Researchers at Google DeepMind and Yale University have unveiled a groundbreaking artificial intelligence system that could expedite cancer treatment. The C2S-Scale 27B model is capable of identifying intricate biological patterns within cells and predicting how various drugs will affect tumors.
This AI has helped uncover a mechanism that allows "cold" tumors to become visible to the immune system. Typically, these tumors evade recognition by immunity, complicating treatment options, especially immunotherapy. The new discovery may pave the way for combination treatment methods that were previously deemed impossible.
C2S-Scale 27B analyzed over 4000 drugs on patient tumor samples and in laboratory settings. The model was able to pinpoint substances that selectively enhance immune response rather than affect all cells indiscriminately.
Among the most intriguing findings is the CK2 kinase inhibitor silmitasertib (CX-4945). The AI predicted that the drug is effective only in the presence of a small amount of interferon in the cells. Either the drug or interferon alone had minimal effects, but their combination boosted immune system activity by 50%, effectively "warming up" the tumors and making them visible to the body's defense cells.
The AI's predictions were validated in the lab on neuroendocrine human cells that the model had not previously encountered. This demonstrates that the system can not only process data but also draw conclusions based on the cellular context.
According to scientists, large AI models could serve as "virtual laboratories" conducting thousands of simulations and revealing unknown connections between drugs, cells, and the immune system.
"This discovery could represent a new avenue for developing cancer treatment methods," said Google's CEO Sundar Pichai.
The success of C2S-Scale 27B illustrates that AI can accelerate scientific progress and assist in creating effective treatment methods far more rapidly than traditional approaches.
Recently, researchers from New York University Abu Dhabi (NYUAD) also introduced a new type of brain implant capable of delivering drugs to multiple areas of the brain with high precision.