๐ง BraTS-MEN 2023
๐ง BraTS-MEN 2023
Pre-operative meningioma segmentation ยท multiparametric MRI
๐ Publications and records
| Quick facts | |
|---|---|
| Cohort | 1,424 MRI examinations from 1,344 patients |
| Input | T1, T1ce, T2, FLAIR |
| Institutional origin | The dataset paper reports six contributing academic medical centers: Duke University, Thomas Jefferson University, Yale University, University of California San Francisco, University of Missouri, and University of Pennsylvania. Scientific Data article. |
| Official source | BraTS-MEN Synapse dataset ยท BraTS 2023 challenge page |
๐งพ Dataset overviewโ
The BraTS pre-operative meningioma dataset brings together multi-institutional, pre-operative MRI with expert-annotated tumor subregions and clinical metadata. It is intended for segmentation of meningioma components and supports evaluation across institutions.
๐๏ธ Official split and label availabilityโ
| Cohort | Cases | Label status |
|---|---|---|
| Training | 1,000 | Expert segmentations provided |
| Validation | 141 | Challenge leaderboard cohort; public data availability depends on the release/version |
| Test | 283 | Private holdout for unbiased evaluation; test labels are not public |
The Scientific Data article reports that training and validation data are publicly available, while the 283-case test cohort remains private. During the original challenge, validation labels were used for leaderboard evaluation. Because a post-challenge Synapse package may contain more than the original leaderboard bundle, verify the exact annotation files and terms for the version you download.
๐งฌ Modalities and annotationโ
The four MRI sequences are T1, post-contrast T1 (T1ce), T2, and FLAIR. The annotation protocol identifies enhancing tumor, non-enhancing tumor/core, and surrounding edema. Preserve the release-specific voxel-label mapping when deriving WT/TC/ET-style regions.
๐ Access and termsโ
- Synapse dataset record, DOI 10.7303/syn51514106
- Official challenge / access information
- Synapse registration and acceptance of the applicable post-challenge data terms may be required.
๐ Related challengeโ
ASNRโMICCAI BraTS 2023 Intracranial Meningioma Challenge evaluated multi-compartment tumor segmentation from pre-operative T1, T1ce, T2, and FLAIR MRI.
๐ Publications and citationโ
Dataset paper: LaBella et al. โA multi-institutional meningioma MRI dataset for automated multi-sequence image segmentation.โ Scientific Data, vol. 11, no. 1, article 496, 2024. DOI: 10.1038/s41597-024-03350-9.
Challenge analysis: LaBella et al. โAnalysis of the BraTS 2023 Intracranial Meningioma Segmentation Challenge.โ arXiv preprint, 2024. arXiv:2405.09787.
The dataset article and challenge analysis are separate references; the split notes above distinguish the original challenge cohorts from later public release contents.
๐ Sources last checked: 27 October 2026 ยท Back to the dataset catalog