June 2, 2020


In response to the continuing COVID-19 epidemic in France, CNES and Toulouse University Hospital have teamed up to develop an artificial intelligence solution that aids interpretation of lung scans.

When a patient with COVID-19 is admitted to hospital, the standard examination for determining the seriousness of the infection and monitoring their condition is a thoracic scan. This complex radiation scan requires fragile and contagious patients to be transported in a medical vehicle. At a time when agility, time and the availability of healthcare personnel are at a premium, performing a thoracic scan becomes a delicate proposition.

Lung scan - A promising tool under study

Toulouse University Hospital has assembled a pool of 200 patients who are being subjected to a thoracic scan and a lung scan. The images acquired as part of the patients’ treatment are anonymized and sent to CNES’s Toulouse Space Centre. Using these data and artificial intelligence, the agency’s teams are developing a system to automatically detect the quantity and type of pulmonary lesions and predict the infection’s likely evolution with scans.

Eric Morand, in charge of CNES’s Computing, Software Engineering and Data Exploitation department, commented: “There are many cases of space research bringing significant benefits in our daily lives. The current health crisis we’re going through is no exception. Applying an effective solution that combines an ultrasound scan with artificial intelligence would ease the pressure in hospitals monitoring the health of a patient who has tested positive for COVID-19. A lung scan can be performed without having to move the patient from their bed, which solves all of the issues that go with transport, contagiousness and mobilization of healthcare personnel.”

Professor Stein Silva, an intensive care physician and the project’s clinical coordinator, added: “This novel disease is forcing us to innovate, and that means we have to forge new partnerships. The collaboration between Toulouse University Hospital and CNES is a perfect example of this in terms of its synergies and efficacy.”

Ultimately, if the ultrasound scan solution proves fully conclusive, it could enable better treatment of COVID-19 patients on admission to hospital, particularly in regions of the world where access to thoracic scanners is limited. The algorithm based on ultrasound scan data could also facilitate monitoring of hospital patients by detecting any respiratory complications.

This project shows how by combining the know-how of the space and health sectors, it is possible to develop new tools for citizens. Creating such synergies is central to the Connect by CNES initiative and to the MEDES space medicine and physiology institute, the CNES health subsidiary of which Toulouse University Hospital is a member.

Pascale Bresson    CNES    Tel. +33 (0)1 44 76 75 39    pascale.bresson@cnes.fr
Raphaël Sart    CNES    Tel. +33 (0)1 44 76 74 51    raphael.sart@cnes.fr

Dominique Soulié    CHU Toulouse    Tél. 05 61 77 83 49    soulie.d@chu-toulouse.fr
Mathilde Ratineaud    CHU Toulouse    Tél. 06 09 64 27 52    ratineaud.m@chu-toulouse.fr