Brain MRI Scans and Radiologist Reports for 2 Million Studies

Over 2 million medical studies pair brain MRI scans with radiologists' reports, including descriptions, conclusions, and recommendations. The imaging data features high-quality 1 mm slice thickness and an average of 30 slices per scan. This dataset, created by ud-medical and last updated in August 2025, is designed for detection, classification, and segmentation tasks across more than 50 pathologies.

Use Cases

• Train pathology detection models based on the paired MRI scans and diagnostic reports.
• Develop classification systems for brain tumors and lesions based on the described 50+ pathologies.
• Build segmentation models for brain structures using the high-resolution 1 mm slice thickness scans.
• Generate automated radiology report summaries based on the detailed conclusions and recommendations.

Strengths

• Contains over 2 million medical studies, providing a large-scale resource.
• Pairs high-quality MRI scans with 1 mm slice thickness to detailed textual radiologist reports.
• Covers more than 50 pathologies, including brain tumors and lesions.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Freshness should be verified as the last update was 2025-08-02.

Provenance

Source

ud-medical on Hugging Face

Freshness

Last updated 2025-08-02 00:23:08