# Type B aortic dissection CTA (Graz): 40 volumes with expert true-lumen and false-lumen annotations

## Facts

- Provider: Christian Mayer, Antonio Pepe, Jan Egger, Heinrich Mächler et al. (Medical University of Graz); mirrored by MedOtter
- License: CC-BY-4.0 (https://api.figshare.com/v2/articles/22269091)
- Size: 284 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/MedOtter/Aortic-Dissection-Graz
- Access: public, readable without an account
- DOI: https://doi.org/10.6084/m9.figshare.22269091
- Category: Health & Biomedical
- Source: https://doi.org/10.1038/s41597-024-03284-2
- Page: https://scigantic.com/archive/bb06af73-ea6e-52ee-9fa2-227911784ec5
- Keywords: Health & Biomedical, CT, ECG, image segmentation, Parquet, Hugging Face

## About

40 computed tomography angiography (CTA) volumes of type B aortic dissection collected in clinical routine between 2005 and 2021 across the Medical University of Graz regional hospital network, each with expert segmentations of the true lumen and the false lumen. Scans are arterial phase and not ECG-gated, mostly 512x512xZ with in-plane spacing 0.525 to 0.965 mm and slice spacing 1.25 to 3.0 mm; head and face were cropped for anonymization. Thrombosed sections, calcifications and atherosclerotic changes were not segmented. The repository holds the image volumes and masks as NRRD with surface meshes (STL) under dataset/, the untouched source masks under original_masks/, and train.jsonl with per-case shape, spacing and true and false lumen volumes in mL; about 3.8 GB (measured 2026-11-17). Use for lumen segmentation and false-lumen volumetry. Read NRRD with SimpleITK or pynrrd and meshes with trimesh. Mirrored on Hugging Face by MedOtter from the authors' release, with a card documenting provenance and any normalization. The original figshare record is CC BY 4.0; cite Mayer et al., Type B Aortic Dissection CTA Collection with True and False Lumen Expert Annotations, Scientific Data 11, 596 (2024).

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("MedOtter/Aortic-Dissection-Graz")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `preview/train-00000-of-00001.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/bb06af73-ea6e-52ee-9fa2-227911784ec5
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/bb06af73-ea6e-52ee-9fa2-227911784ec5
- Over MCP: `get_schema_card` with id `bb06af73-ea6e-52ee-9fa2-227911784ec5` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Image segmentation datasets](https://scigantic.com/topics/image-segmentation)
- [Health & Biomedical datasets](https://scigantic.com/topics/health-and-biomedical)
- [CT datasets](https://scigantic.com/topics/ct)
- [ECG datasets](https://scigantic.com/topics/ecg)
