Overview

Schema Version

https://atlas.rsna.org/schemas/2025-11/dataset.json

Name

RSNA Lumbar Degenerative Imaging Spine Classification (LumbarDISC) Dataset

Link

https://mira.rsna.org/dataset/6

Indexing

Keywords: Degenerative Lumbar Spine Classification, Lumbar Spondylosis, MRI, Spinal Stenosis, Neural Foraminal Stenosis, Subarticular Stenosis, Image Annotation, Multi-institutional Dataset, Degenerative Spine Disease, Low Back Pain, Radiology

Content: MK, MR, NR

RadLex: RID28675, RID5085

Author(s)

Tyler J. Richards

Adam E. Flanders

Errol Colak

Luciano M. Prevedello

Robyn L. Ball

Felipe Kitamura

John Mongan

Maryam Vazirabad

Hui-Ming Lin

Anne Kendell

Thanat Kanthawang

Salita Angkurawaranon

Emre Altinmakas

Hakan Dogan

Paulo Eduardo de Aguiar Kuriki

Arjuna Somasundaram

Christopher Ruston

Deniz Bulja

Naida Spahović

Jennifer Sommer

Sirui Jiang

Eduardo Moreno Júdice de Mattos Farina

Eduardo Caminha Nunes

Michael Brassil

Megan McNamara

Johanna Ortiz

Jacob Peoples

Vinson L Uytana

Anthony Kam

Venkata N.S. Dola

Daniel Murphy

David Vu

Dataset Curator Group

Dataset Contributor Group

Dataset Annotator Group

Jason F. Talbott

Organization(s)

Radiological Society of North America

American Society of Neuroradiology (ASNR)

Chiang Mai University Faculty of Medicine

Diagnósticos da América S.A. (DASA)

Gold Coast University Hospital

Koc University School of Medicine

University of Sarajevo, Bosnia and Herzegovina

Thomas Jefferson University

Universidade Federal de São Paulo (Unifesp)

University of California - San Francisco

University of Utah School of Medicine

Queen's University at Kingston, Ontario, Canada

Tallaght University Hospital, Dublin, Ireland

University Hospitals Cleveland Medical Center

Version

1.0

License

Text: RSNA MIRA DATASET RESEARCH USE AGREEMENT

URL: https://docs.google.com/document/d/1r8_0yW-5XqxSqhFzFq2fV6L4NxIQ6drF0sBjXXJevXU/edit?usp=sharing

Ethical review

Local institutional review board approval obtained at each contributing site.

Comments

The RSNA Lumbar Degenerative Imaging Spine Classification (LumbarDISC) dataset is composed of MRI studies of the lumbar spine from 2,697 patients with a total of 8,593 image series from 8 institutions across 6 countries and 5 continents. It was created for the RSNA 2024 Lumbar Spine Degenerative Classification AI Challenge and future research.

Dataset

Motivation

To create the largest, most diverse, expertly annotated MRI lumbar spondylosis dataset for the RSNA 2024 Lumbar Spine Degenerative Classification AI Challenge and future research.

Sampling

Test set cases were selected using stratified random sampling, stratified on contributing site, age group, sex, and severity at each lumbar spine level. Contributing sites also supplemented the dataset with high-grade disease examples at under-represented levels.

Partitioning scheme

The dataset was divided into training, public test, and private test sets using stratified random sampling based on contributing site, age group, sex, and disease severity classes. High-grade stenosis cases at L1/L2, L2/L3, and L5/S1 were augmented in test sets.

Missing information

Low representation of high-grade disease at L1/L2 and L2/L3 levels, which may limit model performance at these naturally under-represented levels.

Noise

Studies with substantial motion degradation, severe scoliosis, or technical artifacts that limited diagnostic evaluation were excluded.

Re-identification

All identifying DICOM elements, including Patient ID, Study Instance UID, Series Instance UID, and SOP Instance UID, were de-identified using the pydicom Python library.

Sensitive data

Basic demographic information including patient age, sex, and ethnicity (if available) was collected. MRI images were de-identified. All identifying DICOM elements (Patient ID, Study Instance UID, Series Instance UID, SOP Instance UID) were removed.