Dbcm: 2D Slice-Based Multisequence Breast MRI Dataset with Lesion Segmentation

Derived from the Duke Breast Cancer MRI collection on The Cancer Imaging Archive, this dataset provides 2D slices from multi-sequence 3D breast MRI volumes. The original source includes pathologically confirmed lesions with bounding-box annotations and associated clinical, pathological, and molecular metadata. The dataset was last updated on 2026-06-18.

Use Cases

• Training segmentation models based on 2D MRI slices with lesion annotations.
• Developing multi-sequence MRI analysis algorithms based on the described T1, T2, and DCE sequences.
• Conducting radiomics research based on pathologically confirmed lesion data.
• Benchmarking medical image classification models on a dataset with comprehensive clinical metadata.

Strengths

• Derived from a large-scale, pathologically confirmed breast MRI collection.
• Includes comprehensive clinical, pathological, and molecular metadata from the original source.
• Provides lesion segmentation annotations for 2D MRI slices.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

The Cancer Imaging Archive (TCIA), Duke Breast Cancer MRI collection.

Collection Method

Derived and processed from original 3D MRI volumes.

Freshness

Last updated 2026-06-18 15:16:28; freshness should be verified.