Cover Image

Improving risk assessment of local failure in brain metastases patients using vision transformers: a multicentric development and validation study

Background and purpose - This study investigates the use of Vision Transformers (ViTs) to predict Freedom from Local Failure (FFLF) in patients with brain metastases using pre-operative MRI scans. The goal is to develop a model that enhances risk stratification and informs personalized treatment str...

Full description

Saved in:
Bibliographic Details
Main Authors: Erdur, Ayhan Can (Author) , Scholz, Daniel (Author) , Nguyen, Q. Mai (Author) , Buchner, Josef (Author) , Mayinger, Michael Christian (Author) , Christ, Sebastian M. (Author) , Brunner, Thomas B. (Author) , Wittig, Andrea (Author) , Zimmer, Claus (Author) , Meyer, Bernhard (Author) , Guckenberger, Matthias (Author) , Andratschke, Nicolaus (Author) , El-Shafie, Rami (Author) , Debus, Jürgen (Author) , Rogers, Susanne (Author) , Riesterer, Oliver (Author) , Schulze, Katrin (Author) , Feldmann, Horst J. (Author) , Blanck, Oliver (Author) , Zamboglou, Constantinos (Author) , Bilger-Zähringer, Angelika (Author) , Grosu, Anca-Ligia (Author) , Wolff, Robert Jürgen (Author) , Eitz, Kerstin Anne (Author) , Combs, Stephanie (Author) , Bernhardt, Denise (Author) , Wiestler, Benedikt (Author) , Rückert, Daniel (Author) , Peeken, Jan Caspar (Author)
Format: Article (Journal)
Language:English
Published: 5 July 2025
In: Radiotherapy and oncology
Year: 2025, Volume: 210, Pages: 1-9
ISSN:1879-0887
DOI:10.1016/j.radonc.2025.111031
Online Access:Verlag, kostenfrei, Volltext: https://doi.org/10.1016/j.radonc.2025.111031
Verlag, kostenfrei, Volltext: https://www.sciencedirect.com/science/article/pii/S0167814025045359
Get full text
Author Notes:Ayhan Can Erdur, Daniel Scholz, Q. Mai Nguyen, Josef A. Buchner, Michael Mayinger, Sebastian M. Christ, Thomas B. Brunner, Andrea Wittig, Claus Zimmer, Bernhard Meyer, Matthias Guckenberger, Nicolaus Andratschke, Rami A. El Shafie, Jürgen Debus, Susanne Rogers, Oliver Riesterer, Katrin Schulze, Horst J. Feldmann, Oliver Blanck, Constantinos Zamboglou, Angelika Bilger-Z., Anca L. Grosu, Robert Wolff, Kerstin A. Eitz, Stephanie E. Combs, Denise Bernhardt, Benedikt Wiestler, Daniel Rueckert, Jan C. Peeken
Description
Summary:Background and purpose - This study investigates the use of Vision Transformers (ViTs) to predict Freedom from Local Failure (FFLF) in patients with brain metastases using pre-operative MRI scans. The goal is to develop a model that enhances risk stratification and informs personalized treatment strategies. - Materials and methods - Within the AURORA retrospective trial, patients (n = 352) who received surgical resection followed by post-operative stereotactic radiotherapy (SRT) were collected from seven hospitals. We trained our ViT for the direct image-to-risk task on T1-CE and FLAIR sequences and combined clinical features along the way. We employed segmentation-guided image modifications, model adaptations, and specialized patient sampling strategies during training. The model was evaluated with five-fold cross-validation and ensemble learning across all validation runs. An external, international test cohort (n = 99) within the dataset was used to assess the generalization capabilities of the model, and saliency maps were generated for explainability analysis. - Results - We achieved a competent C-Index score of 0.7982 on the test cohort, surpassing all clinical, CNN-based, and hybrid baselines. Kaplan-Meier analysis showed significant FFLF risk stratification. Saliency maps focusing on the BM core confirmed that model explanations aligned with expert observations. - Conclusion - Our ViT-based model offers a potential for personalized treatment strategies and follow-up regimens in patients with brain metastases. It provides an alternative to radiomics as a robust, automated tool for clinical workflows, capable of improving patient outcomes through effective risk assessment and stratification.
Item Description:Gesehen am 17.12.2025
Physical Description:Online Resource
ISSN:1879-0887
DOI:10.1016/j.radonc.2025.111031

Similar Items