人工智能发现新的癌症治疗组合 可能帮助免疫系统瞄准肿瘤 实验室测试显示 AI finds new cancer treatment combo that may help immune system target tumors, lab tests show.

Google DeepMind和耶鲁大学开发了27亿参数的AI模型Cell2Sentence-session-Service 27B,其中确定了一种潜在的新的癌症治疗方法:将Silmitasertib与低剂量干涉结合,使“冷”肿瘤在免疫系统中更加明显。 Google DeepMind and Yale University have developed a 27-billion-parameter AI model, Cell2Sentence-Scale 27B, which identified a potential new cancer therapy: combining silmitasertib with low-dose interferon to make "cold" tumors more visible to the immune system. AI预测,在使用人体细胞进行的实验室试验中,只有在免疫活性条件下,抗原的展示才会产生协同增强作用。 The AI predicted a synergistic boost in antigen presentation—only under immune-active conditions—validated in lab tests using human cells. 这一发现过去没有记录,表明AI有能力产生新的、可测试的生物假设。 The discovery, not previously documented, demonstrates AI’s ability to generate novel, testable biological hypotheses. 模型和工具可以公开提供,但专家强调,这是早期研究,需要在临床使用前进行多年进一步研究。 The model and tools are publicly available, though experts stress this is early-stage research requiring years of further study before clinical use.