Automated segmentation of esophageal squamous cell carcinoma in contrast-enhanced free-breathing 3D-GRE: a comparative study of UNet, nnUNet, and UMamba for tumor delineation
UMamba demonstrates superior performance for automated ESCC segmentation on high-resolution MRI, providing a robust tool that has the potential to reduce manual segmentation time and inter-observer variability in clinical practice.
Authors
- Jinrong Qu (ORCID: https://orcid.org/0000-0001-8751-9988)
- Chenxin Hu
- Bingmei Bai
- Shuting Wang (ORCID: https://orcid.org/0000-0003-1461-3898)
- Zhaoqi Wang (ORCID: https://orcid.org/0000-0003-0894-8165)
- Liyi Shi (ORCID: https://orcid.org/0000-0003-1551-3768)
- Binbin Han
- Mengzhu Wang (ORCID: https://orcid.org/0009-0004-9431-1383)
- Yanyan Zhao (ORCID: https://orcid.org/0000-0001-9865-2741)
- Funing Chu
- Zaiyi Liu (ORCID: https://orcid.org/0000-0002-9307-8522)
- Keke Zhao
Institutions
- Guangdong General Hospital (CN)
- Zhengzhou University (CN)
- Henan Cancer Hospital (CN)
- Guangdong Academy of Medical Sciences (CN)
- Siemens Healthcare (United States) (US)
Publication Details
- Journal
- BMC Medical Imaging
- Published
- 2025-11-11
- DOI
- https://doi.org/10.1186/s12880-025-02011-6
- Primary Topic
- Esophageal Cancer Research and Treatment
- Type
- article
- Field-Weighted Citation Impact
- 0.00
Funders
- National Natural Science Foundation of China