# LyNSeC: diffuse large B-cell lymphoma nucleus segmentation in H&E and IHC (CD3, Ki67, ERG), 161k nuclei

## Facts

- Provider: Hussein Naji, Reinhard Büttner et al.; mirrored by MedOtter
- License: CC-BY-4.0 (https://zenodo.org/records/8065174)
- Size: 5 files
- Format: Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/MedOtter/LyNSeC
- Access: public, readable without an account
- DOI: https://doi.org/10.5281/zenodo.8065174
- Category: Health & Biomedical
- Source: https://www.sciencedirect.com/science/article/pii/S0010482524000623
- Page: https://scigantic.com/archive/98a50862-fb2c-5911-ae23-00a2a28cdf9f
- Keywords: Health & Biomedical, histopathology, image segmentation, Parquet, Hugging Face

## About

LyNSeC (Lymphoma Nuclear Segmentation and Classification) is a histopathology dataset of diffuse large B-cell lymphoma (DLBCL) lymph node tissue with per-nucleus instance masks across H&E and immunohistochemistry. It has 699 RGB tiles of 512x512 pixels at 40x magnification with 161,247 annotated nuclei in three subsets: lynsec1_ihc (379 IHC tiles stained for CD3, Ki67 or ERG, 87,316 nuclei, marker-positive versus marker-negative class per nucleus), lynsec2_he (280 H&E tiles, 65,479 nuclei, instances only) and lynsec3_he_expert (40 expert-refined H&E tiles, 8,452 nuclei, tumor versus non-tumor class). Instance masks are 16-bit PNGs with one ID per nucleus; class masks are 8-bit. Parquet files under data/, about 360 MB (measured 2026-11-17); the source has no predefined train/test split. Use for nucleus instance segmentation, IHC positivity scoring and lymphoma cell classification. Mirrored on Hugging Face by MedOtter from the authors' release, with a card documenting provenance and any normalization. The authors' Zenodo record 8065174 is CC BY 4.0; cite Naji et al., HoLy-Net, Computers in Biology and Medicine 2024.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("MedOtter/LyNSeC")
print(ds)
```

## Data structure

File types in sampled files: Parquet

Example file: `data/lynsec1_ihc-00000-of-00001.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/98a50862-fb2c-5911-ae23-00a2a28cdf9f
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/98a50862-fb2c-5911-ae23-00a2a28cdf9f
- Over MCP: `get_schema_card` with id `98a50862-fb2c-5911-ae23-00a2a28cdf9f` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Image segmentation datasets](https://scigantic.com/topics/image-segmentation)
- [Health & Biomedical datasets](https://scigantic.com/topics/health-and-biomedical)
- [Histopathology datasets](https://scigantic.com/topics/histopathology)
