AI Digital Twins Could Revolutionise Drug Trials and Slash Costs

Artificial intelligence-generated digital twins are emerging as a transformative tool in drug development, potentially saving the biopharma industry up to USD 30 billion annually by 2030.

These virtual patients, built from real medical data, simulate how a person’s health might evolve without treatment, offering a new way to test drug safety and efficacy without relying solely on traditional control groups.

According to ClinicalLeader.com, AI is expected to be used in 60–70% of clinical trials by 2030, significantly reducing trial sizes and timelines.

Digital twins are already being integrated into trial designs and evaluated by regulators. Companies like Unlearn are leading the charge, using disease-specific neural networks to generate these virtual models.

Unlearn’s founder and machine learning scientist, Aaron Smith, said their digital twins provide comprehensive, individualised predictions of a patient’s future health outcomes, helping researchers understand disease progression and improve trial efficiency.

The technology supports two main applications: acting as simulated controls in single-arm trials and serving as prognostic covariates in randomised controlled trials. Both methods are gaining regulatory acceptance.

The European Medicines Agency has issued a formal qualification opinion for Unlearn’s PROCOVA method, which uses digital twin data to enhance trial power without increasing sample size.

Smith noted that even a 10% reduction in sample size for a Phase 3 trial could shorten enrolment time by four months and save tens of millions of USD. This is especially critical for patients with rare diseases or aggressive cancers.

Digital twins also offer a compassionate alternative in ethically sensitive trials, such as those involving children or terminally ill patients, where withholding treatment for control purposes is problematic.

While often misunderstood as replacements for real patients, digital twins in randomised trials instead provide additional prognostic data, improving the reliability of trial results.

Smith emphasised the importance of transparency and validation, stating that Unlearn documents every aspect of its models and validates them within each trial’s specific context.

Despite their promise, digital twins face challenges due to fragmented and inconsistent healthcare data. Bias, inaccuracies and missing values can affect model reliability.

Scaling the use of digital twins will require not just technological innovation but also a cultural shift among trial sponsors, clinical research organisations and regulators.

Smith said the industry is at an inflection point, with digital twins already delivering value and gaining trust.

---

• AI digital twins could save the biopharma industry up to USD 30 billion annually by 2030

• Digital twins simulate patient outcomes and reduce reliance on traditional control groups

• Unlearn’s technology is gaining regulatory support, including from the European Medicines Agency

Source: FORBES