How data and AI are accelerating cancer clinical trials

Phesi is leading the charge in using data and AI to put patients at the center of innovation. The company has spent two decades building the industry’s most comprehensive, contextualized clinical trials database, empowering partners to design smarter, faster, and more equitable studies.

Through its Trial Accelerator™ platform, Phesi leverages real-world data and AI-driven analytics to create digital patient profiles, digital twins, and predictive models that improve trial efficiency and success rates.

Dr. Gen Li, Phesi’s founder, shared his insights into how patient-centric data science is transforming global healthcare outcomes, the persistent challenges in oncology research, and how AI is reshaping the future of clinical trials.

How does Phesi’s focus on patient-centric data analytics fit into the bigger picture of improving global healthcare outcomes?

GL: Phesi’s guiding principle is that the patient must be at the heart of clinical development; everything starts with gaining a deep understanding of the individual subject and targeted population. With patient-centered data science, sponsors can optimize clinical trial design, reduce patient burden and select investigator sites and enrolling countries with precision. This data-driven approach is essential for sponsors to withstand the increasing complexity of clinical development, accelerate cycle times and control costs, helping get much-needed treatments to patients faster.

Tell us about your Trial Accelerator™ platform. What makes this approach to clinical trial data unique?

GL: Trial Accelerator is Phesi’s clinical trials database. With a global reach spanning over 300 million patients across 195 countries, we capture a diverse and comprehensive picture of disease, covering over 4,000 indications, including rare and ultra rare diseases. We use these data to create Digital Patient Profiles and our unique Patient Access Score. Digital Patient Profiles are the foundation from which we generate Digital Twins and External Control Arms.

Our approach is unique because it combines scale, context and intelligence. We maintain the industry’s largest source of contextualized real-world data, where every piece of patient information comes with context on who collected it, when it was collected, where it was collected, and how it was collected and the design of the study.

This breadth of data is coupled with AI-powered technology and proprietary algorithms we’ve been developing for twenty years, enabling insights that drive smarter trial design and accelerate development. Our data has already proven its impact in helping to deliver innovative medicines such as Keytruda® and Besponsa®.

Your most recent annual global analysis showed that breast cancer is the world’s most studied disease for the fourth consecutive year. What are the key reasons for this?

GL: Breast cancer remains the most studied disease because of persistent unmet needs of patients and continuous medical innovation resulting in improved understanding of this disease. As the leading cause of cancer-related death among women, it represents both a major clinical challenge and an area of significant research opportunity. The combination of clinical urgency, investment and a large patient population continues to drive extensive research.

Can you provide any other insights from the report about breast cancer and the other highly researched cancers?

GL: Today, we have a much deeper understanding of the molecular subtypes of breast cancer, including Luminal A and B, HER2-positive, and Triple-Negative. These insights have driven an increase in biomarker-specific trials. A similar progression has happened in NSCLC (non-small-cell lung cancer), which explains its re-entry into the most studied diseases of last year (2024).

However, despite such advances, breast cancer trial expertise remains disproportionately concentrated in just two countries – the US (35%) and China (15%) – while lower-income countries continue to suffer from increasing mortality rates. Recent data from the WHO showed that by 2050, there will be 1.1 million breast cancer-related deaths per year, with the increase disproportionately affecting low Human Development Index countries.

If sponsors combine the growing knowledge of disease biology with advances in AI and clinical data analytics, trial design will be more precise and effective. The benefits are twofold. Firstly, more efficient and successful trials, with increased ROI. Second, and most importantly, this approach ensures that breast cancer clinical development is patient-centric and distributed equitably around the globe.

Phase II attrition rates remain stubbornly high. Why does this trend persist?

GL: Phase II attrition rates remain stubbornly high because sponsors have historically taken a scattergun approach to clinical development – beginning trials without fully understanding the patient population, the most appropriate investigators, or prior study data. This approach leads to overly complex trial designs and investigator site saturation.

A large portion of Phase II clinical trials, particularly in oncology and rare diseases, are single arm trials. This has led to various trial design problems and challenges in trial results interpretation. This has not only caused higher Phase II attrition rates, but also inevitably contributed to higher failure rates in Phase III trials.

These inefficiencies mean many programs progress to later phases already at high risk of severe enrolment difficulties and trial failure. Sponsors are under increasing pressure to deliver cost-effective therapies faster but risk missing their commercial objectives if they continue to neglect real-world data and AI-driven clinical data science to support the design, execution and interpretation of trials.

How do you envision data science and AI helping turn the tide on these clinical trial inefficiencies, particularly in cancer research?

GL: Oncology R&D is increasingly driven by precision and understanding the molecular mechanisms of cancer. However, clinical trials, from planning to implementation, have yet to reflect the same level of precision. By harnessing clinical data analytics and AI, sponsors can perform detailed scenario and prediction modelling before a wet trial ever gets underway.

AI-powered digital twins will be particularly valuable for biomarker-specific oncology trials, where patient populations are often limited. Digital twins minimize or even eliminate the need for external control arms, accelerating recruitment and reducing patient burden. Sponsors that proactively use AI and clinical data science will overcome many of the longstanding issues faced in oncology clinical development – and, most importantly, get life-saving treatments to patients faster.

It is hard to exaggerate the power of AI driven clinical data science. It allows you to meet your patients you are inspired to treat even before you have designed your clinical trial. We can effortlessly simulate the entire clinical development and or clinical trial in the planning stage to prevent costly protocol amendments in flight and mitigate potential risk in operational implementation, reducing costs and shortening cycle time.

Finally, what excites you most about the future of work at Phesi and its potential impact on cancer research?

GL: The growing demand for AI-driven solutions in clinical development is incredibly exciting. At Phesi, we’re constantly evolving Trial Accelerator to incorporate the growing source of real-world data available and provide deeper insights for sponsors. This platform demonstrates the power of clinical trial simulation, offering deep visibility at every stage, from the patient profile to trial outcomes and FDA approvals. Trial Accelerator reimagines a future where every development candidate benefits from data-driven foresight and precision, helping to accelerate cycle times and transform the landscape of cancer research.

The application of artificial intelligence in clinical trials is transforming clinical development. You are either a part of it, or will be left out.

Explore further advancements in cancer research in our current editorial feature.