EUVIMEDEUVIMED entdecken
Crossref

CMHF-3DNet: A Transformer-Based Framework for Improved Brain Tumor Segmentation Across Modalities

Abid Hussain, Wu Jungshen, Ali Turab, Yasir Gilani, Kazi Istiaque Ahmed, Ahsan Wajahat

Academic Radiology · 2026-09 · 33 (9) · 4042-4059

DOI: 10.1016/j.acra.2026.04.045 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Academic Radiology, 2026-09. Verfasst von Abid Hussain und 5 weiteren Autor:innen. Beteiligte Einrichtung: School of Software, Northwestern Polytechnical University, Xi'an 710000, China (A.H., W.J., A.T., Y.G., I.A., A.W.). Electronic address: abidhussain@…. Hauptthemen laut MEDLINE: Brain Neoplasms, Imaging, Three-Dimensional, Magnetic Resonance Imaging, Image Interpretation, Computer-Assisted. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 50 Literaturangaben.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

BACKGROUND: Accurate segmentation of brain tumor subregions in multi-modal MRI is essential for diagnosis, treatment planning, and longitudinal monitoring. However, delineating heterogeneous regions such as the enhancing tumor (ET) and tumor core remains challenging due to weak boundaries and substantial appearance variability across imaging modalities and scanners. METHODS: To address these challenges, we propose Cross-Modal Hierarchical Fusion U-Net 3D (CMHF-3DNet), a 3D encoder-decoder framework that performs voxel-wise cross-modal transformer fusion to explicitly model inter-modal dependencies and employs hierarchy-aware multi-task learning to enforce anatomical consistency constraints (ET ⊂ TC ⊂ WT). RESULT: The proposed method was evaluated on the BraTS 2023, BraTS 2024, and BraTS 2025 validation datasets. CMHF-3DNet achieved a mean Dice Similarity Coefficient (DSC) of 0.8527 on BraTS 2025 and 0.8526 on BraTS 2023, along with a mean 95th percentile Hausdorff Distance (HD95) of 6.47 mm on BraTS 2025. CONCLUSION: These results indicate improved boundary delineation for clinically relevant tumor subregions, particularly the ET and tumor core. Overall, CMHF-3DNet demonstrates robust performance across multiple BraTS benchmarks and suggests potential utility in supporting automated brain tumor assessment.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Brain NeoplasmsImaging, Three-DimensionalMagnetic Resonance ImagingImage Interpretation, Computer-AssistedHumansAlgorithmsMultimodal ImagingReproducibility of Results

Stichwörter der Autor:innen

3D U-NetBraTS benchmarkBrain tumor segmentationCross-modal transformerHierarchy aware learningMedical image segmentationMulti-modal MRI

Autor:innen und Institutionen

  • School of Software, Northwestern Polytechnical University, Xi'an 710000, China (A.H., W.J., A.T., Y.G., I.A., A.W.). Electronic address: abidhussain@nwpu.edu.cn.
  • School of Software, Northwestern Polytechnical University, Xi'an 710000, China (A.H., W.J., A.T., Y.G., I.A., A.W.).

Zitieren

APA

Hussain, A., Jungshen, W., Turab, A., Gilani, Y., Ahmed, K. I., & Wajahat, A. (2026). CMHF-3DNet: A Transformer-Based Framework for Improved Brain Tumor Segmentation Across Modalities. Academic Radiology, 33(9), 4042-4059. https://doi.org/10.1016/j.acra.2026.04.045

Vancouver

Hussain A, Jungshen W, Turab A, Gilani Y, Ahmed KI, Wajahat A. CMHF-3DNet: A Transformer-Based Framework for Improved Brain Tumor Segmentation Across Modalities. Academic Radiology. 2026;33(9):4042-4059. doi:10.1016/j.acra.2026.04.045

BibTeX
@article{Hussain2026,
  author = {Hussain, Abid and Jungshen, Wu and Turab, Ali and Gilani, Yasir and Ahmed, Kazi Istiaque and Wajahat, Ahsan},
  title = {CMHF-3DNet: A Transformer-Based Framework for Improved Brain Tumor Segmentation Across Modalities},
  journal = {Academic Radiology},
  year = {2026},
  volume = {33},
  number = {9},
  pages = {4042--4059},
  doi = {10.1016/j.acra.2026.04.045}
}
Kostenlos starten →