Introduction 

In the ever-evolving landscape of medical research, the integration of artificial intelligence (AI) in clinical investigation holds immense potential to revolutionise the way clinical trials are conducted.¹ With its ability to analyse vast amounts of data, identify patterns, and provide increasingly accurate predictions, AI can streamline processes, enhance efficiency and decision-making, optimise dose-finding, identify novel biomarkers and enable real-time trial data monitoring.^2–4 These capabilities offer immense potential for accelerating the drug development process, identifying patient subpopulations that may benefit most from a treatment, and enhancing the overall efficiency and success rates of clinical studies.^2–5 

Here, we explore the various ways AI can be incorporated into clinical trials, with a particular focus on the groundbreaking concept of digital twins and synthetic control arms. 

The advent of digital twins and synthetic control arms 

One key example of the potential impact of AI on clinical trials is the substitution of traditional control arms with digital twins or synthetic control arms.^6,7 A digital twin can be defined as the digital mapping of an individual – a model that simulates the behaviour of its physical counterpart with high statistical probability;⁸ a synthetic control arm is generated collating patient-level data (for example, electrocardiogram readings or optical coherence tomography [OCT] scans) and the efficacy of a therapeutic option is evaluated against an external database.⁹ By leveraging deep learning algorithms, large synthetic control arms can allow for more-robust comparisons, reducing the need for placebo or control groups.¹⁰ Professor Aaron Lee from Washington University is at the forefront of this groundbreaking development in the retinal health space, collaborating with OCT reading centres, pharmaceutical companies and the FDA to establish, validate and integrate these novel processes into the regulatory environment.¹¹  

The potential of this technology extends beyond retinal health. In rare diseases and in other therapy areas where recruitment is particularly challenging, ethical concerns are present, and the current standard of care, when available, falls short, AI can harness the wealth of existing research and monitoring data to create digital control arms. This approach allows for a more comprehensive analysis of treatment efficacy, potentially uncovering novel insights and expediting the discovery of effective therapies.¹ Importantly, embracing this paradigm shift requires a change in mindset from pharmaceutical companies, fostering a culture of data-sharing while ensuring intellectual property protection.¹² 

Validating the process 

To build confidence and validate the integration of digital twins or synthetic control arms in clinical trials, a cautious approach where they are employed alongside traditional control arms will likely be required. This proof-of-concept phase enables the evaluation of AI-driven technology’s accuracy, reliability and feasibility. Because most AI solutions evolve independently through machine learning algorithms, that is, through iterative refinement processes that enable the AI to become more precise and accurate over time as it receives new data, pharmaceutical companies can gain the necessary trust and momentum to fully embrace this transformative approach, by gathering real-world evidence and demonstrating concrete benefits.¹³ 

Shaping the future of clinical research 

Looking ahead, the integration of AI holds tremendous promise for shaping the future of clinical research. By removing the need for placebo or control groups, AI technology has the potential to eliminate psychological bias, ensuring that all patients receive the active compounds under investigation.¹ This shift in approach has the power to enhance patient experiences, improve treatment adherence, and ultimately lead to better outcomes and a preserved or elevated quality of life. 

Moreover, AI-powered algorithms can continuously learn from real-time data, adapting treatment protocols to individual patient needs.¹³ This personalised approach has the potential to optimise dosing, minimise adverse events and improve overall treatment efficacy. As AI technologies evolve, researchers can unlock a treasure trove of insights, empowering them to make informed decisions, accelerate drug development and shape the future of precision medicine.^2–5 

Medical Affairs: from data to evidence to action 

AI is poised to revolutionise the landscape of clinical trials and holds an unprecedented potential to enhance patient outcomes and improve treatment efficacy and safety.¹ By seizing the myriad of opportunities AI can offer, we can harness its power to uncover novel insights and ultimately shape the future of medical research.^6,7 Some pharmaceutical companies, such as Sanofi, are already starting to take digital transformation in their stride, developing and integrating AI-powered solutions in their operational framework in a cross-functional manner – from research to clinical development to manufacturing and supply.¹⁴ Medical Affairs teams hold the perfect compass to help navigate this exciting frontier, spearheading collaboration between academia, industry and regulatory authorities to ensure that a successful integration of AI in drug development is translated into tangible initiatives and leads to concrete actions. Together, we can unlock the full potential of AI and pave the way for a new era of evidence-based, patient-centred medicine.  

If you have any questions or would like to hear more from our Medical team, reach out to Joanne Pearson via email at joanne.pearson@cognite.co. 

References 

1. Bajwa J, et al. Future Healthc J. 2021; 8(2):e188–e194.  
2. Askin S, et al. Health Technol (Berl). 2023;13(2):203–213. 
3. Goldenberg SL, et al. Nat Rev Urol. 2019;16(7):391–403. 
4. Yaghy A, et al. Exp Eye Res. 2022;220:109092. 
5. US Food and Drug Administration. https://www.fda.gov/medical-devices/digital-health-center-excellence. Accessed 26 July 2023 
6. Venkatesh KP, et al. NPJ Digit Med. 2022;5:150. 
7. Armeni P, et al. J Pers Med. 2022;12(8):1255. 
8. Yao JF et al. Vis Comput Ind Biomed Art. 2023;6(1):10. 
9. Thorlund K, et al. Clin Epidemiol. 2020;12:457–467. 
10. Lee CS, et al. Biomed Opt Express. 2017;8(7):3440–3448. 
11. Lee AY, et al. Diabetes Care. 2021;44:1168–1175. 
12. Applied Clinical Trials Online. https://www.appliedclinicaltrialsonline.com/view/why-decentralized-clinical-trials-are-the-way-of-the-future. Accessed 26 July 2023. 
13. National Library of Medicine. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7325854/. Accessed 26 July 2023. 
14. Sanofi. https://www.sanofi.com/assets/dotcom/pressreleases/2023/2023-06-13-12-00-00-2687072-en.pdf. Accessed 26 July 2023.