DeepMind, the AI research lab owned by Google’s parent company Alphabet, says its machine‑learning tool has identified vorinostat – a drug already approved for a type of lymphoma – as a candidate for treating liver fibrosis. The discovery marks a step in the search for anti‑fibrotic therapies, a field where few effective options exist.
How the AI made the find
The tool analyzed vast datasets of molecular interactions and disease pathways, narrowing down compounds that might reverse or slow the scarring of liver tissue. Vorinostat, a histone deacetylase inhibitor used in cutaneous T‑cell lymphoma, surfaced as a strong match. DeepMind researchers said the model could reduce the time needed to repurpose existing drugs for new conditions.
What liver fibrosis is
Liver fibrosis is the buildup of scar tissue caused by chronic injury – from hepatitis, alcohol abuse, or fatty liver disease. Over time it can lead to cirrhosis and liver failure. There are no FDA‑approved drugs specifically for reversing fibrosis; current treatment focuses on removing the underlying cause. That makes a potential therapy like vorinostat a significant prospect, though years of clinical testing lie ahead.
The drug’s track record
Vorinostat has been on the market since 2006 for cancer. Its safety profile in cancer patients is known, but that doesn’t guarantee it will work or be safe for liver fibrosis patients. The AI’s prediction will need to be validated in animal models and then human trials. If successful, the drug could be repurposed much faster than a brand‑new compound, because some safety data already exists.
DeepMind has not yet announced a timeline for the next steps. The company typically publishes its findings in peer‑reviewed journals, but a preprint or study release has not been confirmed. Researchers in the field will be watching for the data behind the AI’s selection.
The discovery adds to a growing list of drug candidates that AI tools have proposed, though few have reached patients. Whether vorinostat becomes an exception will depend on the results of experiments that, for now, remain inside the lab.
