Developing virtual twin hearts with Hybrid Cloud

YellowDog has worked closely with ELEM Biotech to grow their population of virtual twin hearts which can be used to test new devices and drug treatments in the human cardiovascular system.

Deaths from cardiovascular disease increased globally from 12.1 million in 1990 to 20.5 million in 2021 (the leading cause of death worldwide), according to the World Heart Federation.

ELEM Biotech has developed algorithms and a platform for the simulation of various systems in the human body. One of these algorithms was developed to produce virtual twins of the human heart. By being able to simulate the heart and effects of treatments on it, we are significantly better equipped to develop drugs, medical devices and surgical techniques, improving current therapies and designing new ones.

ELEM conducts clinical trials on human heart “avatars” in the cloud to predict outcomes in real clinical trials, resulting in a massive acceleration in the development of medical treatments. ELEM’s trials are validated against retrospective and prospective data from real trials. Using YellowDog’s platform, ELEM uses simulations to accurately target cardiac treatments and assess drug cardiac safety.

What will digital twin hearts be used for?

To complement and expand real clinical trials, with the following impacts:

• Speeding up the therapy safety & efficacy assessment, and the ulterior regulatory approval.
• Leading to future elimination of animal tests
• Providing powerful tools that can be used to explore new therapies at much lower costs, patient risks and times, impacting in an improved healthcare.
• Help de-risk medical device and pharmaceutical businesses
• Increase the representativity of demographics currently excluded from clinical trials.

Mariano Vazquez, Co-founder at ELEM, stated “The human heart is a highly complex organ. In particular, the contraction phase, which controls the pumping action, is challenging to model. Therefore, our virtual heart population requires exhaustive verification and validation, to ensure an accurate representation of the virtual population compared to the real one. This validation process requires massive testing. Without YellowDog, it would have been difficult to verify and validate such large number of virtual human twins”.

So how did ELEM use the YellowDog platform?

The onboarding and adjustment of existing techniques to take advantage of cloud compute was highly expedient, allowing for results to be delivered quickly. In total, ELEM ran 239 simulations, consuming 72,000 core hours on OCI (Oracle Cloud Infrastructure) Bare Metal 3.36 machines with 36 OCPUs, 512GB RAM and 3.8TB SSDs in two cloud regions in Europe.

ELEM was able to generate hundreds of data points, which made the importance of this massive calibration paramount. In the future, these tests will allow ELEM to use Machine Learning techniques to analyse results and feed them back, to adjust the virtual populations to the real ones even better.

Intense computing needs will be a requirement going forward for ELEM Biotech. Future work will use Machine Learning to create digital twins using the data points developed in previous simulations.

Large step changes in ground-breaking projects can get lost in the demand for compute. This risks three things: 1) project viability may not be uncovered in a timely manner, 2) time is lost usually in not being able to investigate and test different approaches, 3) sufficient confidence in technique may not be reached due to lack of testing resource.

These challenges are avoidable

Workloads of an infrequent but large nature, invariably require compute resources which are beyond the inventory that an organisation may have on-premise or be prepared to allocate to these workloads. Step in Hybrid Cloud and “pay as you go” pricing models which can sate the needs of these workloads but also be cost effective versus increasing on-premise spend to accommodate.

YellowDog’s differentiated value is enabling any organisation to rapidly onboard, scale and execute through intelligent compute workload management, operating small and large clusters in multiple clouds underpinned by automated DevOps and FinOps.

Removing the complexity enables teams like ELEM to concentrate the development and testing of virtual twin hearts to benefit us all. We’re delighted to have supported the ELEM team’s incredible work.