Skip to main content
Healthcare Healthcare

“Super-brain” of Robovision Helps to Detect COVID-19

Date Section Blog

The most commonly used test to determine whether someone has the COVID-19 is the cotton swab test. However, it takes at least a few hours to receive an accurate result. A CT-scan of the lungs can provide a more reliable indication of infection more quickly. European radiologists therefore want to use artificial intelligence (AI) to be able to work faster, and are therefore calling on a Ghent technology company to develop a kind of “super-brain”.

CT-scanners are sufficiently accurate to recognise the damage in the lungs from the COVID-19. We learnt that from the experiences in China and other areas where the disease is more common. We can even see abnormalities before the patient has symptoms.

We wondered whether we could perhaps better analyse those CT scans with digital tools to be able to make that diagnosis with a certain degree of certainty. And with that thought in mind, we came up with the idea of developing an algorithm. A software that helps us.

Dr Erik Ranschaert, radiologist at the Elisabeth-TweeSteden Hospital (ETZ), Tilburg, the Netherlands

Ranschaert joined forces with Jonathan Berte, the founder of the Ghent-based Robovision, one of the leaders in Europe in the field of AI. Together with researcher Laurens Topff, they set up an international consortium of thirty European hospitals and the authoritative Netherlands Cancer Institute (NKI).

People are very good at recognizing patterns, but are less good at quantitatively calculating damaged lungs. And that is exactly what this AI model will be able to do very well.

Jonathan Berte, CEO and co-founder of Robovision

The large-scale project with Robovision has meanwhile also attracted attention in the United States. The two main American radiologists’ associations, Radiological Society of North America (RSNA) and American College of Radiology (ACR), have already expressed their willingness to cooperate with European-based projects.

The radiology lab of the renowned Stanford University led by the Belgian professor Olivier Gevaert is also interested. In addition, Belgian Federal Minister of the  Digital Agenda, Telecommunications & Post, Philippe De Backer, is trying to convince European Commissioner for Internal, Market Thierry Breton, to involve even more European member states.

Radiologists point out that the AI tool is not intended to replace them, but that it could primarily be a tool to provide more guidance in a rapid diagnosis of Covid-19. Especially now that hospitals are becoming increasingly overloaded. “We will still be required to fully interpret the scans. Because don’t forget that artificial intelligence cannot currently detect Covid-19 and a lung tumor at the same time,” says Dr Ranschaert.

Reference

VAN DE WEGHE, T (2020) “Superbrein” van Gents technologiebedrijf moet coronavirus helpen detecteren [Online] 8 April 2020. Available from: https://www.vrt.be/vrtnws/nl/2020/04/03/superbrein-moet-coronavirus-helpen-detecteren/[Accessed 8 April 2022]