machine learning for preoperative risk prediction of thymoma. METHODS: This retrospective study included 173 pathologically confirmed thymoma patients from our institution in the training group and 93 patients from two external centers in the external validation group. Tumors were classified according to the World Health Organization simplified criteria as low‑risk types (A, AB, and B1) or high‑risk types (B2 and B3). Radiomics features and deep transfer learning features were extracted from venous‑phase contrast‑enhanced CT images by using a modified Inception V3 network. Principal component analysis and least absolute shrinkage and selection operator regression identified 20 key predictors. Six classifiers-decision tree, gradient boosting machine, k‑nearest neighbors, naïve Bayes, random forest (RF), and support vector machine-were trained on five feature sets: CT imaging model, radiomics feature model, deep transfer learning feature model, combined feature model, and combined model. Interpretability was assessed with SHapley Additive exPlanations (SHAP), and an interactive web application was developed for real‑time individualized risk prediction and visualization. RESULTS: In the external validation group, the RF classifier achieved the highest area under the receiver operating characteristic curve (AUC) value of 0.956. In the training group, the AUC values for the CT imaging model, radiomics feature model, deep transfer learning feature model, combined feature model, and combined model were 0.684, 0.831, 0.815, 0.893, and 0.910, respectively. The corresponding AUC values in the external validation group were 0.604, 0.865, 0.880, 0.934, and 0.956, respectively. SHAP visualizations revealed the relative contribution of each feature, while the web application provided real‑time individual prediction probabilities with interpretative outputs. CONCLUSION: We developed a CT‑based deep transfer learning radiomics model combined with explainable machine learning and an interactive web application; this model achieved high accuracy and transparency for preoperative thymoma risk stratification, facilitating personalized clinical decision‑making.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

ThymomaTomography, X-Ray ComputedThymus NeoplasmsDeep LearningMachine LearningHumansMaleFemaleMiddle AgedRetrospective StudiesAdultRisk AssessmentAgedRadiographic Image Interpretation, Computer-AssistedRadiomics

Stichwörter der Autor:innen

Computed tomographyDeep transfer learningMachine learningRadiomicsThymoma

Autor:innen und Institutionen

  • Department of Radiology, The First Affiliated Hospital of Wannan Medical College, Wuhu, Anhui 241001, China.
  • School of Medical Imageology, Wannan Medical College, Wuhu, Anhui 241002, China.
  • Department of Ultrasound, WuHu Hospital, East China Normal University (The Second People's Hospital, WuHu), Wuhu, Anhui 241001, China.
  • Department of Radiology, WuHu Hospital, East China Normal University (The Second People's Hospital, WuHu), Wuhu, Anhui 241001, China.
  • Department of Radiology, WuHu Hospital, East China Normal University (The Second People's Hospital, WuHu), Wuhu, Anhui 241001, China. Electronic address: m18055373782@163.com.
  • School of Medical Imageology, Wannan Medical College, Wuhu, Anhui 241002, China. Electronic address: xuzy@wnmc.edu.cn.

Förderung

  • Wannan Medical College
  • Scientific Research Foundation of Education Department of Anhui Province of China

Zitieren

APA

Wu, S., Fan, L., Wu, Y., Xu, J., Guo, Y., Zhang, H., & Xu, Z. (2025). Deep transfer learning radiomics combined with explainable machine learning for preoperative thymoma risk prediction based on CT. European Journal of Radiology. https://doi.org/10.1016/j.ejrad.2025.112266

Vancouver

Wu S, Fan L, Wu Y, Xu J, Guo Y, Zhang H, et al. Deep transfer learning radiomics combined with explainable machine learning for preoperative thymoma risk prediction based on CT. European Journal of Radiology. 2025. doi:10.1016/j.ejrad.2025.112266

BibTeX
@article{Wu2025,
  author = {Wu, Shujian and Fan, Lifang and Wu, Yimin and Xu, Jingya and Guo, Yong and Zhang, Hu and Xu, Zhengyuan},
  title = {Deep transfer learning radiomics combined with explainable machine learning for preoperative thymoma risk prediction based on CT},
  journal = {European Journal of Radiology},
  year = {2025},
  doi = {10.1016/j.ejrad.2025.112266}
}
Kostenlos starten →