摘要：美国西达 - 赛奈医疗中心（Cedars-Sinai）获美国卫生高级研究计划局（ARPA-H）高达 505.4235 万美元拨款，开发 AI 驱动的药物安全平台 KronosRx。该平台核心是将 AI 工具与源自人类干细胞的类器官、器官芯片 “患者虚拟化身” 结合，通过海量匿名电子健康记录训练 AI 模型，可在临床试验前预测药物毒性与不良反应，动态模拟不同年龄、健康状况及合并用药患者的用药风险演变，旨在解决 30% 临床试验因药物不良反应失败的行业痛点，减少对动物实验的依赖，缩短新药研发周期。
AI-Powered Pharma: “Digital Doctors” Predict Drug Efficacy, “Smart Hunters” Unlock Healing Keys
A Cedars-Sinai medical center has secured over $5 million in government funding to develop an AI-driven drug safety platform called KronosRx—with its core being a group of “patient avatars” grown from human stem cells! These mini stand-ins are sophisticated organoids and organ-on-chip systems that can mimic the functions of human organs like the heart and liver with just a few cells, accurately reflecting the effects of experimental drugs.
Trained on millions of anonymous patients’ electronic health records, the AI model can not only determine drug safety but also dynamically simulate changes in medication risks based on factors like age, health status, and even other drugs a patient may be taking. Notably, 30% of clinical trials historically fail due to adverse drug reactions, as animal tests cannot accurately predict long-term drug performance in humans. With this “digital pre-screening system,” new drugs will undergo a “stress test” before human trials in the future, reducing failure and waste while getting life-saving treatments to patients faster.
美国西达 – 赛奈医疗中心（Cedars-Sinai）拿到了一笔超 500 万美元的政府拨款，要打造一款名叫 KronosRx 的 AI 药物安全平台 —— 核心竟是一群由人类干细胞培育的 “患者虚拟化身”！这些迷你替身是复杂的类器官和器官芯片，用少量细胞就能模拟心脏、肝脏等人体器官的功能，还能精准反应实验药物的作用效果。
AI 模型通过分析数百万匿名患者的电子健康记录 “学成毕业” 后，不仅能判断药物安全与否，还能根据患者年龄、健康状况甚至正在服用的其他药物，动态模拟用药风险的变化过程。要知道，过去 30% 的临床试验都栽在药物不良反应上，动物实验根本没法精准预测药物在人体的长期表现。有了这个 “数字预检系统”，未来新药在进入人体试验前就能完成一轮 “压力测试”，既减少失败浪费，还能让救命药更快抵达患者手中～