On October 18th at 3pm CEST, the winners of STOIC2021 received their prizes and presented their solutions.

3.00 | Welcome and introduction to STOIC2021
3.05 |  The STOIC Project: clinical background
3:10 |  Award ceremony
3:15 |  Team #3 Solution: how did they do it?
3:20 |  Team #2 Solution: how did they do it?
3:25 |  Team #1 Solution: how did they do it?
3:30 |  Outline of the paper, next steps & closing
3:40 |  Q&A

STOIC2021 - COVID-19 AI Challenge

Can you predict from a CT scan who will develop severe COVID-19?

We are launching an artificial intelligence challenge aimed at predicting the severe outcome of COVID-19, based on the largest dataset of Computed Tomography (CT) images of COVID-19 suspects and patients collected to date. Participants will have access to data from the STOIC project, recently published in Radiology. 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. 

The focus of the challenge is the prediction of severe COVID-19, defined as intubation or death within one month from the acquisition of the CT scan (AUC, primary metric). COVID19 positivity will be assessed as a secondary metric in the leaderboard.

The STOIC 21 challenge will be organized in two phases, a qualification round and a final round.

In the qualification round, participating teams can use a publicly available sample of the STOIC dataset, consisting of 2000 scans and released under the CC BY-NC 4.0 license, for training. Participants must upload their solutions as a Docker container to https://grand-challenge.org/algorithms/.   

The test set consists of 1000 scans from the STOIC dataset. During the qualification round, a subset of this test set will be used to compute entries on the initial leaderboard. At the end of the qualification round, each team can submit one Algorithm to the final leaderboard where the full test set will be processed. 

The best-performing teams will be invited to participate in the final round. For this round, the finalists will upload a Docker container that can train an improved model, using the model from the qualification round as a starting point. The code for this container needs to be released under a permissive open-source license. This container will be trained on the grand-challenge.org platform, hosted on AWS, in a secure and protected environment with all scans from the entire STOIC dataset (minus the 1000 test scans, of course). The resulting solutions will be made available on https://grand-challenge.org/algorithms/ for research use.


Three top-performing solutions will be rewarded with AWS credits as follows.

🥇1st place:  AWS credits to the value of $10,000.

🥈2nd place:  AWS credits to the value of $6,000.

🥉3rd place:  AWS credits to the value of $4,000.

For detailed information see the Challenge Overview.

This challenge is sponsored by AWS and endorsed by MICCAI.