Digital twin for healthcare

How regulators treat healthcare digital twins

Regulators assess each model for a defined use. They have not accepted digital twins as a category. The FDA's final guidance on the credibility of computational modelling in medical device submissions dates from November 2023. It covers physics-based and mechanistic models, and does not cover standalone AI models or AI-generated control arms in clinical trials. In Europe, the EMA has issued a qualification opinion on Unlearn's PROCOVA method, which applies to that statistical method in certain trials. Neither document means virtual patients can generally replace real control groups. Each company still has to agree its approach with the regulator trial by trial.

State of the art (2026)

As of mid-2026, healthcare digital twins are split between a commercially scaled cardiac-imaging tier and an earlier-stage trial-and-research tier. HeartFlow FFRct and Cleerly process large per-patient volumes under existing FDA clearances and US reimbursement, making coronary plaque and flow modelling the proof-of-scale use case. In clinical trials, AI virtual control arms are the fastest-moving application: Unlearn.AI demonstrated up to a 33% control-arm reduction with Johnson & Johnson, and the EMA has issued a favourable qualification opinion for digital twins in Phase 2-3 trials - the first such endorsement by a major regulator. Most other organ and whole-body twins remain exploratory or supplemental, with the FDA still anchored on its computational-model credibility framework rather than a settled twin-specific pathway.

Cardiac twins are the proof-of-scale application; oncology and neurology follow

The cardiac-imaging digital twin sub-category is the first commercially scaled example of per-patient digital twins. HeartFlow's FFRct analysis (non-invasive computational fluid dynamics over a patient-specific coronary digital twin) and Cleerly's plaque-characterisation approach now process more than 1M patients per year combined; HeartFlow listed on Nasdaq (HTFL) in August 2025, raising roughly $317M and giving the category its first pure-play public comparable. Dassault Systemes' Living Heart Project entered its next phase in February 2025 with AI-powered virtual twins customisable to individual patients and patient populations - among the most clinically validated heart-modelling platforms globally with 100+ research partners and a track record of FDA collaboration (the ENRICHMENT in silico device study). Siemens Healthineers partnered with Mayo Clinic in September 2025 to develop AI-enhanced cardiovascular digital twins integrating real-time records and imaging. The structural template - high-quality imaging plus disease-specific physics-and-AI models - is now beginning to replicate into oncology (tumour growth, immunotherapy response), neurology (Alzheimer's progression, stroke rehabilitation), and pulmonology (COPD, asthma), with first regulatory submissions for these organ systems expected from 2027.

Clinical-trial virtual control arms are the largest near-term commercial application

The fastest-growing commercial application of digital twins in healthcare is virtual control arms in clinical trials. Vendors such as Unlearn.AI generate AI digital-twin models from historical trial data and use them to predict how enrolled patients would have responded under placebo or standard-of-care; in published Phase 3 collaborations with AbbVie and Johnson & Johnson, Unlearn's prognostic twins cut required control-arm size by roughly 19-33% depending on endpoint, at statistically equivalent power. Real-world-evidence and synthetic-control players (Atropos Health, Aetion, Phesi, Owkin) address the adjacent external-comparator market. The structural payoff is reduced enrolment requirements - smaller actual control arms, faster timelines, fewer patients on placebo. Clinical trials are a large multi-billion-dollar annual market and even single-digit efficiency improvement compounds significantly. The post-FDA-guidance window (2026-2028) is when virtual control arms move from pilot adoption at a handful of trial sponsors toward default consideration across major pharma's late-phase trial portfolios.