AI Identifies Hidden Drug Treatments for Liver Disease and Cancer

Opening

Researchers using artificial intelligence systems have identified previously overlooked drug candidates and compounds to treat killer diseases including liver fibrosis and acute myeloid leukaemia, according to two new studies published in the journal Nature. The breakthrough emerged from AI systems designed to analyse vast volumes of medical literature and identify connections humans may have missed. Firms are racing to deploy AI in medical research, with scientists emphasizing the technology's ability to dramatically accelerate discoveries that traditionally require years of work.

Google DeepMind AI Co-Scientist System

One study focused on an AI co-scientist built by Google DeepMind, designed to generate its own scientific theories and identify connections buried within large volumes of scientific research.

Professor Gary Peltz from Stanford University School of Medicine used the system while researching existing drugs that might help tackle liver fibrosis, a condition that can be deadly. Peltz selected two drugs based on decades of expertise; the AI suggested three additional possibilities. Testing showed Peltz's choices provided no benefit, but two of the AI's suggestions worked. One treatment blocked approximately 90% of fibrosis damage.

Peltz stated: "It was really quite striking. I kind of fell off my chair." He added: "It's like talking to an oracle that's read everything, knows all the literature, but also has some reasoning capabilities and finds connections that we may miss."

Kira6 Compound and Leukaemia Research

The AI system also flagged a compound called Kira6, which scientists believe could help combat acute myeloid leukaemia, an aggressive blood cancer. Early laboratory tests found the compound suppressed cancer cells.

Expert Assessment

Professor Filippo Menolascina of Edinburgh University stated that AI would not replace scientists but could enhance research capabilities. He said: "It's like strapping on a jet pack."

Clinical Status

Scientists stressed that any treatments identified through this research remain a long way from being clinically proven.