A startup looking to develop digital copies of patients in clinical trials—allowing computer models of study participants to be generated ahead of time—has raised $50 million for its machine learning-powered efforts.
Unlearn works with biopharma companies to take historical patient data and create a “digital twin” before predicting the course of their disease when given a well-known control treatment or a placebo.
The goal is to help drugmakers and academic researchers complete clinical trials while requiring fewer overall participants by integrating digital twins into relatively smaller control arms.
Reducing the size of the control arm also allows more patients to access a potentially beneficial experimental treatment, said Dylan Morris, managing director at Insight Partners, the firm that led Unlearn’s venture capital round.
“Trials can be run faster and with the same resources so that patients can receive access to more effective treatments sooner,” added Morris, who joined Unlearn’s board of directors. The series B fundraising was also backed by Radical Ventures, 8VC, DCVC, DCVC Bio and Mubadala Capital Ventures.
Unlearn previously gathered $15 million through an extended series A round in late 2020, which included investments from Eisai, Epic Ventures and Alumni Ventures Group.
“Unlearn has proven how generative AI-models trained on clinical data can be successfully utilized to create digital twins of patients for complex populations such as Alzheimer’s disease,” Shun Asami, senior director of Eisai Innovation, said at the time. “We believe that the Unlearn technology and its intelligent control arms have the potential to solve the enrollment challenges, timeline delays and high failure rates that too often burden clinical trials.”
Unlearn has said that its artificial intelligence-driven models can generate digital twins that, on paper, can be “statistically indistinguishable” from actual patients in terms of following the progression of Alzheimer’s disease.
The company collects readouts from previously completed clinical trials, pools together data on patients in their control arms and then uses that information to train a machine learning model. That model is then used to analyze the baseline particulars gathered when a new participant first enrolls in a study and creates a digital twin with forecasts that will be compared to the patient’s records at the end of the trial.
So far Unlearn has worked in neurodegenerative diseases such as Alzheimer’s as well as Parkinson’s disease, Huntington disease, ALS and multiple sclerosis. It has also explored inflammatory conditions such as lupus, rheumatoid arthritis, psoriasis, ulcerative colitis and Crohn’s disease.
Most recently, the company signed a multi-year collaboration agreement in February with Merck KGaA to place digital twins into late-stage studies within the German drugmaker’s immunology pipeline.