CellVoyant, a U.K.-based startup, has launched an AI platform that allows scientists to predict the future health and performance of human cells based on their appearance in microscope images—a breakthrough that could sharply reduce the cost of cell-based medical therapies.
A growing number of treatments for diseases ranging from cancer to Parkinson’s to diabetes depend on modified human cells as the core component of the treatment. These include CAR-T, where a patient’s own immune cells are extracted, genetically reprogrammed outside the body so that they can recognize and destroy cancer cells, and then reinjected into the patient. The category also includes various stem cell-based therapies.
Currently, these cell-based therapies are among the most expensive medications available. CAR-T treatments, for example, often cost hundreds of thousands of dollars per dose.
Part of the reason they are so costly is due to the labor-intensive, highly sensitive, and relatively wasteful process that is required to produce them. Scientists have to culture a lot more cells than they need, because some of the cells will not be healthy enough or exhibit the right properties to make a good treatment dose. Although it is possible to determine some aspects of cell health simply by “eyeballing” the living cells under a microscope, scientists usually cannot gauge which cells are best without performing tests on samples, usually killing those cells being tested in the process. Then they have to hope that the cells they tested are actually representative of the other cells in the culture, which is not always the case. And while tests assess a cell’s current condition, they can’t predict how the cell will develop in the future. Sometimes entire cell lines fail to perform and have to be discarded. Sometimes an entire treatment dose winds up being ineffective. All of this adds to the cost of the treatments.
CellVoyant’s new product, a platform called FateView, aims to significantly reduce the waste in this process by using AI models it has trained to classify cells by their current qualities and, critically, predict which cells will possess the right qualities in the future, simply by analyzing microscope-based imagery of the cells that use regular, white light. Currently the platform can do this for 10 different cell types—including stem cells, T-cells, cardiac cells, and blood cells—and the company is planning on training its models to work with more in the future.
Predicting cells’ behavior
The company’s platform can, according to CellVoyant, instantly identify which cells are currently exhibiting certain biomarkers, predict how well individual cells will express certain genes in the future, and forecast how well stem cells will differentiate into specific cell types—all from white light microscope images, without having to perform chemical tests that are time-consuming and can destroy cells in the process.
“We can see, understand and predict how cells behave without having to destroy them,” Rafael Carazo Salas, CellVoyant’s founder and CEO, said. He said FateView could predict a cell’s quality hours, days, or even weeks out from its present state. That should allow scientists and the companies producing cell lines for therapy to be much more selective in deciding which cells should progress, eliminating waste, improving the chances of success, and ultimately, lowering costs.
Carazo Salas gives the example of scientists who produce specific cell-types from stem cells—which has the potential to revolutionize the treatment of everything from Type 1 Diabetes to heart disease. This complex process can take weeks and the yields of usable cells tend to be low, he said. But he said CellVoyant had been able to reduce the costs of what’s called cell derivation—making those specific cell types from stem cells—by up to 80% simply by better predicting at each stage of the process which cells are most likely to progress well. Because some of these cell therapies currently have price tags approaching $1 million, that cost savings is game-changing, meaning that many more people (or their insurance companies) will be able to pay for these treatments, he said.
And it’s not just cell-based therapies that depend on healthy cells. The challenge of predicting cell health is also relevant for many “biologic” drugs, which are usually proteins produced by bacteria or other kinds of cells that are then harvested, and even in the case of cells needed to test the effects of the small molecules that make up the majority of pharmaceuticals. “Whether using cells to discover drugs, as a measuring device, or using cells as a factory to produce biologics, or cells as a drug, as in case of cell therapy, the unit economics is defined by cells,” Carazo Salas told Fortune. “The cost is defined by cells, batch to batch. Variation is what accounts for a lot of the cost in the industry.”
CellVoyant is making its FateView system available through a simple online interface that lets scientists upload microscope images of cells to be analyzed, as well as through an API (application programming interface) for companies that need to analyze high volumes of samples, possibly as part of a robotic laboratory workflow. Academic users will be able to access the platform for a nominal fee, while biotech and pharmaceutical companies are charged an annual subscription, which gives them the right to store their data securely, as well as a relatively low per-use charge.
Carazo Salas, who is also a professor of cellular and molecular medicine at the University of Bristol, in England, said CellVoyant was able to train AI models to characterize cells and predict their behavior because it had access to a large database of microscope images of the same cells taken over time, as well as the results of traditional chemical assays on cells taken at different stages of development. This time-series data allows the models to learn how the shape and visual characteristics of a cell at any point relate to its current function, as well as how it relates to its future appearance and function. The company trains a specific model for each cell type it works on—for instance, a separate model for cardiac cells and one for metabolic cells—although it is possible that in the future a single foundation model might be able to learn how to make predictions about any cell type, Carazo Salas said.
CellVoyant, whose name is a portmanteau derived from the words “cell” and “clairvoyant,” was spun out of the University of Bristol in 2021. In 2023, it received £7.6 million ($10.1 million) in seed funding from Octopus Ventures, Horizon Ventures, Verve Ventures, and Air Street Capital.
FateView marks CellVoyant’s first major commercial product release. Previously, the company has worked with specific biotech and pharma partners, only some of which it can name, Carazo Salas said.
One of its early customers is Rinri Therapeutics, a biotech company in Sheffield, England, that is working on a cell-based regenerative therapy for hearing loss. Terri Gaskell, the chief technology officer at Rinri Therapeutics, said in a statement that CellVoyant’s platform had enabled it to predict “cell behavior in ways that haven’t been possible before.” Gaskell said that with help from CellVoyant, the company “hope[s] it will be possible to scale production [of cells] more efficiently and make it significantly more cost effective, [and] ultimately bring restorative cell therapies closer to those with hearing loss that need them most.”
