As the global community prepares to mark Rare Disease Day on 28 February 2026, the pharmaceutical industry is witnessing a pivotal shift in how treatments for rare, often neglected diseases are developed.
For decades, the primary hurdle in rare disease research has been a lack of sufficient patient populations to power traditional randomised controlled trials. However, Dr Gen Li, CEO, Phesi, asserts that the maturation of AI and digital twins is finally providing a viable path forward.
From data scarcity to digital evidence
Traditional trial models often fail in rare disease contexts because recruiting a statistically significant control arm is ethically and logistically challenging. Logistically, the extreme scarcity and geographic dispersal of patients make it nearly impossible to hit the ideal numbers required for a traditional study; while ethically, it is difficult to justify placing patients with life-threatening, untreated conditions into a placebo group when a potentially life-saving therapy is available.
Organisations like Phesi are now leveraging digital patient profiles to bridge this gap. These profiles use hundreds of millions of global patient records to create digital twins for even the rarest diseases. By modelling how a condition naturally progresses over time, researchers can accurately simulate a control group, removing the need for real patients to take a placebo.
“Advances in analytics now make it possible to create digital patient profiles and digital twins for any indication, including rare diseases,” explains Dr Gen Li, CEO, Phesi. “While they are not a silver bullet, digital twins offer new avenues of hope where traditional trial designs are impractical, and regulators are increasingly open to alternative forms of evidence.”
Regulatory shifts and technological integration
The momentum behind these technologies is bolstered by a changing regulatory landscape. The FDA’s exploration of plausible mechanism pathways for bespoke medicines suggests a growing acceptance of external control arms and digital evidence.
This shift enables the approval of highly personalised genetic therapies based on mechanistic evidence that a drug targets the known biological cause of a disease, rather than requiring large-scale, multi-year population data.
Another advantage is that by integrating AI-supported protocol optimisation and investigator site selection, sponsors can significantly reduce trial timelines and costs, making them accessible to smaller biotech firms.
“With the right data and AI to support curating, modelling and analysing outcomes, the industry has an opportunity to run faster, more efficient trials,” Dr Li says. “Ultimately, the goal is to reduce uncertainty and help bring new treatments to patients sooner.”
