An attention-based multimodal approach for predicting overall survival in patients with glioblastoma using clinical and imaging data
Polish Journal of Radiology · 09.10.2026 · 91 (1) · 484-493
DOI: 10.5114/pjr/220559 ↗Kurz gefasst
Fachartikel in der Zeitschrift Polish Journal of Radiology, online erschienen am 09.10.2026. Verfasst von Saman Mohammadpour und 5 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben.
Abstract (Originalsprache)
Purpose
To develop and evaluate an attention-based multimodal model for improving the prediction of overall survival in patients with glioblastoma by integrating clinical and imaging features. Material and methods Data from 358 glioblastoma patients were obtained from The Cancer Imaging Archive. The dataset included clinical data and three magnetic resonance imaging (MRI) sequences. Three categories of features – clinical, radiomics, and deep features – were used for survival prediction. Radiomics features were extracted using PyRadiomics, while deep features were extracted from MRI images using a Swin UNEt TRansformers. An attention-based multimodal predictive model was designed to integrate these three feature sets. The DeepHit framework was used for survival modeling. Model performance was evaluated using the concordance index (C-index) and the Brier score. In addition, accuracy, precision, recall, area under the curve (AUC), and F1-score were calculated to predict mortality risk at different time intervals.
Results
The proposed model achieved a C-index of 0.91 and a Brier score of 0.14. For identifying mortality risk within 6 months, the model achieved an accuracy of 0.93, precision of 0.87, recall of 0.92, AUC of 0.96, and an F1-score of 0.89. For mortality risk within 12 months, the corresponding values were 0.91, 0.87, 0.85, 0.92, and 0.88, respectively. For mortality risk within 24 months, the model achieved an accuracy of 0.75, precision of 0.71, recall of 0.70, AUC of 0.94, and an F1-score of 0.80.
Conclusions
The proposed model, leveraging a combined radiomics-based approach, demonstrated the capability to predict overall survival in patients with glioblastoma. However, to ensure generalizability, further validation using multicenter datasets is required.
Zitieren
APA
Mohammadpour, S., Emami, H., Rabiei, R., Roshanpoor, A., Tabatabaeefar, M., & Fariborz, F. (2026). An attention-based multimodal approach for predicting overall survival in patients with glioblastoma using clinical and imaging data. Polish Journal of Radiology, 91(1), 484-493. https://doi.org/10.5114/pjr/220559
Vancouver
Mohammadpour S, Emami H, Rabiei R, Roshanpoor A, Tabatabaeefar M, Fariborz F. An attention-based multimodal approach for predicting overall survival in patients with glioblastoma using clinical and imaging data. Polish Journal of Radiology. 2026;91(1):484-493. doi:10.5114/pjr/220559
BibTeX
@article{Mohammadpour2026,
author = {Mohammadpour, Saman and Emami, Hassan and Rabiei, Reza and Roshanpoor, Arash and Tabatabaeefar, Morteza and Fariborz, Faeghi},
title = {An attention-based multimodal approach for predicting overall survival in patients with glioblastoma using clinical and imaging data},
journal = {Polish Journal of Radiology},
year = {2026},
volume = {91},
number = {1},
pages = {484--493},
doi = {10.5114/pjr/220559}
}
EUVIMED