# SPIDER: lumbar spine MRI segmentation of vertebrae, discs and spinal canal

## Facts

- Provider: Jasper W. van der Graaf et al., Radboud University Medical Center (mirrored by MedOtter)
- License: CC-BY-4.0 (https://zenodo.org/records/10159290)
- Size: 900 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/MedOtter/SPIDER
- Access: public, readable without an account
- DOI: https://doi.org/10.5281/zenodo.10159290
- Category: Health & Biomedical
- Source: https://doi.org/10.1038/s41597-024-03090-w
- Page: https://scigantic.com/archive/bbf348a1-679e-5d09-8cd6-9a693e06ec36
- Keywords: Health & Biomedical, MRI, image segmentation, benchmark, Parquet, Hugging Face

## About

SPIDER is a public lumbar spine MRI segmentation dataset and benchmark from Radboud University Medical Center, published in Scientific Data (van der Graaf et al. 2024). It contains 447 sagittal lumbar spine MRI series from 218 patients with a history of low back pain (196 T1, 210 T2 and 41 T2-SPACE), each with a reference segmentation of the vertebrae, the intervertebral discs and the spinal canal; vertebrae are numbered top-down, discs are labelled from 201 upward, and the canal is label 100. T2-SPACE masks are fully manual; T1 and T2 masks were refined semi-automatically under musculoskeletal radiologist supervision. overview.csv gives per-series metadata and the training (360) and validation (87) split, and radiological_gradings.csv gives per-disc gradings such as Modic changes, Pfirrmann grade and herniation. The hidden challenge test set is not included. This MedOtter repository reproduces the Zenodo v4 release; as measured 2026-11-14 it holds .mha volumes under images/images and masks/masks plus Parquet indexes, about 3.8 GB in all. Read .mha with SimpleITK or itk. Released under CC BY 4.0 on Zenodo; cite van der Graaf et al. 2024.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("MedOtter/SPIDER")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `data/training-00000-of-00001.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/bbf348a1-679e-5d09-8cd6-9a693e06ec36
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/bbf348a1-679e-5d09-8cd6-9a693e06ec36
- Over MCP: `get_schema_card` with id `bbf348a1-679e-5d09-8cd6-9a693e06ec36` at https://api.scigantic.com/mcp

## Related topics

- [MRI datasets](https://scigantic.com/topics/mri)
- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Benchmark datasets](https://scigantic.com/topics/benchmark)
- [Image segmentation datasets](https://scigantic.com/topics/image-segmentation)
- [Health & Biomedical datasets](https://scigantic.com/topics/health-and-biomedical)
