MR-Eye atlas: sex-specific and combined T1-weighted templates in common MRI spaces
Poster No:
1633
Submission Type:
Abstract Submission
Authors:
Jaime Barranco1, Adrian Luyken2, Philipp Stachs3, Oscar Esteban4, Yasser Alemán-Gómez5, Sönke Langner2, Oliver Stachs2, Benedetta Franceschiello6, Meritxell Bach-Cuadra7
Institutions:
1CIBM, CHUV, UNIL, HES-SO, Lausanne, Vaud, 2Department of Ophthalmology, Rostock University Medical Center, Rostock, Germany, 3Karlsruhe Institute of Technology (KIT), Karlsruhe, Germany, 4Department of Radiology, Lausanne University Hospital and University of Lausanne, Lausanne, Vaud, 5Centre hospitalier universitaire vaudois (CHUV), Lausanne, VT, 6Institute of Systems Engineering, School of Engineering, HES-SO Valais-Wallis, Sion, Valais, 7CIBM Center for Biomedical Imaging, Lausanne, Switzerland
First Author:
Co-Author(s):
Oscar Esteban
Department of Radiology, Lausanne University Hospital and University of Lausanne
Lausanne, Vaud
Department of Radiology, Lausanne University Hospital and University of Lausanne
Lausanne, Vaud
Benedetta Franceschiello
Institute of Systems Engineering, School of Engineering, HES-SO Valais-Wallis
Sion, Valais
Institute of Systems Engineering, School of Engineering, HES-SO Valais-Wallis
Sion, Valais
Introduction:
Brain atlases have long supported neuroimaging by enabling spatial normalization and analysis across populations (Dickie, 2017) (Fonov, 2009) (Holmes, 1998). These templates aid in understanding neurological disorders and brain function. However, standardized tools are lacking for dedicated reference imaging of the eye and orbit structures. A recent study (Lee, 2024) introduced eye MRI (MR-Eye) atlases (N=100), but larger-scale atlases capturing anatomical variability and sex-specific differences are needed, e.g. in endocrine orbitopathy (Hierl, 2022). In this context, we present the first large-scale male (N=594), female (N=616), and combined (N=1210) T1w MR-Eye atlases, with detailed healthy eye and orbit structure labels mapped onto two volumetric coordinate systems (VCS) (Fonov, 2009) (Holmes, 1998). These atlases will support the standardization of spatial references for MR-Eye.
Methods:
Data. We used the SHIP dataset (Schmidt, 2019) (Volzke, 2011): 1245 T1w scans (56±13 y.o., 620M, 625F) from a 1.5T Magnetom Avanto. Experts annotated 74 subjects' right eyes, including lens, globe, optic nerve, intraconal/extraconal fat, and rectus muscles (RM). MNI152 (Fonov, 2009) (152 subjects) and Colin27 (Holmes, 1998) (1 male) VCS were used for alignment.
MR-Eye segmentation method. Labels for unannotated SHIP data were generated using nnU-Net (Isensee, 2021), trained on 35 cases (4 validation) and tested on 39. Non-training data (N=1210) contributed to atlas creation.
Template construction. We employed ANTs toolkit (Avants, 2009) for iterative image registration to build average maps. Eye-cropped images, guided by nnU-Net segmentations, served as inputs.
Labels generation. Templates were aligned to subjects, and nnU-Net-derived labels were projected onto template space. Majority voting yielded maximum probability maps, while voxel-wise label probabilities reflected uncertainty (gray for low confidence).
Registration to common VCS. Eye regions in templates (Fonov, 2009) (Holmes, 1998) were cropped using modified antsBrainExtraction masks and registered to the combined and male eye atlas, respectively. Labels were projected onto cropped VCS spaces and transposed back.
MR-Eye segmentation method. Labels for unannotated SHIP data were generated using nnU-Net (Isensee, 2021), trained on 35 cases (4 validation) and tested on 39. Non-training data (N=1210) contributed to atlas creation.
Template construction. We employed ANTs toolkit (Avants, 2009) for iterative image registration to build average maps. Eye-cropped images, guided by nnU-Net segmentations, served as inputs.
Labels generation. Templates were aligned to subjects, and nnU-Net-derived labels were projected onto template space. Majority voting yielded maximum probability maps, while voxel-wise label probabilities reflected uncertainty (gray for low confidence).
Registration to common VCS. Eye regions in templates (Fonov, 2009) (Holmes, 1998) were cropped using modified antsBrainExtraction masks and registered to the combined and male eye atlas, respectively. Labels were projected onto cropped VCS spaces and transposed back.
Results:
We introduce large-scale unbiased male, female, and combined MR-Eye atlases with probability maps (Barranco, 2024). Figure 1 highlights male and female differences, with fat volume larger in males.
Accurate labels were mapped to MNI152 and Colin27 VCS. Figure 2 shows projections, with Colin27 volumes closer to reference and MNI152 generally larger. Lens and fat volumes showed notable differences.
Accurate labels were mapped to MNI152 and Colin27 VCS. Figure 2 shows projections, with Colin27 volumes closer to reference and MNI152 generally larger. Lens and fat volumes showed notable differences.
Conclusions:
Sex-based differences, especially larger fat volume in males, highlight the need for separate male and female atlases. Consistent volumes for Colin27 and MNI152 demonstrate the atlases' utility. MR-Eye atlases are a step forward in ophthalmic imaging standardization, supporting sex-specific disease research and improved diagnostics.
Modeling and Analysis Methods:
Image Registration and Computational Anatomy 2
Segmentation and Parcellation 1
Neuroinformatics and Data Sharing:
Brain Atlases
Databasing and Data Sharing
Keywords:
Segmentation
STRUCTURAL MRI
Vision
Other - eye; atlas
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Provide references using APA citation style.
2. Barranco, J., (2024). Eye-Opening Advances: Automated 3D Segmentation, Key Biomarkers Extraction, and the First Large-Scale MRI Eye Atlas. bioRxiv.
3. Dickie, D. A., (2017). Whole Brain Magnetic Resonance Image Atlases: A Systematic Review of Existing Atlases and Caveats for Use in Population Imaging. Frontiers in Neuroinformatics, 11, 1.
4. Fonov, V., (2009). Unbiased nonlinear average age-appropriate brain templates from birth to adulthood. NeuroImage, 47, S102.
5. Hierl, K. V., (2022). 3-D cephalometry of the the orbit regarding endocrine orbitopathy, exophthalmos, and sex. PLOS ONE, 17(3), e0265324.
6. Holmes, C. J., (1998). Enhancement of MR Images Using Registration for Signal Averaging. Journal of Computer Assisted Tomography, 22(2), 324.
7. Isensee, F., (2021). nnU-Net: a self-configuring method for deep learning-based biomedical image segmentation. Nature Methods, 18(2), 203–211.
8. Lee, H. H., (2024). Super-resolution multi-contrast unbiased eye atlases with deep probabilistic refinement (No. arXiv:2401.03060). arXiv.
9. Schmidt, P., (2019). Association of anthropometric markers with globe position: A population-based MRI study. PLoS ONE, 14(2), e0211817.
10. Volzke, H., (2011). Cohort Profile: The Study of Health in Pomerania. International Journal of Epidemiology, 40(2), 294–307.