STOIC (Study of Thoracic CT in COVID-19)

License

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

Citation

Paper BibTeX:

@article{revel2021study,
  title={Study of thoracic CT in COVID-19: the STOIC project},
  author={Revel, Marie-Pierre and Boussouar, Samia and de Margerie-Mellon, Constance and Saab, In{\`e}s and Lapotre, Thibaut and Mompoint, Dominique and Chassagnon, Guillaume and Milon, Audrey and Lederlin, Mathieu and Bennani, Souhail and others},
  journal={Radiology},
  volume={301},
  number={1},
  pages={E361--E370},
  year={2021},
  publisher={Radiological Society of North America}
}

Dataset description

Collected during March–April 2020 in France, this dataset contains chest CT scans from 10,735 individuals suspected of COVID-19 infection. For each patient in the training set, binary labels indicate COVID-19 presence (RT-PCR confirmed) and severity (intubation or death within one month). The public subset used here contains 2,000 scans as provided for the STOIC2021 challenge.

Challenge homepage: https://stoic2021.grand-challenge.org/

Number of CT volumes: 2000

Contrast: Mostly non-contrast; contrast-enhanced CT performed when pulmonary embolism was suspected

CT body coverage: Chest

Does the dataset include any ground truth annotations?: No (only labels for classification tasks)

Original GT annotation targets: -

Number of annotated CT volumes: -

Annotator: -

Acquisition centers: 20 university hospitals: 15 from Assistance Publique des Hôpitaux de Paris and five from other cities (Strasbourg, Lyon, Rennes, and Montpellier).

Pathology/Disease: COVID-19 pneumonia, suspected of being infected with SARS-COV-2 during the first wave of the pandemic

Original dataset download link: https://registry.opendata.aws/stoic2021-training/

Original dataset format: .mha

Note

This version uses the publicly available 2,000-scan subset released for the STOIC2021 challenge.