The introduction of digital twins is creating opportunities to rethink rare disease studies. Here, Gen Li, CEO and Founder of Phesi, shares his thoughts on how digital twins are offering new hope for patients with rare conditions.
What are the biggest unmet needs in rare disease clinical development, and how can digital twins help address them?
Many patients still face long diagnostic timelines and limited treatment options. Clinical development relies on strong evidence, but rare diseases often lack the volume and quality of data needed for traditional randomized controlled trials. In addition, rare diseases show high variability in symptoms, progression, and treatment response, making it difficult to define consistent endpoints and study criteria.
According to Orphanet, 72 percent of rare diseases have genetic origins and extremely low prevalence rates (fewer than one per 100,000 people). This makes patient recruitment slow and challenging. Limited trial activity also means there is often little historical data to guide study design, increasing the risk of protocol issues such as unclear endpoints, dosing strategies, or control group selection.
Advances in clinical data science are beginning to address these challenges. Digital patient profiles (DPPs) and digital twins can now be developed for rare diseases using large datasets and artificial intelligence. These approaches can help overcome recruitment limitations and reduce reliance on traditional comparator arms, which may be impractical or unethical in rare disease settings.
There is also increasing focus on mechanism-based evidence, alongside new regulatory pathways – such as the FDA’s approach to personalized genetic therapies – which may not require traditional randomized trial designs. This reflects growing acceptance of alternative evidence sources, including digital twins and external control arms, particularly in studies with small patient populations.
Regulatory agencies, including the FDA, have long allowed the use of retrospective patient data in rare diseases. Digital twin approaches now make it possible to use these data more efficiently to support clinical development.
How can digital twins and patient profiles help reduce diagnostic delays and uncertainty in rare and ultra-rare diseases?
Each rare disease has its own diagnostic, treatment, and care landscape. Clinical data science can help define and measure these differences. DPPs, which underpin digital twins, provide a detailed view of the target patient population, including factors such as age, sex, ethnicity, comorbidities, medications, and the clinicians involved in care.
Context is essential. A DPP captures where and when data were generated, who collected them, and how. This helps ensure accuracy and relevance for clinical development.
By providing a clearer understanding of patient populations and care pathways, these insights can help reduce diagnostic delays and support more data-driven decisions in study design and execution, including adapting to changes in standards of care.
When we last spoke, about modernizing clinical trials, you emphasized data-driven trial design. How can those principles be applied in rare diseases, where patient numbers are small and diagnostic definitions may still be evolving?
Rare disease trials are less common, which means sponsors often have limited historical data to guide study design. Bringing together and contextualizing patient data – such as care settings, research environments, and prior clinical trials – can help build a more complete understanding of rare disease populations, even when data sources are limited. This combined approach supports more informed decision-making in protocol design.
Small patient populations also make predictive analysis more important. For example, traditional methods for selecting investigator sites are often less effective in rare diseases, where patients are geographically dispersed. This can lead to selecting “generalist” sites with limited experience in the specific condition, resulting in slower recruitment, higher costs, and potential trial delays or failure.
Using predictive approaches, such as those based on DPPs, can improve site selection by identifying investigators with relevant experience and a track record of enrolling patients with the specific rare disease. This can help streamline recruitment and reduce delays for both patients and investigators.
Which types of diagnostic data are most critical for building clinically meaningful digital twins?
The importance of diagnostic data – such as histopathology, genomic sequencing, proteomics, and routine laboratory tests – depends on the specific disease being studied.
Our Trial Accelerator uses real-world data from more than 90,000 continuously updated sources, including observational studies, electronic health records, retrospective and cohort studies, disease registries, pharmacy records, claims data, and randomized clinical trials. These data are combined to build digital twins that reflect real patient populations.
Using a data-driven approach, patient information is analyzed to identify consistent patterns within a disease, including differences between subtypes. This helps create a more accurate and detailed representation of the condition being studied.
What safeguards are needed to ensure that digital twin–derived diagnostic insights are transparent, reproducible, and trusted by pathologists and regulators?
The quality of the underlying data is critical. Like other scientific fields, clinical data science follows established principles to ensure objectivity, transparency, and reproducibility. Sponsors and regulators need clear information on where and when data were generated, who collected them, and how, so results can be verified and trusted.
Ongoing engagement with regulatory agencies is also important. As approaches continue to evolve, regulators are becoming more open to new technologies and are often willing to work with sponsors to support the development of regulatory-grade digital twin models.
What practical steps should pathology laboratories take now to prepare for the integration of digital twins into rare disease diagnostics and clinical development?
Just embrace it. This is no longer a futuristic concept, it is here now. The choice is to adopt and engage with it now, or risk falling behind.
Newsletters
Receive the latest pathologist news, personalities, education, and career development – weekly to your inbox.
