AWS helps accelerate scientific initiative to predict COVID-19 patients’ trajectory
A new, multi-center Italian research study that aims to predict the prognosis of hospitalised COVID-19 patients, is accelerating its work with support from AWS. The non-profit project, called AIforCOVID, is the result of a public private partnership between the Centro Diagnostico Italiano, diagnostic imaging company Bracco Imaging, and a number of Italian Institutes and hospitals. The project’s primary purpose will be to create an artificial intelligence model able to forecast the trajectory of a patient’s disease, using chest x-ray images and medical records as training input.
It will provide doctors with early indications of a patient’s future clinical course, helping them to identify those likely to experience a serious form of COVID-19, and intervene faster with an appropriate level of care. The technology will also help hospitals triage patients more effectively, and to more accurately allocate resources by planning for the right number of intensive and sub-intensive care units.
The project, which will use a range of AWS technologies including advanced machine learning and artificial intelligence, as well as data and analytics services, has three main phases. The current first phase will be the development of a technology platform hosted on AWS for hospitals and universities to collect and store anonymised chest x-ray imagery. The second will see the creation, training, and distribution of machine learning models to analyze these images by leveraging Amazon SageMaker, a fully managed service to build, train, and deploy machine learning models, which will be providing indications on the patient's future clinical course.
Finally, the third phase aims to create an analytics system to generate reports and interactive dashboards, enabling the discovery of hidden trends and outliers, identification of patterns across time-series data, and forecasting.
The ultimate goal is for the project to serve as a technological and collaborative model that may be applied to future disease outbreaks, not only COVID-19.
The great value of the AIforCOVID project, lies not only in the immediate use of the platform to respond to COVID-19, but in the fact that we have built a new technological, regulatory, and operational framework that will be a point of reference for future projects," explains Fulvio Renoldi Bracco, Bracco Imaging CEO. “This is a pioneering initiative, which we have been able to launch thanks to support from the AWS Diagnostics Development Initiative.
“There is still a great deal that we need to understand about COVID-19, and it is exciting to see the AlforCOVID project tackle the project in a way that advances our understand of this virus in the near-term while setting up a framework that will have a lasting benefit to the scientific community,” said Teresa Carlson, Vice President of the Worldwide Public Sector at AWS. “We are taking a similar approach at AWS by staying focused on using the power of the cloud to tackle the challenges posed by the pandemic, while also thinking about the ways we can help the global recovery effort, and continue to innovate for the long-term.”
The project is being supported by donated computing credits provided by the AWS Diagnostic Development Initiative, a $20 million global program launched in March to support AWS customers working to bring more accurate diagnostics solutions to market faster.