Asan Medical Center SWI Microbleed Set (ARIA-H)
Asan Medical Center SWI microbleed dataset for ARIA-H severity assessment
In-house set of 565 SWI brain MRI scans from Asan Medical Center, Seoul (429 with microbleeds), labelled by one neuroradiologist to train microbleed detection for ARIA-H severity grading. Described in a 2025 AAIC abstract; not publicly released.
Overview
This is an in-house collection of susceptibility-weighted brain MRI from Asan Medical Center in Seoul, used by VUNO Inc. and Asan researchers to train an automatic cerebral microbleed detector. The goal is to count microbleeds for ARIA-H severity grading during anti-amyloid therapy, where fewer than 5, 5 to 9 and 10 or more microbleeds mean mild, moderate and severe. The only public description is a poster abstract from AAIC 2025 published in Alzheimer's & Dementia.
Composition
565 SWI scans: 429 with at least one microbleed and 136 without. The cohort has a mean age of 71.9 years (SD 10.2) and includes 363 women and 202 men. The scans were split 3:1:1 into training, validation and test sets; the test set has 114 scans (86 positive with 158 microbleeds, 28 negative). The abstract does not state whether the patients were on anti-amyloid therapy or what their diagnoses were.
Acquisition
All scans are SWI with 2 mm slices from a single centre. Scanner vendor, field strength and sequence parameters are not reported.
Annotations
One neuroradiologist with 14 years of experience labelled microbleeds, defined as hypointense lesions 2 to 10 mm across on SWI. The model was scored with Dice and lesion-level metrics using a 3 mm centre distance, which suggests lesion-level labels, but the label format (masks or points) is not described.
Known limitations
- Not publicly available, with no published access route or terms.
- Single centre, single reader, and only a short abstract describes the data.
- No scanner, field strength or clinical diagnosis information.
Cohort
Aggregate numbers from the sources below. Bars are relative to the largest value.
Age
mean 71.9± 10.2No age bins reported.
Contrast / sequence
scans, values can overlap
- Susceptibility-weighted 565 100%
Condition
scans, values can overlap
- Cerebral microbleeds 429 76%
Country
scans
- South Korea 565 100%
Split
scans
- Test 114 20%
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 abstract has no data availability statement and no data sharing route is published. The CC BY 4.0 license of the abstract covers the text only.
What you can do
- Commercial use 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
Kim S, Jung W, Lee SH, Suh CH. AI-based Microbleed Detection for ARIA Radiographic Severity Assessment. Alzheimer's & Dementia 21(Suppl 2), e100169 (2025). doi:10.1002/alz70856_100169
All numbers
Every number on this page, as stored in stats.csv, with its source.
| Measure | Breakdown | Value | Source |
|---|---|---|---|
| Scans | total SWI scans with 2 mm slices | 565 | kim2025 Method |
| Scans | contrast=swi | 565 | kim2025 Method |
| Scans | condition=cerebral_microbleeds positive cases; 136 scans are negative | 429 | kim2025 Method |
| Scans | country=KR Asan Medical Center | 565 | kim2025 Method |
| Scans | sex=female the abstract gives sex for the cohort; one scan per person is not stated | 363 | kim2025 Method |
| Scans | sex=male | 202 | kim2025 Method |
| Scans | split=test 86 positive (158 microbleeds) and 28 negative; train, validation and test were split 3:1:1 | 114 | kim2025 Result |
| Mean age | total | 71.9 | kim2025 Method |
| Age SD | total | 10.2 | kim2025 Method |
Sources
The keys used in the table above.