SPIDER: lumbar spine MRI (T1, T2, T2 SPACE) of 218 patients with vertebra, disc and canal masks

pl.load("spider_lumbar_mri") loads this dataset. 447 sagittal lumbar MRI series (T1, T2, some T2 SPACE) of 218 patients with low back pain, with masks of vertebrae, intervertebral discs and spinal canal, as MetaImage. Two zips on Zenodo (images, masks) read series by series with HTTP Range. Label values: vertebrae numbered from L5 upward (1 = L5), 100 = spinal canal, discs numbered from L5/S1 upward (201 = L5/S1) - mapping TO CONFIRM against the dataset paper for the upper levels. Radiological gradings are in a CSV on the record.

import pardesline as pl

ds = pl.load("spider_lumbar_mri")
print(ds.describe())  # roles, frames, units, labels, licence, citation
s = ds[0]
obj = s.t1  # fetched now, cached under ~/.cache/pardesline/datasets/spider_lumbar_mri

Provenance and licence

Source	remote archive read by HTTP Range: https://zenodo.org/records/10159290/files/images.zip, https://zenodo.org/records/10159290/files/masks.zip
Scheme	archive
Size of the whole source	3.7 GB
Licence	CC-BY-4.0
Commercial use	allowed by the licence
Redistribution	allowed
Licence note	CC BY 4.0 on the Zenodo record (checked 2026-10-14).
Terms	https://zenodo.org/records/10159290
Homepage	https://spider.grand-challenge.org/
Cite	van der Graaf J.W., van Hooff M.L., Buckens C.F.M., et al. Lumbar spine segmentation in MR images: a dataset and a public benchmark. Sci Data 11:264, 2024.
Samples grouped by	id

Roles (what each sample contains)

Role	Type	Frame	Units	Labels / layout	Pattern or file
t1	volume (MR)	from the file header	from the file header		images/{id}_t1.mha Sagittal T1
t2	volume (MR)	from the file header	from the file header		images/{id}_t2.mha Sagittal T2
t2_space	volume (MR)	from the file header	from the file header		images/{id}_t2_SPACE.mha 3-D T2 SPACE when acquired
t1_mask	labels	from the file header	from the file header	1=vertebra_L5, 2=vertebra_L4, 3=vertebra_L3, 4=vertebra_L2, 5=vertebra_L1, 6=vertebra_T12, 7=vertebra_T11, 8=vertebra_T10 … (19 labels)	masks/{id}_t1.mha
t2_mask	labels	from the file header	from the file header	1=vertebra_L5, 2=vertebra_L4, 3=vertebra_L3, 4=vertebra_L2, 5=vertebra_L1, 6=vertebra_T12, 7=vertebra_T11, 8=vertebra_T10 … (19 labels)	masks/{id}_t2.mha
t2_space_mask	labels	from the file header	from the file header	1=vertebra_L5, 2=vertebra_L4, 3=vertebra_L3, 4=vertebra_L2, 5=vertebra_L1, 6=vertebra_T12, 7=vertebra_T11, 8=vertebra_T10 … (19 labels)	masks/{id}_t2_SPACE.mha

Standardized layout

Files are fetched on first access into the dataset cache and stay in their original format; ds.export(folder) writes every sample as standard files (PLY for points and meshes, MetaImage for volumes and label maps) with a .pardesline.json sidecar carrying frame, units, labels, licence and provenance, plus a manifest for the folder. Label maps become exact surfaces with to_mesh(label=k), DICOM series become one volume, raw arrays are decoded through the layout above.

ds = pl.load("spider_lumbar_mri").take(2)
ds.export("./standardized")        # PLY / MetaImage + sidecars + manifest.json

Algorithms and models that run on it

• labels roles: mask.to_mesh(label=k), mask.to_numpy(), metrics: dice, hd95 (evaluation-metrics-3d) ; catalog modules: pyvista_volume_processing, simpleitk_registration, simpleitk_segmentation
• volume roles: volume.align(moving), volume.segment(lower, upper), volume.to_mesh(isovalue=...), volume.to_sitk() ; catalog modules: pyvista_volume_processing, simpleitk_registration, simpleitk_segmentation
• Supervised segmentation: prepare_nnunet_dataset.py writes this dataset in the nnU-Net v2 raw layout; PointNet trains on the label surfaces.
• Self-supervised pretraining: 3D JEPA reads the volumes directly.

Related

• Datasets catalog · How to download a medical imaging dataset in Python · Add a dataset
• Tags: medical, mri, spine, lumbar, segmentation, metaimage

_{Generated from src/pardesline/data/manifests/spider_lumbar_mri.json on 2026-10-14.}