RSNA joins ‘Imaging COVID-19 AI Initiative’ to Use AI for Beating COVID-19

It takes a virus to throw the world into disarray. It takes an international network of top doctors, physicists, mathematicians, engineers and entrepreneurs to work against the clock to pull it together.

Monday, the Radiological Society of North America (RSNA) announced its decision to partner with the ‘imaging COVID-19 AI initiative’ in applying AI to detect advanced cases of COVID-19.

By doing so, this international society of more than 54,000 members joins 28 other organizations that include the European Society of Medical Imaging Informatics (EUSOMII) and the Netherlands Cancer Institute (NKI), along with entrepreneurs and hospitals from Europe and the Middle East.

The Imaging COVID-19 AI initiative

The story began July 2018 when Robovision.ai CEO and Founder, Jonathan Berte met Erik Ranschaert, President of the European Association of Radiologists (Eusomii) at a Medical Imaging with Deep Learning (MIDL) conference in Amsterdam. Two weeks ago, they launched their Imaging COVID-19 AI initiative to train AI to detect COVID-19 in lung scans.

Spain-based Quibim, a company that specializes in medical imaging data and 28 other hospitals from countries that included Italy, Spain, Belgium, England, the Netherlands and Iran pitched in. Entrepreneurs such as Microsoft, Google and Nvidia also registered to help with extra computing power.

The COVID-19 Scanning Problem

While computed tomography (CT) scans are faster and more reliable than X-rays in detecting lung problems, they’re still too slow, especially in overcrowded and short-staffed hospitals such as those in Italy, Madrid and New York.

RoboVision’s AI platform has successfully helped hundreds of companies in Agriculture, Manufacturing, Healthcare and Governance detect production flaws and with intelligent automation of complex, in-line decisions.

Berte and Ranschaert now used their consortium of 30 hospitals and growing – half of which are linked to academic institutions – to supply them with hundreds of CT scans of patients diagnosed with COVID-19 infections. This data is used by Robovision.ai to train its AI algorithm, after which it will be able to perform faster analyses of new scans. Quibim created the web interface to upload deliberately anonymized patient data.

Monday, the Initiative expanded to receive the Radiological Society of North America (RSNA), an organization of radiologists, medical physicists and other medical professionals. This prestigious organization publishes two peer-reviewed journals and has funded over $60 million in radiology research around the world.

According to its website:

  • The RSNA is committed to connecting radiologists and the radiology community to the most timely and useful COVID-19 information and resources
  • The Team’s Pledge: To Accurately Detect the Disease in Legions of Untested Patients
  • The AI Initiative-RSNA team expressed their unflagging determination to train AI algorithms, without spreading the infection through contaminated equipment.

 

As Peter Van Wijnaerde, Marketing Manager at Robovision.ai said:

We will collaborate to enable hospitals to provide imaging data securely and efficiently with the researchers, respecting privacy and ethical principles. We will use protocols for selecting and labeling imaging data to create a tool for researchers and medical practitioners to be used on lung images in the fight against COVID-19.

On top of that: “We will also invite other interested organizations to join this coalition to share information and facilitate a rapid response to COVID-19 by the imaging and AI research community.”

Robovision.ai works to have a reliable and tested tool in three weeks’ time, sped up to help the waves of patients.

“In medical terms, that’s almost the speed of light,” Van Wijnaerde said.

“It is wonderful, “ Berte added, “to see how fast many academics and entrepreneurs are joining forces in a short time to achieve something for the common good.”

Maybe it will be this unifying effort that defeats this divisive virus.