Project EQUIVALENCE: Revolutionising Medical Device Testing with Virtual Populations Medical device testing is at a critical juncture. Traditional clinical trials, while valuable, are often limited in scope, time-consuming, and expensive. They may not adequately represent diverse populations, potentially leading to unforeseen complications when devices are widely deployed. EQUIVALENCE aims to address these challenges by refining, translating, and demonstrating the efficacy of a unique approach to creating high-quality, high-volume synthetic virtual populations for medical device testing. At the heart of EQUIVALENCE is a groundbreaking in-silico trials capability that leverages advanced generative AI modelling. This innovative technology creates synthetic virtual patients that closely mimic the diversity and complexity of real-world populations. Our aim is to demonstrate that these in-silico trials can replicate key safety and efficacy outcomes of conventional clinical trials, but at a fraction of the cost and time. The project focuses on three main objectives: 1. Refinement: We will enhance our existing virtual population generation technology, improving its accuracy, diversity, and scalability. This will involve integrating new data sources and advanced AI algorithms to create more comprehensive and representative virtual patient cohorts. 1. Translation: We will develop protocols and workflows to translate traditional clinical trial designs into in silico formats. This includes creating virtual analogues of medical interventions and defining virtual endpoints that correspond to real-world clinical outcomes. 1. Demonstration: We will conduct a series of virtual trials, paralleling completed real-world clinical trials for specific medical devices. By comparing the outcomes, we aim to validate the equivalence of our in-silico approach to traditional methods. The potential impact of EQUIVALENCE is substantial. Successful demonstration of this technology could accelerate medical device development, reduce costs, and most importantly, improve patient safety by enabling more comprehensive testing across diverse virtual populations. It could allow for the exploration of rare but critical scenarios that are difficult to encounter in traditional trials. Moreover, this approach aligns with ethical imperatives to reduce animal testing and minimise risk to human participants in early-stage trials. It also offers the potential to democratise medical device development, allowing smaller companies to conduct comprehensive testing that was previously only feasible for large corporations. EQUIVALENCE represents a significant step towards a future where medical devices are developed faster, tested more thoroughly, and designed with greater consideration for diverse patient populations. Our goal is to set a new standard in medical device testing, ultimately benefiting patients worldwide through safer, more effective medical technologies.

**ad**silico, a spin-out from the University of Leeds (UoL), is aiming to Define, Disrupt and Dominate the in-silico trials space to accelerate medical device innovations. It will become the world's first end-to-end in-silico trials service provider. adsilico's technology for creating virtual populations (VP) of anatomy and physiology builds upon 15+ years of research conducted at CISTIB, Centre for Computational Imaging and Simulation Technologies in Biomedicine from UoL. The team conducted the first in-silico trial (IST) for devices used to treat cerebral aneurysms (Sarrami-Foroushani, A, et al, 2021). The performance of flow diverters was assessed in a VP comprising cerebral vessel and aneurysm geometries and associated blood flow waveforms. The results demonstrated that ISTs can replicate and expand on the insights gained from equivalent real clinical trials, at a fraction of the cost and time, and with no real patient involvement. Impact through industrial translation of the research requires enhancement of the prototype image computing workflows for creating the vessel/aneurysm VPs, so that outputs are of suitable quality and reliability for commercial use. The team has a proven track record of conducting large-scale image analysis studies and building virtual patient populations, especially in the cardiovascular and neurovascular domains. For example, we recently analysed 40k subjects' cardiac magnetic resonance (CMR) images across 50 time points of the cardiac cycle, amounting to 2xmillion CMR image volumes (Xia, Y., et al, 2022). We are the only group in the world to have undertaken such analyses at this scale. This resulted in a curated data set of 2 million 3D full heart geometries and the associated, validated workflows used to create the same. This project aims to refine and translate a prototype for creating high-quality cerebrovascular and aneurysm virtual patient populations that are suitable for use within in-silico studies. **Key deliverables:** 1\. Securing early adopters. 2\. Define QA processes for compliance with applicable regulations/standards and for audit purposes. 3\. Cerebral aneurysm patient data access, curation/preparation. 4\. Strategy for automating management of primary/derived data. 5\. Enhancement/validation of workflow for creating cerebrovascular and aneurysm VPs and design-controlled integration (to execution platform). 6\. Create simulation ready VPs for neurovascular devices (target 5000 instances). 7\. Validation of VPs with customer-driven exemplar. 8\. Design/implement secure environment to access/store/manage digital assets. 9\. Contractual frameworks (**ad**silico access to data and customer use of derived data). 10\. Determine commercial value of cerebrovascular VPs. 11\. Validation of commercial model.