RSNA-CSF: Cervical Spine Fracture CT Scans with Expert Annotations

Approximately 2,000 CT studies from twelve sites on six continents, collected by the Radiological Society of North America in partnership with the ASNR and ASSR. The dataset contains expert image-level annotations indicating the presence, vertebral level, and location of cervical spine fractures. Over 1.5 million spine fractures occur annually in the United States, with cervical spine fractures being the most common site.

Use Cases

• Train fracture detection models based on expert image-level annotations.
• Develop models for localizing fractures to specific vertebral levels based on the provided location labels.
• Study imaging patterns in elderly populations where fractures can be difficult to detect due to degenerative disease and osteoporosis.
• Benchmark diagnostic AI performance against specialist radiologist annotations.

Strengths

• Approximately 2,000 CT studies provide a substantial imaging base.
• Annotations were provided by spine radiology specialists from the ASNR and ASSR.
• Data was collected from twelve sites across six continents, suggesting geographic diversity.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and dataset size are unknown, which may limit suitability assessment.
• Last update date is unknown; freshness unverified.

Provenance

Source

Radiological Society of North America (RSNA), American Society of Neuroradiology (ASNR), American Society of Spine Radiology (ASSR)

Collection Method

CT studies collected from twelve sites, with expert annotations provided by specialist radiologists.

Time Range

null

Freshness

Last updated is unknown.

Geography

Twelve sites on six continents