OXVASC T2*-GRE Microbleed Set
Oxford Vascular Study (OXVASC) T2*-weighted GRE cerebral microbleed segmentation set
Brain MRI of patients after minor stroke or TIA with manual microbleed segmentations, used to train and test a deep learning microbleed detector. 74 patients from a population-based Oxford study, 366 microbleeds in 36 of them. Shared on request to the principal investigator.
Overview
This entry covers a labelled subset of the Oxford Vascular Study (OXVASC), not the whole OXVASC imaging cohort. OXVASC is a population-based study at the Wolfson Centre for the Prevention of Stroke and Dementia in Oxford that has followed patients with vascular events from eight Oxfordshire general practices since 2002. Sundaresan et al. (2023) used T2*-weighted gradient echo scans of 74 OXVASC patients, with manual cerebral microbleed segmentations, to train and test a deep learning microbleed detector. The rest of the OXVASC imaging data is not part of this entry.
Composition
74 patients who had recently had a minor non-disabling stroke or a transient ischaemic attack; the paper does not split them by diagnosis. 36 are women and 38 are men, aged 39.6 to 91.2 years (mean 69.8, median 67.3). 36 patients have at least one microbleed, 366 microbleeds in total (mean 10.2, median 3 per patient). The other 38 have none and serve as negative cases. The paper describes T2*-weighted GRE images for every patient.
Acquisition
Scans were acquired on a 3 T Siemens Verio: 2D single-echo T2*-weighted GRE with GRAPPA 2, TR 504 ms, TE 15 ms, flip angle 20 degrees, 0.9 x 0.8 mm in-plane resolution, 5 mm slices and a 640 x 640 x 25 matrix. The study was approved by the South Central Oxford A Research Ethics Committee (05/Q1604/70).
Annotations
Microbleeds were segmented manually on the T2*-GRE images for all 36 positive patients, giving voxel masks. The paper does not say who drew them or which rating rules were used. The authors' detection code is public at github.com/v-sundaresan/microbleed-detection, without a license file.
Known limitations
- Not public. Requests go to the OXVASC principal investigator and are judged case by case; no data use terms are published.
- One scanner model and one protocol, with thick 5 mm slices.
- Rater, protocol and inter-rater agreement for the masks are not reported.
- The paper does not say whether other sequences or clinical data come with the T2*-GRE scans.
Cohort
Aggregate numbers from the sources below. Bars are relative to the 74 subjects.
Sex
- Female 36 49%
- Male 38 51%
Section 3.2 in Sundaresan et al. 2023, Frontiers in Neuroinformatics (PMC10363739)
Condition
Groups can overlap
- Cerebral microbleeds 36 49%
Section 3.2 in Sundaresan et al. 2023, Frontiers in Neuroinformatics (PMC10363739)
Field strength
- 3 T 74 100%
Section 3.2 in Sundaresan et al. 2023, Frontiers in Neuroinformatics (PMC10363739)
Age
mean 69.8 ± 14.6, median 67.3, range 39.6 to 91.2
No age bins reported.
Section 3.2 in Sundaresan et al. 2023, Frontiers in Neuroinformatics (PMC10363739)
Contrast / sequence
Groups can overlap
- T2*-weighted 74 100%
Section 3.2 in Sundaresan et al. 2023, Frontiers in Neuroinformatics (PMC10363739)
Scanner vendor
- Siemens Healthineers 74 100%
Section 3.2 in Sundaresan et al. 2023, Frontiers in Neuroinformatics (PMC10363739)
Country
- United Kingdom 74 100%
From Oxford Vascular Study page, Nuffield Department of Clinical Neurosciences
Contrast combinations
How many subjects have exactly each set of contrasts.
| T2starw | Subjects with exactly this set |
|---|---|
74 |
Section 3.2 in Sundaresan et al. 2023, Frontiers in Neuroinformatics (PMC10363739)
License and access
Our reading of the license, not legal advice. Before you use the data, read the original license and confirm that your use is allowed. We take no responsibility for how you use a dataset. Full disclaimer
Ask the authors or the data holder. No published process or terms, and access is at their discretion
No published license or data use terms
The data holder has published no license and no data use terms. Any use, sharing or commercial right has to be agreed with the data holder, and nothing can be assumed to be allowed. Default copyright and data protection law still apply.
The paper's data availability statement says requests for OXVASC data are considered by the PI, Peter Rothwell, in line with data protection laws. The OXVASC web page adds that fully anonymised data "may be shared with other organisations outside the University of Oxford for research to improve healthcare delivery (not NHS data) and subject to appropriate agreements being in place and approval by the Principal Investigator". No agreement text is published.
What you can do
- Commercial use Not stated
- Not stated
- Train ML models Not stated
- Create derived data Not stated
- Publish results Not stated
What you can share
- Share the data Not stated
- Share derived data Not stated
- Share trained models Not stated
What you must do
- Cite or credit Not stated
- Share alike Not stated
- Sign an agreement Not stated
- Ethics approval Not stated
- Manuscript review Not stated
- Release code Not stated
- Return results Not stated
- Delete after use Not stated
Limits
- No re-identification Not stated
- Location limits Not stated
Citation
Sundaresan V, Arthofer C, Zamboni G, Murchison AG, Dineen RA, Rothwell PM, Auer DP, Wang C, Miller KL, Tendler BC, Alfaro-Almagro F, Sotiropoulos SN, Sprigg N, Griffanti L, Jenkinson M. Automated detection of cerebral microbleeds on MR images using knowledge distillation framework. Frontiers in Neuroinformatics 17, 1204186 (2023). doi:10.3389/fninf.2023.1204186
Sources
Every number on this page comes from one of these documents. Each chart names the table or page it is taken from. The raw numbers are in stats.csv.