Speed is vital when triaging COVID-19, as quick detection saves lives. When the WHO declared the SARS COV-2 outbreak a pandemic in early March, Delft Imaging Systems realised their CAD4TB technology used to detect tuberculosis could easily be adapted to triage COVID-19. By the end of the month, they had adapted their artificial intelligence software to be able to detect COVID-19 abnormalities in the lungs.
CAD4COVID supports triage in settings where resources are otherwise limited. It is the first AI software that has been proven to detect COVID-19 to the same level as human radiologists. The system is suitable for use in patients who present with moderate to severe respiratory symptoms. The system allocates a score between 0 and 100 to indicate the extent of the damage to the lungs caused by the novel coronavirus. Using a heat map the colour image quantifies the percentage of damage. So damaged tissue is displayed in yellow and red while healthy tissue shows up blue.
The CAD4COVID technology is based on a revolutionary system to diagnose tuberculosis CAD4TB developed by Dutch company Delft Imaging Systems in collaboration AI startup Thirona and the Radboud University in Nijmegen. The software can be used for digital X-rays or CT scans. By partnering with German company Smart Reporting, Thirona has made the CT software accessible to over 10,000 radiologists.
Throughout the corona crisis resources for testing and personal protective equipment have been scarce. Using CAD4COVID technology means patients with suspected COVID-19 can be triaged rapidly and safely, without have to wait for test results. It also saves medical staff time and enables them to track the progress of patients on the mend.
The CAD4COVID technology was first used at a hospital in the Netherlands’ worst-hit region Brabant. Since then the system has been helping hospitals and medical centres handle COVID patients as far as Peru, Canada and Hungary. In Ghana, CAD4COVID has been installed in 55 general hospitals with the help of the Netherlands Enterprise Agency through its Dutch Good Growth Fund (DGGF).
Mieke Flierhuis of DGGF at Netherlands Enterprise Agency explains, “Delft Imaging has developed an artificial intelligence tool to support the triage of covid-19 cases in resource-constrained settings. In close cooperation with the Ministry of Health in Ghana, CAD4COVID will be installed in 55 district hospitals across the country. CAD4COVID analyses images of the lungs and detects COVID-19 related abnormalities. The Netherlands Enterprise Agency, through its Dutch Good Growth Fund, supports this project.”
Join the programme
The collaboration with the various hospitals and medical centres across the world will help to validate the technology. With the consent of the participants, the data produced and learnings will be used to constantly update the software.
To help save lives, Delft Imaging and Thirona have made use of the technology available across the world free-of-charge. If you work for a medical institution or hospital, you can click on the button below to join the programme.