A Benchmark Framework for the Right Atrium Cavity Segmentation From LGE-MRIs

Jieyun Bai; Jinwen Zhu; Zhiting Chen; Ziduo Yang; Yaosheng Lu; Lei Li; Qince Li; Wei Wang; Henggui Zhang; Kuanquan Wang; Jie Gan; Jichao Zhao; Hua Lu; Suining Li; Jiawen Huang; Xiaoming Chen; Xiaoshen Zhang; Xiaowei Xu; Lulu Li; Yanfeng Tian; Víctor M. Campello; Karim Lekadir

doi:10.1109/TMI.2025.3590694

A Benchmark Framework for the Right Atrium Cavity Segmentation From LGE-MRIs

Jieyun Bai^*
, Jinwen Zhu
, Zhiting Chen
, Ziduo Yang
, Yaosheng Lu
, Lei Li
, Qince Li
, Wei Wang
, Henggui Zhang
, Kuanquan Wang
, Jie Gan
, Jichao Zhao
, Hua Lu^*
, Suining Li
, Jiawen Huang
, Xiaoming Chen
, Xiaoshen Zhang
, Xiaowei Xu
, Lulu Li
, Yanfeng Tian

Víctor M. Campello, Karim Lekadir

^*Corresponding author for this work

Jinan University
The University of Auckland
National University of Singapore
School of Computer Science and Technology, Harbin Institute of Technology
University of Manchester
University of Sydney
The First Affiliated Hospital of Jinan University
Division of Thoracic Surgery
The First Affiliated Hospital of Harbin Medical University
University of Barcelona

Research output: Contribution to journal › Article › peer-review

3 Scopus citations

Abstract

The right atrium (RA) is critical for cardiac hemodynamics but is often overlooked in clinical diagnostics. This study presents a benchmark framework for RA cavity segmentation from late gadolinium-enhanced magnetic resonance imaging (LGE-MRIs), leveraging a two-stage strategy and a novel 3D deep learning network, RASnet. The architecture addresses challenges in class imbalance and anatomical variability by incorporating multi-path input, multi-scale feature fusion modules, Vision Transformers, context interaction mechanisms, and deep supervision. Evaluated on datasets comprising 354 LGE-MRIs, RASnet achieves SOTA performance with a Dice score of 92.19% on a primary dataset and demonstrates robust generalizability on an independent dataset. The proposed framework establishes a benchmark for RA cavity segmentation, enabling accurate and efficient analysis for cardiac imaging applications.

Original language	English
Pages (from-to)	5290-5305
Number of pages	16
Journal	IEEE Transactions on Medical Imaging
Volume	44
Issue number	12
DOIs	https://doi.org/10.1109/TMI.2025.3590694
State	Published - 2025
Externally published	Yes

Keywords

Right atrium
cardiac segmentation
challenge
late gadolinium-enhanced magnetic resonance imaging
segment anything model

Access to Document

10.1109/TMI.2025.3590694

Cite this

Bai, J., Zhu, J., Chen, Z., Yang, Z., Lu, Y., Li, L., Li, Q., Wang, W., Zhang, H., Wang, K., Gan, J., Zhao, J., Lu, H., Li, S., Huang, J., Chen, X., Zhang, X., Xu, X., Li, L., ... Lekadir, K. (2025). A Benchmark Framework for the Right Atrium Cavity Segmentation From LGE-MRIs. IEEE Transactions on Medical Imaging, 44(12), 5290-5305. https://doi.org/10.1109/TMI.2025.3590694

@article{807be5a649fc464891728cae11107134,

title = "A Benchmark Framework for the Right Atrium Cavity Segmentation From LGE-MRIs",

abstract = "The right atrium (RA) is critical for cardiac hemodynamics but is often overlooked in clinical diagnostics. This study presents a benchmark framework for RA cavity segmentation from late gadolinium-enhanced magnetic resonance imaging (LGE-MRIs), leveraging a two-stage strategy and a novel 3D deep learning network, RASnet. The architecture addresses challenges in class imbalance and anatomical variability by incorporating multi-path input, multi-scale feature fusion modules, Vision Transformers, context interaction mechanisms, and deep supervision. Evaluated on datasets comprising 354 LGE-MRIs, RASnet achieves SOTA performance with a Dice score of 92.19\% on a primary dataset and demonstrates robust generalizability on an independent dataset. The proposed framework establishes a benchmark for RA cavity segmentation, enabling accurate and efficient analysis for cardiac imaging applications.",

keywords = "Right atrium, cardiac segmentation, challenge, late gadolinium-enhanced magnetic resonance imaging, segment anything model",

author = "Jieyun Bai and Jinwen Zhu and Zhiting Chen and Ziduo Yang and Yaosheng Lu and Lei Li and Qince Li and Wei Wang and Henggui Zhang and Kuanquan Wang and Jie Gan and Jichao Zhao and Hua Lu and Suining Li and Jiawen Huang and Xiaoming Chen and Xiaoshen Zhang and Xiaowei Xu and Lulu Li and Yanfeng Tian and Campello, \{V{\'i}ctor M.\} and Karim Lekadir",

year = "2025",

doi = "10.1109/TMI.2025.3590694",

language = "英语",

volume = "44",

pages = "5290--5305",

journal = "IEEE Transactions on Medical Imaging",

issn = "0278-0062",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "12",

}

Bai, J, Zhu, J, Chen, Z, Yang, Z, Lu, Y, Li, L, Li, Q, Wang, W, Zhang, H, Wang, K, Gan, J, Zhao, J, Lu, H, Li, S, Huang, J, Chen, X, Zhang, X, Xu, X, Li, L, Tian, Y, Campello, VM & Lekadir, K 2025, 'A Benchmark Framework for the Right Atrium Cavity Segmentation From LGE-MRIs', IEEE Transactions on Medical Imaging, vol. 44, no. 12, pp. 5290-5305. https://doi.org/10.1109/TMI.2025.3590694

TY - JOUR

T1 - A Benchmark Framework for the Right Atrium Cavity Segmentation From LGE-MRIs

AU - Bai, Jieyun

AU - Zhu, Jinwen

AU - Chen, Zhiting

AU - Yang, Ziduo

AU - Lu, Yaosheng

AU - Li, Lei

AU - Li, Qince

AU - Wang, Wei

AU - Zhang, Henggui

AU - Wang, Kuanquan

AU - Gan, Jie

AU - Zhao, Jichao

AU - Lu, Hua

AU - Li, Suining

AU - Huang, Jiawen

AU - Chen, Xiaoming

AU - Zhang, Xiaoshen

AU - Xu, Xiaowei

AU - Li, Lulu

AU - Tian, Yanfeng

AU - Campello, Víctor M.

AU - Lekadir, Karim

PY - 2025

Y1 - 2025

N2 - The right atrium (RA) is critical for cardiac hemodynamics but is often overlooked in clinical diagnostics. This study presents a benchmark framework for RA cavity segmentation from late gadolinium-enhanced magnetic resonance imaging (LGE-MRIs), leveraging a two-stage strategy and a novel 3D deep learning network, RASnet. The architecture addresses challenges in class imbalance and anatomical variability by incorporating multi-path input, multi-scale feature fusion modules, Vision Transformers, context interaction mechanisms, and deep supervision. Evaluated on datasets comprising 354 LGE-MRIs, RASnet achieves SOTA performance with a Dice score of 92.19% on a primary dataset and demonstrates robust generalizability on an independent dataset. The proposed framework establishes a benchmark for RA cavity segmentation, enabling accurate and efficient analysis for cardiac imaging applications.

AB - The right atrium (RA) is critical for cardiac hemodynamics but is often overlooked in clinical diagnostics. This study presents a benchmark framework for RA cavity segmentation from late gadolinium-enhanced magnetic resonance imaging (LGE-MRIs), leveraging a two-stage strategy and a novel 3D deep learning network, RASnet. The architecture addresses challenges in class imbalance and anatomical variability by incorporating multi-path input, multi-scale feature fusion modules, Vision Transformers, context interaction mechanisms, and deep supervision. Evaluated on datasets comprising 354 LGE-MRIs, RASnet achieves SOTA performance with a Dice score of 92.19% on a primary dataset and demonstrates robust generalizability on an independent dataset. The proposed framework establishes a benchmark for RA cavity segmentation, enabling accurate and efficient analysis for cardiac imaging applications.

KW - Right atrium

KW - cardiac segmentation

KW - challenge

KW - late gadolinium-enhanced magnetic resonance imaging

KW - segment anything model

UR - https://www.scopus.com/pages/publications/105011734682

U2 - 10.1109/TMI.2025.3590694

DO - 10.1109/TMI.2025.3590694

M3 - 文章

AN - SCOPUS:105011734682

SN - 0278-0062

VL - 44

SP - 5290

EP - 5305

JO - IEEE Transactions on Medical Imaging

JF - IEEE Transactions on Medical Imaging

IS - 12

ER -

A Benchmark Framework for the Right Atrium Cavity Segmentation From LGE-MRIs

Abstract

Keywords

Access to Document

Other files and links

Fingerprint

Cite this