EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Illustration of transfer learning from breast cancer detection to risk prediction: adaptation to local data and local objectives

Tobias Wagner iD, Zan Klanecek iD, Yao-Kuan Wang iD, Lesley Cockmartin iD, Nicholas W Marshall iD, Andrej Studen iD, Robert Jeraj iD, Hilde Bosmans iD

Journal of Medical Imaging · 27.06.2026 · 13 (03) · 1-22

DOI: 10.1117/1.jmi.13.3.034504 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Medical Imaging, erschienen am 27.06.2026. Verfasst von Tobias Wagner und 7 weiteren Autor:innen. Beteiligte Einrichtung: KU Leuven, Division of Medical Physics & Quality Assessment, Department of Imaging and Pathology, Leuven, Belgium. Die Arbeit wurde bisher 1-mal zitiert (Crossref). Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

PurposeWe aim to investigate whether a breast cancer risk model can be trained with transfer learning from a breast cancer detection model.ApproachAn existing open-source breast cancer detection model was used to extract the latent space representation of a local dataset composed of images from 17,878 breast cancer screening participants. An autoencoder network was used for dimensionality reduction. A final model training step mapped the reduced representation to 4-year breast cancer risk. Two experimental studies were performed: first, a local dataset was used to investigate the impact of transfer learning to local data. Second, the training was executed for breast cancer risk by excluding cases with obvious signs of cancer in the model training process, to avoid the influence of cancer signs. In addition, the Mirai model was evaluated.ResultsThe open-source model achieved a baseline performance of 0.72 with 95%CI [0.66, 0.78] for 4-year risk prediction when applied to local test cases of breast cancer risk data. This performance was improved after training on local breast cancer detection data to 0.75 with 95%CI [0.69, 0.81], although not statistically significant (p=0.09). Further training with local breast cancer risk data achieved an area under the curve (AUC) of 0.77 with 95%CI [0.71, 0.82], which is higher than the original model (p

Schlagwörter

Stichwörter der Autor:innen

breast cancerdeep learningmammographyrisk assessmenttransfer learning

Autor:innen und Institutionen

  • KU Leuven, Division of Medical Physics & Quality Assessment, Department of Imaging and Pathology, Leuven, Belgium.
  • University of Ljubljana, Faculty of Mathematics and Physics, Ljubljana, Slovenia.
  • UZ Leuven, Department of Radiology, Leuven, Belgium.
  • Jožef Stefan Institute, Ljubljana, Slovenia.
  • University of Wisconsin-Madison, Department of Medical Physics, Madison, Wisconsin, United States.

ORCID-Profile: Tobias Wagner, Zan Klanecek, Yao-Kuan Wang, Lesley Cockmartin, Nicholas W Marshall, Andrej Studen, Robert Jeraj, Hilde Bosmans

Zitieren

APA

Wagner, T., Klanecek, Z., Wang, Y. K., Cockmartin, L., Marshall, N. W., Studen, A., Jeraj, R., & Bosmans, H. (2026). Illustration of transfer learning from breast cancer detection to risk prediction: adaptation to local data and local objectives. Journal of Medical Imaging, 13(03), 1-22. https://doi.org/10.1117/1.jmi.13.3.034504

Vancouver

Wagner T, Klanecek Z, Wang YK, Cockmartin L, Marshall NW, Studen A, et al. Illustration of transfer learning from breast cancer detection to risk prediction: adaptation to local data and local objectives. Journal of Medical Imaging. 2026;13(03):1-22. doi:10.1117/1.jmi.13.3.034504

BibTeX
@article{Wagner2026,
  author = {Wagner, Tobias and Klanecek, Zan and Wang, Yao-Kuan and Cockmartin, Lesley and Marshall, Nicholas W and Studen, Andrej and Jeraj, Robert and Bosmans, Hilde},
  title = {Illustration of transfer learning from breast cancer detection to risk prediction: adaptation to local data and local objectives},
  journal = {Journal of Medical Imaging},
  year = {2026},
  volume = {13},
  number = {03},
  pages = {1--22},
  doi = {10.1117/1.jmi.13.3.034504}
}
Kostenlos starten →