Insight at the University of Galway Pioneers the Future of Aortic Care with Cloud-Based Diagnostics on AWS

How we helped

The project's goals were multifaceted, but two main challenges stood out: the daunting computational process of modeling blood flow in the aorta and efficiently handling the massive, intricate three-dimensional CT scan images. Niamh Hynes, Consultant Vascular Surgeon and Senior Research Fellow at the University of Galway and Galway Clinic, shed light on the size of the task: “What you have are very large files, hundreds of gigabytes for a full CT scan. If we want to look at the biomechanical properties of the aorta, currently that takes up to 150 hours using a supercomputer.”

Before the partnership, research was often constrained by the limitations of local computing resources, both in cost and technical capability. But with the backing of Rackspace Technology and AWS's cloud infrastructure, the Insight team found solutions for secure storage and scalability. The offerings weren't just about storage space; they also focused on cost optimization, specialized consultation, and valuable training.

The ability to scale computational resources on-demand was a game-changer during the aortic blood flow modeling. Now, the team could seamlessly and securely engage with patient data across various storage solutions. There was a time when processing these models took around 120 hours for each sample. Today, with the leverage of public cloud scalability coupled with AI/ML services, a more streamlined architecture is in place. The use of Amazon SageMaker and Apache Spark has expedited the computational modeling process, offering avenues to fine-tune model construction and achieve enhanced predictive results.

Andy Donald, Research Fellow at the Insight SFI Research Centre for Data Analytics, University of Galway, highlighted the drastic improvements and their implications: “We are researching how to emulate the complex computational modeling process as a ‘surrogate model’ to reduce processing time from approximately 150 hours to around 10 minutes. This near real-time potential advance would allow a surgical team to progress with more informed data.”

The secure storage of diverse patient data — spanning medical images to clinical information — was a priority. Traditionally, such data found its home on university server-based physical hard drives. Cloud storage has transformed this, offering not only fortified security but also quicker, more efficient data distribution to collaborators. Donald went on to emphasize the security aspect: “Given the sensitive nature of this project involving medical patient data, it is especially critical to ensure complete security and anonymity. We store full-body CT scan files of hundreds of gigabytes in Glacier to optimize efficiency and rapid access for further analytics. With Amazon SageMaker, we tackle the most demanding workloads, easily scaling resources as needed.”

While cloud technology offers tremendous potential, harnessing it requires a specific skill set. This led to tailored training for Insight’s data scientists and engineers, ensuring they were proficient with vital cloud technologies for the project.

Curry pointed out the challenges and the solutions. "Ensuring our team had the necessary skills and were up to date in using cloud technologies was vital,” he said. “Partnering with Rackspace Technology and AWS has given us access to the training our data scientists and engineers needed, enabling them to develop cutting-edge AI/ML solutions alongside our industrial partners."