2D Mammograms with DeepSight Cancer Annotations from OMAMA-DB

165,368 2D mammogram images from 157,810 patients, annotated for cancer detection by the Oregon-Massachusetts Mammography Database (OMAMA-DB). The dataset includes metadata files with lesion coordinates, confidence scores, and cancer classifications, with 7,351 cancer cases and 4,371 images having a validated DeepSight score. It was last updated on July 15, 2026.

Use Cases

• Training object detection models for lesion localization based on the provided bounding box coordinates.
• Developing image classification models for cancer diagnosis based on the provided 'Label' field.
• Analyzing the impact of image windowing settings ('WindowCenter', 'WindowWidth') on model performance.
• Validating AI model predictions against the provided 'Score' confidence values.

Strengths

• Large scale with 165,368 images from 157,810 patients.
• Includes cancer annotations for 7,351 cases and validated scores for 4,371 images.
• Metadata provides detailed technical parameters like pixel spacing and windowing settings.
• Balanced laterality with 81,523 left-side and 82,045 right-side images.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• Data may reflect geographic or institutional bias inherent to the OMAMA-DB source.

Provenance

Source

The Oregon-Massachusetts Mammography Database (OMAMA-DB), authored by Daniel Haehn.

Collection Method

Likely collected from clinical mammography screenings.

Time Range

null

Freshness

Last updated 2026-07-15 15:16:09; freshness should be verified.

Geography

null