Charles Fisher, Ph.D., is the CEO and Founding father of Unlearn, a platform harnessing AI to tackle a few of the biggest bottlenecks in clinical development: long trial timelines, high costs, and unsure outcomes. Their novel AI models analyze vast quantities of patient-level data to forecast patients’ health outcomes. By integrating digital twins into clinical trials, Unlearn is in a position to speed up clinical research and help bring life-saving latest treatments to patients in need.
Charles is a scientist with interests on the intersection of physics, machine learning, and computational biology. Previously, Charles worked as a machine learning engineer at Leap Motion and a computational biologist at Pfizer. He was a Philippe Meyer Fellow in theoretical physics at École Normale Supérieure in Paris, France, and a postdoctoral scientist in biophysics at Boston University. Charles holds a Ph.D. in biophysics from Harvard University and a B.S. in biophysics from the University of Michigan.
You’re currently within the minority in your fundamental belief that mathematics and computation must be the inspiration of biology. How did you originally reach these conclusions?
That’s probably simply because mathematics and computational methods haven’t been emphasized enough in biology education lately, but from where I sit, individuals are starting to alter their minds and agree with me. Deep neural networks have given us a latest set of tools for complex systems, and automation helps create the large-scale biological datasets required. I feel it’s inevitable that biology transitions to being more of a computational science in the subsequent decade.
How did this belief then transition to launching Unlearn?
Up to now, a lot of computational methods in biology have been seen as solving toy problems or problems far faraway from applications in medicine, which has made it difficult to exhibit real value. Our goal is to invent latest methods in AI to resolve problems in medicine, but we’re also focused on finding areas, like in clinical trials, where we will exhibit real value.
Are you able to explain Unlearn’s mission to eliminate trial and error in medicine through AI?
It’s common in engineering to design and test a tool using a pc model before constructing the true thing. We’d prefer to enable something similar in medicine. Can we simulate the effect a treatment could have on a patient before we give it to them? Although I feel the sector is pretty removed from that today, our goal is to invent the technology to make it possible.
How does Unlearn’s use of digital twins in clinical trials speed up the research process and improve outcomes?
Unlearn invents AI models called digital twin generators (DTGs) that generate digital twins of clinical trial participants. Each participant’s digital twin forecasts what their final result can be in the event that they received the placebo in a clinical trial. If our DTGs were perfectly accurate, then, in principle, clinical trials might be run without placebo groups. But in practice, all models make mistakes, so we aim to design randomized trials that use smaller placebo groups than traditional trials. This makes it easier to enroll within the study, speeding up trial timelines.
Could you elaborate precisely on what’s Unlearn’s regulatory-qualified Prognostic Covariate Adjustment (PROCOVA™) methodology?
PROCOVA™ is the primary method we developed that permits participants’ digital twins to be utilized in clinical trials in order that the trial results are robust to mistakes the model may make in its forecasts. Essentially, PROCOVA uses the undeniable fact that a few of the participants in a study are randomly assigned to the placebo group to correct the digital twins’ forecasts using a statistical method called covariate adjustment. This permits us to design studies that use smaller control groups than normal or which have higher statistical power while ensuring that those studies still provide rigorous assessments of treatment efficacy. We’re also continuing R&D to expand this line of solutions and supply much more powerful studies going forward.
How does Unlearn balance innovation with regulatory compliance in the event of its AI solutions?
Solutions geared toward clinical trials are generally regulated based on their context of use, which suggests we will develop multiple solutions with different risk profiles which are geared toward different use cases. For instance, we developed PROCOVA since it is amazingly low risk, which allowed us to pursue a qualification opinion from the European Medicines Agency (EMA) to be used as the first evaluation in phase 2 and three clinical trials with continuous outcomes. But PROCOVA doesn’t leverage all of the data provided by the digital twins we create for the trial participants—it leaves some performance on the table to align with regulatory guidance. In fact, Unlearn exists to push the boundaries so we will launch more revolutionary solutions geared toward applications in earlier stage studies or post-hoc analyses where we will use other kinds of methods (e.g., Bayesian analyses) that provide far more efficiency than we will with PROCOVA.
What have been a few of the most important challenges and breakthroughs for Unlearn in utilizing AI in medicine?
The most important challenge for us and anyone else involved in applying AI to problems in medicine is cultural. Currently, the overwhelming majority of researchers in medicine specifically should not extremely conversant in AI, and so they are often misinformed about how the underlying technologies actually work. Consequently, most individuals are highly skeptical that AI will likely be useful within the near term. I feel that can inevitably change in the approaching years, but biology and medicine generally lag behind most other fields with regards to the adoption of latest computer technologies. We’ve had many technological breakthroughs, but an important things for gaining adoption are probably proof points from regulators or customers.
What’s your overarching vision for using mathematics and computation in biology?
In my view, we will only call something “a science” if its goal is to make accurate, quantitative predictions in regards to the results of future experiments. At once, roughly 90% of the drugs that enter human clinical trials fail, actually because they don’t actually work. So, we’re really removed from making accurate, quantitative predictions without delay with regards to most areas of biology and medicine. I don’t think that changes until the core of those disciplines change–until mathematics and computational methods turn out to be the core reasoning tools of biology. My hope is that the work we’re doing at Unlearn highlights the worth of taking an “AI-first” approach to solving a vital practical problem in medical research, and future researchers can take that culture and apply it to a broader set of problems.
