Interview: University of Leeds on Realizing the Vision of In Silico Trials for Medical Devices

Can You Tell Us About Your Group?

We are part of the Centre for Computational Imaging and Simulation Technologies in Biomedicine (CISTIB) at the University of Leeds. The Centre has around 45 members, with backgrounds spanning computer science, engineering, mathematics, informatics, biological sciences, and medicine. In a nutshell, our remit is Computational Medicine, and in this context, we have developed image- and simulation-based solutions for numerous clinical problems, and across various clinical domains: cardiovascular, neurology, orthopedics, and oncology.

What Are the Main Themes/Goals of Your Research?

An overarching goal of our work is to realize the vision of in silico trials (ISTs) in the medical devices sector. ISTs, as virtual counterparts to conventional clinical trials, use computational modelling & simulation to assess the safety and efficacy of medical devices – and, critically, without direct involvement of real patients. Our goal therefore is to have ISTs established as integral parts of medical device lifecycles and embraced by industry, regulators, the clinic, and patients alike. 
The main work of our group is focused on creating the technologies that must underpin this vision. We address three main areas:

1. Creating virtual patient populations at scale, reliably capturing population anatomical, physiological, and pathological variations
2. Efficiently conducting virtual interventions on these populations, simulating patient/device interactions and functional performance; and
3. Predicting virtual outcomes, e.g., by establishing and computing surrogate endpoints for known clinical outcomes.

Where Do You See Your Research Having the Most Impact?

If fully realized, ISTs will fundamentally change the way medical devices are developed, assessed, and regulated. Clinical trials, which come at the end of normally lengthy and expensive research and development cycles, are themselves responsible for a large portion of the costs of getting new devices to market. They are also not a flawless means of evidencing device acceptability because they have practical limitations on duration, patient recruitment, and restrictive inclusion/exclusion criteria. The high-profile post-market failures of various devices in recent years are testament to this.

ISTs offer an alternative (cheaper, faster, and safer) means of generating evidence (so-called digital evidence) of performance/safety before a device ever goes near a patient. High-level impacts will include:

• Faster patient access to new devices through shortened development and assessment processes,
• Cheaper devices through concomitant reductions in development costs, and
• Safer and more innovative devices reaching market through more comprehensive testing and earlier de-risking (for manufacturers) of designs.

When Did You First Start Using Simpleware Software, and What Does the Software Help You to Achieve?

We have been using Simpleware tools for around 1½ years. This has been motivated especially by our current efforts to virtually recreate an existing phase III trial of a transcatheter aortic valve implant (TAVI). Simpleware software with the automated Heart CT segmentation and Valve Analysis tools, allow us to generate high-fidelity virtual twins of the trial participants from their CT angiograms, which are then used in simulations of the device deployment and hemodynamic function. These twins include the aortic root and left ventricular outflow tract anatomy, valve calcification patterns, and geometric features like aortic centerlines and key anatomical landmarks. Given the numbers involved (>700 participants), the ability to automate model construction, from images to high-quality computational meshes, is critical.

Where Do You See Your Research Heading over the Next Few Years?

As mentioned, there are some core technical areas we will remain focused on (virtual populations, interventions, etc.). There is much to be done for all of these. In parallel, we also aim to develop further exemplar studies, like the one for TAVI, that demonstrate equivalence of in silico evidence, reaching the same conclusions about device behavior as do comparable real trials. We want to show how ISTs can give us richer insights that are difficult or impossible to get in-vivo, but easy in silico, and how this can benefit the economy and the patients.

What Are Some of the Challenges That You Expect to Face with Your Research?

There will always be technical challenges to solve, but in a sense, these are the “easy” ones; at least, they are the ones we engineers and scientists are trained to address. The bigger challenge is how we build trust in in silico methods across the different stakeholders: manufacturers, regulators, clinicians, and patients. For these groups to have confidence in the reliability and value of ISTs, when clinical trials and other empirical methods have for long been the indispensable gold standards, requires a rigorous and well-targeted body of evidence. This approach needs to be rigorous because it must stand up to whatever scrutiny is imposed. It also needs to be well-targeted because it must answer in an accessible manner the questions those stakeholders are actually interested in. As a technical community, we need to engage with and listen to the end users to achieve these goals.

Learn More

• Read more about the CISTIB's research
• Simpleware software’s applications in the Life Sciences