Projected to surpass $38 billion in value by 2032, artificial intelligence (AI) simulations—particularly digital twins—are reshaping the healthcare landscape. These virtual replicas of individuals possess an uncanny ability to replicate human biological complexities, making them powerful tools for patient care and clinical research.
The Role of Digital Twins
A digital twin, often visualized as a glowing 3D organ pulsating on a dynamic dashboard, can simulate treatment outcomes with remarkable accuracy. By modeling patient outcomes and generating synthetic data, digital twins enhance real-world datasets and streamline clinical trials.
According to Rashidi Hooman, associate dean of AI in Medicine at the University of Pittsburgh, digital twins are transforming life sciences, enabling predictions of outcomes for patients in alternative treatment groups and addressing inefficiencies in research and development.
“We’re on the brink of an ‘AI tsunami,’” Hooman remarked. “Its full impact will hit in the next three to five years.”
Pharma’s Embrace of Digital Twins
Leading pharmaceutical companies, such as Bayer, are early adopters of digital twin technology. Sai Jasti, head of data science and AI at Bayer Research and Development, described digital twins as “transformational,” improving precision in clinical decision-making and streamlining trials.
In 2023, Bayer and AstraZeneca partnered with Altis Labs, a Toronto-based AI company, to enhance cancer trials using digital twin models. Altis trains its AI on historical, de-identified patient data to predict outcomes under standard-of-care treatments. At the 2024 European Society for Medical Oncology (ESMO) conference, Altis demonstrated how its AI models in non-small cell lung cancer could predict overall survival and refine treatment response assessments.
Bridging Data Gaps and Addressing Inequities
Digital twins hold promise in tackling disparities among underrepresented patient groups, such as racial minorities, pregnant women, and children, who are often excluded from clinical trials. Felix Baldauf-Lenschen, founder of Altis Labs, highlighted the potential of AI to address these gaps by generating evidence for subpopulations typically overlooked in traditional studies.
For example, people of color are more likely to be diagnosed with advanced melanoma and face higher mortality rates despite lower overall diagnosis rates. A 2022 study in JMIR Dermatology emphasized the “white lens phenomenon,” wherein AI diagnostic systems trained predominantly on light skin images disadvantage people of color. The study proposed that deep learning techniques could diversify dermatological atlases, improving diagnostic accuracy across skin tones.
Challenges in the Digital Twin Space
Despite its transformative potential, the digital twin field faces significant hurdles. Hooman noted that while the technology is highly innovative, it’s not without flaws.
“These frameworks follow statistical parameters and predictive analysis, not magic,” he said. “Errors are inherent, and in healthcare, mistakes can have life-threatening consequences.”
Key challenges include navigating regulatory requirements and ensuring high-quality, annotated data for training AI models. Creating accurate digital twins also demands enormous amounts of patient data, which can be difficult to obtain, especially for underrepresented groups.
Altis’ Baldauf-Lenschen acknowledged the difficulty of training AI on smaller subsets of data, while Bayer’s Jasti highlighted the importance of collaboration with regulatory authorities and scientific peers to overcome these barriers.
The Road Ahead
While challenges persist, the benefits of digital twins are poised to outweigh their limitations. As technology advances, these AI-powered models are expected to enhance precision medicine, streamline drug development, and address inequities in healthcare.
Hooman expressed optimism about the future: “The positives will far outweigh the negatives. The potential for digital twins to revolutionize healthcare is immense.”
