Clear Cell Renal Cell Carcinoma Nuclei Grading Dataset
Select Region of Interests from WSI
Experienced pathologists are invited to select regions of interest (tumor regions) from ccRCC WSIs.
Mark Every Nucleus by Different Colored Dots
Every nucleus is annotated by three well-trained annotators by the Nuclick pre-trained model. Majority voting is adapted to assign the type of each nucleus, and then these results are reviewed by pathologists.
This dataset is proposed for nuclei grading of ccRCC.
It consists of 1000 H&E stained image patches with a resolution of 512x512.
In total the dataset contains 70945 labeled nuclei, each with an instance segmentation mask and a classification mask.
The original WSIs are derived from the TCGA database.
The dataset provided here is for research purposes only. Commercial uses are not allowed.
If you intend to publish a research work that uses any of these datasets, you must cite our publication.
ISUP/WHO guidelines for grading of RCC
Papers
Nuclei Grading of Clear Cell Renal Cell Carcinoma in Histopathological Image by Composite High-Resolution Network
Zeyu Gao, Jiangbo Shi, Xianli Zhang, Yang Li, Haichuan Zhang, Jialun Wu, Chunbao Wang, Deyu Meng, Chen Li
Ground Truth Demo
Statistics
This dataset contains 70945 annotated nuclei that consist of 16652 endothelial nuclei and 54293 tumor nuclei (45108, 6406, 2779 for grades 1 to 3, respectively), totally in four classes.
Blue: Endothelial nuclei
Green: Grade 1 tumor nuclei
Yellow: Grade 2 tumor nuclei
Red: Grade 3 tumor nuclei
Data Format
This dataset (1000 H&E WSIs) is divided into three subsets, 700 for training, 200 for testing, and 100 for validation. Each sample of this dataset is composed of two parts:
The original ROIs (image patches) selected from WSIs.
Save as png files under the corresponding folder.
The corresponding annotation of each ROI.
Save as mat files under the corresponding folder.
There are three keys in each mat file: 'instance_map' , 'class_map' and 'TCGA_file_name'.
The key 'instance_map' saves a pixel matrix with 0 to N. 0 represents the background and 1-N represent the instance number of each nucleus.
The key 'class_map' saves a pixel matrix with 0 to 4. 0 represents the background, 1-3 represent the grade of each tumor nucleus and 4 represents endothelial nuclei.
The key ‘TCGA_file_name’ records where the image patches cropped from.
Applications
A WSI of ccRCC marked as Grade 1
A WSI of ccRCC marked as Grade 2
A WSI of ccRCC marked as Grade 3
Visualisation of applying nuclei segmentation and classification network trained on ccRCC grading dataset to unseen whole-slide images.