pwoznicki/RadiomicsHub

Radiomics Features from Public Medical Imaging Datasets

This repo gathers together the available open-source datasets suitable for radiomics research.

More information about each dataset and the extracted radiomics features as well as the labels can be accessed at https://radiomics.uk.

Datasets

Dataset Name	Website	Task	Status
LIDC-IDRI	TCIA	binary classification	✔️
LNDb	Zenodo	multiclass classification	✔️
NSCLC-Radiogenomics	TCIA	survival analysis	✔️
NSCLC-Radiomics	TCIA	survival analysis	✔️
LUAD-CT-Survival	TCIA	binary classification	✔️
RIDER-Lung-CT	TCIA	repeatability	✔️
BraTS-2021	Kaggle	binary classification	✔️
UCSF-PDGM	TCIA	binary classification, survival analysis	✔️
UPENN-GBM	TCIA	survival analysis	✔️
Meningioma-SEG-CLASS	TCIA	binary classification	✔️
LGG-1p19qDeletion	TCIA	binary classification	✔️
PI-CAI	Grand Challenge	multiclass classification	✔️
Prostate-MRI-US-Biopsy	TCIA	multiclass classification	✔️
QIN-PROSTATE	TCIA	repeatability	✔️
Head-Neck-Radiomics-HN1	TCIA	survival analysis	✔️
HNSCC	TCIA	survival analysis	✔️
Head-Neck-PET-CT	TCIA	survival analysis	✔️
OPC-Radiomics	TCIA	survival analysis	✔️
QIN-HEADNECK	TCIA	repeatability	✔️
Colorectal-Liver-Metastases	TCIA	survival analysis	✔️
HCC-TACE-Seg	TCIA	survival analysis	✔️
C4KC-KiTS	TCIA	survival analysis	✔️
Soft-Tissue-Sarcoma	TCIA	binary classification	✔️
WORC	GitHub	binary classification	✔️

Folder structure

Each dataset adheres to the following structure, with minor variations:

<dataset_name>
├── raw
│   ├── dicom       # depending on the format of the original dataset
│   └── tables
└── derived
    ├── nifti       # converted to NIfTI format
    └── tables