The Kidney and Kidney Tumor Segmentation Challenge （KiTS21）

License

CC BY-NC-SA 4.0
Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License

Citation

Paper BibTeX:

@article{heller2021state,
  title={The state of the art in kidney and kidney tumor segmentation in contrast-enhanced CT imaging: Results of the KiTS19 challenge},
  author={Heller, Nicholas and Isensee, Fabian and Maier-Hein, Klaus H and Hou, Xiaoshuai and Xie, Chunmei and Li, Fengyi and Nan, Yang and Mu, Guangrui and Lin, Zhiyong and Han, Miofei and others},
  journal={Medical image analysis},
  volume={67},
  pages={101821},
  year={2021},
  publisher={Elsevier}
}

Dataset description

KiTS21 builds on the KiTS19 challenge, which aimed to advance automatic 3D kidney and kidney tumor segmentation in contrast-enhanced CT scans. It provides a curated set of manually annotated volumes for benchmarking deep learning methods and supports an open leaderboard for ongoing evaluation.

KiTS21 challenge homepage: https://kits-challenge.org/kits23/

KiTS21 challenge design: https://zenodo.org/records/4674397

Number of CT volumes: 300

Contrast: Contrast-enhanced

CT body coverage: Abdomen (occasional chest/pelvis coverage)

Does the dataset include any ground truth annotations? Yes

Original GT annotation targets: Kidney, kidney tumor, kidney cyst

Number of annotated CT volumes: 300

Annotator: Human

Acquisition centers: Multiple, with varied scanner brands; predominantly from Minnesota, North Dakota, and western Wisconsin

Pathology/Disease: Kidney tumors

Original dataset download link: https://github.com/neheller/kits21/blob/master/README.md

Original dataset format: nifti

Note

These 300 volumes correspond to the KiTS21 training split, which includes all cases from the train and test splits of KiTS19.