AWS Public Sector Blog

How the cloud is powering fast, scalable diagnostics in the fight against COVID-19

In response to the COVID-19 pandemic, healthcare organizations around the world are focusing on improving and speeding up testing and diagnostics. Digital health companies Smart Reporting in Germany and Thirona in the Netherlands have been working to create a CT-based imaging solution to support COVID-19 diagnosis, enabled by the cloud and the Amazon Web Services (AWS) Diagnostic Development Initiative. The AWS Diagnostic Development Initiative provides support to organizations for innovation in rapid and accurate patient testing for 2019 novel coronavirus (COVID-19) and other diagnostic solutions to mitigate future outbreaks.

Test shortages require new diagnostics

By March 2020, hospitals throughout Europe were filled with suspected COVID-19 patients showing a variety of symptoms, from a dry cough to a fever, making it hard to diagnose without testing. The speed of the outbreak resulted in a severe shortage of tests in many countries, leading many medical experts to look at different ways to diagnose the disease.

“The guidelines were clear: COVID-19 had to be diagnosed with a test, so for everybody, a swab through the throat and the nose was sent to the lab. Even in the best case, it takes hours before you get the results of that test and in reality, it was usually days,” explained Professor Dr. Bram van Ginneken, professor of medical image analysis at Radboud University Medical Center in the Netherlands and co-founder of Thirona.

Dr. van Ginneken established in 2012 to organize machine learning challenges in biomedical image analysis. Today, it brings together more than 45,000 registered researchers and clinicians from all over the world. With his expertise, Dr. van Ginneken started to analyze medical images of confirmed COVID-19 patients, which showed a distinct, recognizable appearance on CT scans.

Fast research to product

In conjunction with Radboud University Medical Center, Thirona created an AI algorithm to diagnose COVID-19 on chest CT scans and X-rays. Thirona automated all the analysis steps from a standardized scoring system—called CORADS—that is now used in every hospital in the Netherlands to allow radiologists to arrive at a diagnosis. Thirona’s AI solutions, CAD4COVID-CT and CAD4COVID-xray, are freely available to anyone in the medical community. Thirona was able to quickly research and develop their algorithms, and make them available to the medical community, thanks to donated computing credits and technical support provided through the AWS Diagnostic Development Initiative.

Meanwhile, Smart Reporting was also investigating the use of CT scans for COVID-19 diagnosis. They designed a template to report findings of pulmonary opacities in CT scans due to COVID-19. This reporting template includes a decision tree based on the latest guidelines and recommendations from the British Society of Thoracic Imaging (BSTI) and the Radiological Society of North America (RSNA). The template is available internationally in English, German, French, Spanish, and Portuguese. The aim was to standardize medical documentation and support image analysis and make medical documents comparable between patients, hospitals, and countries.

Using the cloud to scale

United by the common objective of fighting COVID-19 and the common use of AWS, these projects came together to create machine learning-enabled solutions supporting diagnosis of  COVID-19. The result is SmartCAD | COVID-19. The integrated algorithm by Thirona scans the CT scans for abnormalities. Then radiologists can then choose to integrate this information into Smart Reporting’s structured COVID-19 reporting solution.

Dr. Eva van Rikxoort, managing director and co-founder of Thirona, said, “We used our existing expertise and experience in medical image analysis to rapidly develop a solution to detect COVID-19. By integrating it with the platform offered by Smart Reporting, we were able to scale it quickly to an audience of over 10,000 radiologists.”

The solution makes radiologists’ workflows more efficient by offering automated image analysis. The platform is hosted on AWS and is accessible worldwide at no cost to radiologists. SmartCAD | COVID-19 uses Amazon Simple Storage Service (Amazon S3) with accelerated transfers to gather data from sites globally, while Amazon CloudFront is used for fast data access. Amazon Elastic Compute Cloud (Amazon EC2) is used to deploy image rendering servers across the world that can start containers on demand for each new user and scale compute horizontally in each Region using Auto Scaling groups. To reduce the numbers of services required, Amazon Simple Queue Service (Amazon SQS) is used as message broker.

“It usually takes several years until academic research translates into a real, viable product. Here, we did all of that in three months to yield interacting products. This was possible due to cloud technology since we could all use the same infrastructure,” said Dr. Wieland Sommer, Smart Reporting founder and chief executive officer.

Enabling diagnostic research

Other research groups around the globe have been leveraging the AWS platform to build machine learning diagnostic tools to fight the COVID-19 pandemic including the University of British Columbia (UBC) and Vancouver General Hospital (VGH) and UC San Diego Health.

Learn more about the AWS Diagnostic Development Initiative and healthcare for the public sector.

Razvan Ionasec

Razvan Ionasec

Razvan Ionasec, PhD, MBA, is the technical leader for healthcare at Amazon Web Services in Europe, Middle East, and Africa. His work focuses on helping healthcare customers solve business problems by leveraging technology. Previously, Razvan was the global head of artificial intelligence (AI) products at Siemens Healthineers in charge of AI-Rad Companion, the family of AI-powered and cloud-based digital health solutions for imaging. He holds 30+ patents in AI/ML for medical imaging and has published 70+ international peer-reviewed technical and clinical publications on computer vision, computational modelling, and medical image analysis. Razvan received his PhD in Computer Science from the Technical University Munich and MBA from University of Cambridge, Judge Business School.