This data set is part of the public development data for the 2023 Automated Universal Classification Challenge (AUC23). The data set concerns COVID-19 RT-PCR outcome prediction and prediction of severe COVID-19, defined as death or intubation after one month, from computed tomography (CT). The data set was previously introduced and described by Revel, M. et al (2021). Data was restructured in compliance with the AUC23 challenge format. The STOIC project collected CT images of 10,735 individuals suspected of being infected with SARS-COV-2 during the first wave of the pandemic in France, from March to April 2020. For each patient in the training set, the dataset contains binary labels for COVID-19 presence based on RT-PCR test results, and COVID-19 severity, defined as intubation or death within one month from the acquisition of the CT scan. This data set contains the training sample of the STOIC dataset as used in the STOIC2021 challenge. Images are 3D tensors: 0: 3D CT scan Classification labels: COVID-19: 0: Negative RT-PCR 1: Positive RT-PCR Severe COVID-19: 0: Alive and no intubation after one month 1: Death or intubation after one month imagesTr (root folder with all patients and studies) ├── covid19severity_6_0000.mha (3D CT for study 6) ├── covid19severity_17_0000.mha (3D CT for study 17) ├── ... Please cite the following article if you are using the STOIC2021 training dataset: STOIC2021 Training was accessed on DATE from https://registry.opendata.aws/stoic2021-training. STOIC2021 Training was documented in Thoracic CT in COVID-19: The STOIC Project, Revel, Marie-Pierre, et al. Radiology, 2021, https://doi.org/10.1148/radiol.2021210384. Due to upload size limits, the data set was split into six batches. Batch 1: https://zenodo.org/record/7969800 Batch 2: https://zenodo.org/record/8042589 Batch 3: https://zenodo.org/record/8042817 Batch 4: https://zenodo.org/record/8043089 Batch 5: https://zenodo.org/record/8043216