Published April 9, 2025 | Version v1
Dataset Open

HER2-IHC-40x: High-Resolution Histopathology Image Dataset for HER2 Scoring in Breast Cancer

Description

HER2-IHC-40x: High-Resolution Histopathology Datasets for HER2 IHC Scoring

Please cite the dataset from the journal paper Click :

Nabi, M. S., Fauzi, M. F. A., Rehman, Z. U., Karim, H. B. A., Cheah, P. L., Chiew, S. F., & Looi, L. M. (2025). HER2-IHC-40x: A High-Resolution Histopathology Dataset for HER2 IHC Scoring in Breast Cancer. Data in Brief, 111922.

Here is the experimental paper; you may read and cite it as well for your own experiment.  Click:

Nabi, M. S., Fauzi, M. F. A., Karim, H. B. A., Cheah, P. L., Fan, C. S., & Looi, L. M. (2025). Explainable deep learning models for HER2 IHC scoring in breast cancer diagnosis. Informatics in Medicine Unlocked, 101700.

 

Overview

This dataset contains high-resolution histopathological images of HER2-stained breast cancer tissue sections. Designed for deep learning-based HER2 scoring, the dataset includes two variants:
  • HER2-IHC-40x: Patches extracted after splitting WSIs.
  • HER2-IHC-40x-WSI: Patches extracted before splitting.
Each image patch is categorized into one of four HER2 classes (0, 1+, 2+, 3+), based on staining intensity.

 

Dataset Contents

Dataset Variants

1. HER2-IHC-40x
  • WSI-based 80-20 split before patch extraction.
  • 107 WSIs → 9940 patches (8093 train / 1847 test)
2. HER2-IHC-40x-WSI
  • Patch-based 80-20 split after patch extraction.
  • 107 WSIs → 10,997 patches (8897 train / 2200 test)

 

Folder Structure
HER2-IHC-40x/
├── WSI/              # Original Whole Slide Images (.svs)
├── ROI/              # Expert-annotated tumor regions (.png)
├── Patches/       # 1024x1024 image patches, labeled 0, 1+, 2+, 3+
├── Train/           # 80% training set
└── Test/            # 20% test set

 

HER2-IHC-40x-WSI/
├── Patches/
├── Train/
└── Test/
 

 

HER2 Class Definitions

 

| HER2 Score | Description                                                                 
|----------|-----------------------------------------------------------------------------
| 0           | No observable staining                                                                      
| 1+         | Weak/incomplete membrane staining in ≤10% tumor cells                       
| 2+         | Moderate circumferential staining in >10% tumor cells (Equivocal)           
| 3+         | Strong circumferential staining in >10% tumor cells (Positive)              

 

Preprocessing & Quality Control

  • ROI Selection: Manual annotation by expert pathologists using Cytomine.
  • Color Histogram Filtering: Removed non-tumor/low-quality patches using HSV filtering.
  • Normalization: Intensity normalization across all patches.
  • Patch Extraction: Adaptive 1024×1024 extraction using sliding window method.

 

Usage

This dataset is suitable for:
  • HER2 scoring automation using deep learning
  • Explainable AI (Grad-CAM, attention models)
  • Color normalization and domain adaptation
  • Model benchmarking and generalization research

 

Dataset Statistics

 
HER2-IHC-40x (WSI Split)
| HER2 Class | WSIs | ROIs | Patches |
|----------|------|----- -|---------|
| 0            | 23   | 429   | 3789    |
| 1+         | 26   | 131   | 2153    |
| 2+         | 27   | 483   | 634     |
| 3+         | 31   | 156   | 3364    |
| Total      | 107 | 1199 | 9940    |

 

HER2-IHC-40x (Patch Split)
| HER2 Class | WSIs | ROIs | Patches |
|----------|-----|--------|---------|
| 0           | 23   | 429     | 3789    |
| 1+         | 26   | 131    | 2689    |
| 2+         | 27   | 483    | 1131    |
| 3+         | 31   | 156    | 3388    |
| Total      | 107  | 1199 | 10,997  |
 



Citation

The dataset is already published as a journal article: If you use this dataset, please cite:
@article{nabi2025her2,
  title={HER2-IHC-40x: A High-Resolution Histopathology Dataset for HER2 IHC Scoring in Breast Cancer},
  author={Nabi, Md Serajun and Fauzi, Mohammad Faizal Ahmad and Rehman, Zaka Ur and Karim, Hezerul Bin Abdul and Cheah, Phaik Leng and Chiew, Seow Fan and Looi, Lai Meng},
  journal={Data in Brief},
  pages={111922},
  year={2025},
  publisher={Elsevier}
}
 
APA

Nabi, M. S., Fauzi, M. F. A., Rehman, Z. U., Karim, H. B. A., Cheah, P. L., Chiew, S. F., & Looi, L. M. (2025). HER2-IHC-40x: A High-Resolution Histopathology Dataset for HER2 IHC Scoring in Breast Cancer. Data in Brief, 111922.

 
 
This dataset is part of the research article:
**"Enhancing HER2 IHC Scoring Using HRNet and SwinT with Cross-Dataset Generalization"**  
Authors: Md Serajun Nabi, Mohammad Faizal Ahmad Fauzi, Hezerul Bin Abdul Karim, et al.  
(Preprint server or journal details to be confirmed.)
 
Search the paper for detail data description:
**HER2-IHC-40x: High-Resolution Histopathology Datasets for HER2 IHC Scoring**

The color histogram code source:
* https://github.com/seraju77/HER2-IHC-40x-data-preprocessing.git *

Files

her2-ihc-40x-patch.zip

Files (3.2 GB)

Name Size
md5:c05d1e73335912cbdd8e46e349d5ec72
1.7 GB Preview Download
md5:b34b004628283e41b6f2fc9df1f64f0c
1.5 GB Preview Download