Microsoft Investigator Fellow Dr. Casey Frankenberger at Rush University uses the Azure cloud to deliver better healthcare outcomes

Migrating to Azure

As part of the development of the medical records project, Rush began migrating projects to Azure. “We decided to go to the cloud in January of 2020, and it could not have been a more fortuitous decision,” said Dr. Frankenberger. “This was not in anticipation of COVID, it was just pure luck. But what we found was that everybody on my team was as productive, or more productive, once we started working from home.”

The ability to access and share resources via the cloud was just the beginning. As the pandemic evolved into a full-blown healthcare emergency, Rush became the central data repository for COVID-19-related patient data from Chicago area hospitals. This data portal, called The Commons, was described by Chicago Department of Public Health Commissioner Allison Arwady, MD, as an “inspired new level of collaboration between the Chicago’s many health care organizations.” The common data sets provided healthcare leaders a much better understanding of which hospitals have been treating COVID-19 patients and how the virus was spreading.

Dr. Frankenberger said that being able to use secure cloud computing infrastructure for handling confidential patient information is one of the reasons Azure was a good fit, for both the medical records project and for developing a registry of Chicago area COVID-19 patients. Rush set up the COVID-19 registry in March 2020 and is using the cloud to store, manage, and analyze that data. “We have all those patients consented,” said Dr. Frankenberger, “so we can use all of their data for research. We also have a biorepository on tens of thousands of those patients, with specimens, so we can use those for research, too.”

Using Azure’s AI-ready framework

Facilitating collaboration in the cloud was a key consideration for Dr. Frankenberger and his team even before the pandemic. An important driver in the medical center’s decision to use cloud solutions to integrate genomic data into medical records was Azure’s proven ability to help secure and protect sensitive data.

The challenge was to create an environment in which machine learning models could be developed using anonymized patient data that never leaves the Rush cloud ecosystem. “We wanted to allow people to bring their algorithm into our cloud environment, train the algorithm, and then leave with the algorithm, but not leave with any of the data,” said Dr. Frankenberger.

What Dr. Frankenberger called an “AI-ready framework” emerged from the creation of a customized cloud computing environment in Azure. This is enabling researchers to use data from close to a million oncology patients to mine that treasure trove for insights that will improve the quality and continuity of care.

Azure provides the Rush team with the high-performance computing (HPC) resources needed to use machine learning tools for data analysis and the output of real-time insights. HPC includes a complete set of computing, networking, and storage resources integrated with workload orchestration services for related applications. Azure AI also enables data scientists to collaborate and build models faster. Machine learning models can be built, trained, and deployed using Azure Machine Learning, Azure Databricks and ONNX Runtime. Another valuable tool is Azure Cognitive Search, which enables Rush researchers to uncover latent insights from documents, images, and media. 

Easily enabling data and compute growth

In addition to benefits such as powerful computing capacity and ready-to-use AI tools, migrating to the cloud has been a game changer for Dr. Frankenberger when it comes to deploying new information technology resources. 

Dr. Frankenberger highlighted the scalable, on-demand services available in the cloud. In a medical research environment, those on-demand services range from providing new compute space for a project and extracting structured information from physician notes, to analyzing treatment protocols or outcomes.

Dr. Frankenberger said the Department of Cell and Molecular Medicine has transitioned from using a single server to having infinite computing resources available. Their team now includes experienced cloud architects and data specialists to optimize efficiency, and the cost recovery ratio for the department—which measures how an investment recoups it costs—has gone from zero to 75%.

But for Dr. Frankenberger and his team, better patient outcomes are an important benefit of cloud computing. “We want to find things like a needle in the haystack relative to patient wellness and their genetic constitution, the variability in their care, and then the variability or complexity in their disease phenotype,” he said. 

Data is the key. Dr. Frankenberger and his oncologist colleagues are looking forward to having access to structured genomic data to help them with predicting cancer risk, formulating a prognosis, and assessing how a patient responds to treatment. Researchers have long known that cancer cells contain the clues physicians need to develop treatments that are customized based on a patient’s DNA. Now, integrating genomic test results into electronic medical records gives doctors a powerful new weapon in the fight against cancer.

Supporting academic research

Microsoft is helping academic researchers like Dr. Frankenberger benefit from computational resources in new and innovative ways. Dr. Frankenberger is one of 15 research fellows in the Microsoft Investigator Fellowship program, a two-year program that began in fall 2019 and is one of several Microsoft initiatives designed to accelerate research. Other Microsoft programs include AI for Health—which provides grants that support using AI to address global health challenges—and AI for Earth, which offers funding for research into how AI can change the way people and organizations monitor, model, and manage the Earth’s natural systems. In addition, the Microsoft Research Academic Programs support research opportunities for faculty, students, and collaborating institutions.

“Microsoft is investing in the long-term future of higher education in the United States, and in support of research on artificial intelligence,” said Dr. Lawrence Carin, currently Provost at King Abdullah University of Science and Technology (KAUST) in Thuwal, Saudi Arabia. Dr. Carin helped Microsoft launch the Investigator Fellowship program while serving as Vice President for Research at Duke University in Durham, North Carolina. 

“Research is inherently collaborative. You can't do research and have a gated wall around it,” added Dr. Frankenberger. “When I look at our Microsoft partners, they want to learn about precision medicine, value-based care, and health technology. And when I look at all the people on my team, they want to learn about cloud computing, cloud architecture, and machine learning. Everybody's pulling together and learning—and that's when innovation happens."