QBIN is pleased to announce the result of the special COVID-19 competition. We are pleased to present three grants of $15,000 each to Carl Chartrand-Lefebvre, professor at the University of Montreal, Simon Duchesne, professor at Laval University and to Olivier Lesur, professor at the University of Sherbrooke, for the quality of their project in a competition which proved to be very high despite the short deadlines due to the circumstances.


Michaël Chassé, Joseph Paul Cohen, Carl Chartrand-Lefebvre, Louis-Antoine Mullie, David Buckeridge, Brent Richards, Han Ting Wang, Alexandre Cavayas, Lan Dao

Responding efficiently to a health crisis like the COVID-19 pandemic mandates the collection of thorough, high-quality structured data. Large-scale datasets of patient information can aid to develop classification and prediction models that can facilitate early diagnosis of COVID-19, characterize disease stages, identify patients at high risk of clinical deterioration or death, and understand which patients are most likely to benefit from specific supportive and/or disease-modifying treatment strategies.

The goal of this project is to build a rich database of clinical information from patients with a suspected or confirmed SARS-CoV-2 infection and subsequently use machine learning techniques to build advanced predictive models. The data will include clinical characteristics, laboratory results, ventilator settings, electrocardiogram, and de-identified imaging data. The aim is to build models for various purposes such as predicting COVID-19 status from chest x-rays and laboratory data, predicting key outcomes such as mortality and ICU admission, and identifying disease phenotypes and their treatments responses. Data will be collected from adult patients tested for COVID-19 at five hospital sites in Montreal, and analysis will be conducted in partnership with Mila, a world-leading artificial intelligence research institute.


Simon Duchesne, Patrick Archambault, Nathalie Duchesne, Louis Gagnon, Marie-Hélène Lévesque, Carl Chartrand-Lefebvre, Fabrizio Vecchio

One of the current difficulties experienced by medical staff around the world is a lack of life-saving resources and equipment shortages compared to needs and as a result, a lack of objective measures to allocate these resources based on medical evidence. The main objective of this project is to use artificial intelligence to combine chest radiographs, clinical, and laboratory data to improve the allocation of limited resources (for example ventilators) to individuals presenting with COVID pneumonia in intensive care units.

A decision-making algorithm will be trained using a wide variety of data, including bedside chest x-rays, demographics such as age and sex, mortality due comorbidities (e.g. diabetes, hypertension, history of respiratory problems), and other factors currently used to evaluate pneumonia severity using the SMART-COP score (eg systolic pressure, albumin, respiratory rate, tachycardia, confusion, oxygenation, and arterial pH). The algorithm will be trained using a bank of international data (including data from Italy and China) and then validated on data acquired in Quebec. Once completed, the algorithm and associated data will be prospectively evaluated in Quebec healthcare centers, before being made available to jurisdictions that necessitate this system to guide patient management and transfer decisions worldwide.

Read more about this project in the news (in French):
le Droit - Respirateur artificiel: un outil pour aider à prioriser les patients qui doivent être branchés


Olivier Lesur, Eric Marsault, Jean-Bernard Denault, Djemel Ziou:

Confocal microendoscopy is a technique by which histology-like images can be obtained from inside the human body in real-time. This is done by inserting a special probe that combines optical fibres with microscopy technology into the organ of interest. The technique can be used for the diagnosis of various diseases and abnormalities.

The goal of this project is to develop an optical probe for quantitative intravital, real-time monitoring of the presence of the SARS-CoV-2 virus in the lungs of patients suffering from COVID-19. Using an animal model of rats infected by SARS-CoV-2, the project aims to quantify a direct marker of viral activity (the viral main proteinase 3CLpro) by identifying a compound that becomes fluorescent after interaction with the virus. By validating a specific real-time imaging diagnostic of SARS-CoV-2 within the lungs and without any time-consuming sampling or processing, patients can be formally diagnosed as quickly as possible and then isolated, treated, and monitored. While the project will focus on the SARS-CoV-2 virus, the technology and methods developed have a wide range of potential future applications for the identification and treatment of other diseases.



We thank all the teams for submitting a project for this competition and stress that the work of researchers is essential to the fight against the coronavirus.

Congratulations to the winners and good luck!