PCR : A Progressive Consistency Refinement Network in Medical Decision‐Making System
International Journal of Imaging Systems and Technology · 2026-11 · 36 (6)
DOI: 10.1002/ima.70489 ↗At a glance
Journal article in International Journal of Imaging Systems and Technology, published online on 2026-11-17. Written by Jiaoyi Fang and 4 co-authors. 55 references. Funded by: National Natural Science Foundation of China.
Abstract (original language)
Deep learning has markedly advanced automated segmentation of solid tumors, yet glioblastoma multiforme (GBM) remains exceptionally challenging due to indistinct and heterogeneous tumor boundaries, misalignment between semantic and spatial feature representations, and cross‐scale conflicts in hierarchical feature fusion. To address these issues, we propose a Progressive Consistency Refinement Network (PCR) that formulates GBM segmentation as a three‐stage progressive refinement process. First, PCR establishes semantic‐spatial consistency by amplifying tumor‐specific semantic activations while preserving fine‐grained spatial details through residual connections. Second, it calibrates scale consistency. High‐level tumor semantics are used to guide low‐level spatial details, and hierarchical representations are aligned to reduce semantic and scale mismatch. Third, PCR refines boundary consistency on the calibrated representations. It decomposes infiltrative tumor margins from multiple directions and suppresses irrelevant background responses through lesion‐focused enhancement, thereby improving the delineation of weak and irregular tumor boundaries. We evaluate PCR on the BraTS2021 benchmark and an in‐house clinical dataset. The method achieves average Dice similarity coefficients of 88.79% and 85.40%, respectively, outperforming state‐of‐the‐art approaches while maintaining a low computational footprint. These results demonstrate that PCR effectively tackles the core difficulties in GBM MRI segmentation through its progressive consistency refinement strategy, offering an efficient and robust solution with strong potential for integration into clinical workflows.
Funding
- National Natural Science Foundation of China
Cite
APA
Fang, J., Gou, F., Zhou, J., Wang, Y., & Wu, J. (2026). PCR : A Progressive Consistency Refinement Network in Medical Decision‐Making System. International Journal of Imaging Systems and Technology, 36(6). https://doi.org/10.1002/ima.70489
Vancouver
Fang J, Gou F, Zhou J, Wang Y, Wu J. PCR : A Progressive Consistency Refinement Network in Medical Decision‐Making System. International Journal of Imaging Systems and Technology. 2026;36(6). doi:10.1002/ima.70489
BibTeX
@article{Fang2026,
author = {Fang, Jiaoyi and Gou, Fangfang and Zhou, Jing and Wang, Yisong and Wu, Jia},
title = {PCR : A Progressive Consistency Refinement Network in Medical Decision‐Making System},
journal = {International Journal of Imaging Systems and Technology},
year = {2026},
volume = {36},
number = {6},
doi = {10.1002/ima.70489}
}
EUVIMED