# ORION-CRC: paired 19-channel immunofluorescence and H&E crops of colorectal cancer with nucleus and whole-cell masks

## Facts

- Provider: Jia-Ren Lin, Yu-An Chen, Daniel Campton et al. (ORION); VitaminP processing by Yasin Shokrollahi; curated by SPORA
- License: CC-BY-4.0 (https://zenodo.org/records/21224234)
- Size: 18 files
- Format: CSV, PNG, JSON (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/csmp-hub/orion-crc
- Access: public, readable without an account
- DOI: https://doi.org/10.5281/zenodo.21224234
- Category: Health & Biomedical
- Source: https://doi.org/10.1038/s43018-023-00576-1
- Page: https://scigantic.com/archive/155c4fe6-c117-52ad-b27f-d8f1ad7cbab6
- Keywords: Health & Biomedical, histopathology, image segmentation, CSV, PNG, JSON, Hugging Face

## About

Paired multiplexed immunofluorescence and H&E image crops of human colorectal cancer from the ORION same-section imaging study of Lin, Chen, Campton et al. (Nature Cancer 2023), prepared for cross-modal whole-cell segmentation by the VitaminP project. There are 3,337 crops of 512x512 pixels at 0.325 micrometers per pixel from 41 tissue samples representing 40 cases. Each crop keeps all 19 original ORION fluorescence channels (nuclear, immune and epithelial markers, plus autofluorescence and blank channels) next to the H&E of the same tissue region, with instance masks for nuclei and whole cells. The VitaminP train/val/test assignment is kept with each case in one split: 2,946 / 272 / 119 crops, holding about 550k / 42k / 19k nuclei. Supporting files include the original ORION data-release README, the marker list, split and source audits, and license JSON. About 21 GB (measured 2026-11-17). Images and masks are stored as PNG or TIFF bytes inside Parquet; read them with polars or pandas plus Pillow, or with Hugging Face datasets. Use for virtual H&E to multiplex translation, cross-modal cell segmentation and spatial tumor microenvironment studies. Curated and packaged on Hugging Face by the SPORA project (csmp-hub), which keeps the original license and citation files. The VitaminP Zenodo record 21224234 is CC BY 4.0; cite Lin et al., Nature Cancer 2023 (doi 10.1038/s43018-023-00576-1) and the VitaminP release.

## Load the data

```python
import pandas as pd
df = pd.read_csv("hf://datasets/csmp-hub/orion-crc/crc-source-splits.csv")
df.head()
```

## Data structure

File types in sampled files: CSV, PNG, JSON

Example file: `crc-source-splits.csv`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/155c4fe6-c117-52ad-b27f-d8f1ad7cbab6
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/155c4fe6-c117-52ad-b27f-d8f1ad7cbab6
- Over MCP: `get_schema_card` with id `155c4fe6-c117-52ad-b27f-d8f1ad7cbab6` at https://api.scigantic.com/mcp

## Related topics

- [JSON datasets](https://scigantic.com/topics/json)
- [CSV datasets](https://scigantic.com/topics/csv)
- [Image segmentation datasets](https://scigantic.com/topics/image-segmentation)
- [Health & Biomedical datasets](https://scigantic.com/topics/health-and-biomedical)
- [PNG datasets](https://scigantic.com/topics/png)
- [Histopathology datasets](https://scigantic.com/topics/histopathology)
