CT Lymph Nodes: 176 Chest and Abdomen CT Scans with Voxel-Wise Segmentations

176 chest and abdomen CT scans feature manually-traced voxel-wise mediastinal and abdominal lymph node segmentations. The collection underpins the Roth 2014 detection benchmark and Seff 2015 segmentation benchmark. It remains a widely-cited reference for thoracic-abdominal lymph node computer-aided detection work.

Use Cases

• Training 3D binary segmentation models based on manually-traced lymph node voxels.
• Benchmarking detection algorithms based on the Roth 2014 detection benchmark.
• Benchmarking segmentation algorithms based on the Seff 2015 segmentation benchmark.
• Developing CADe systems for thoracic-abdominal lymph nodes based on the described CT scans.

Strengths

• Contains 176 CT scans, providing a substantial volume of medical imaging data.
• Features manually-traced voxel-wise segmentations, which likely provide high-quality ground truth labels.
• Underpins two established benchmarks (Roth 2014, Seff 2015), indicating its foundational role in the field.

Limitations

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

Provenance

Source

Roth et al., NIH

Collection Method

Manually-traced voxel-wise segmentations on CT scans.

Freshness

Last updated 2026-05-21 12:02:14; freshness should be verified.