What happened
Researchers built an AI system that can simulate how a patient's tumor will change over time in response to specific treatments, rather than just predicting a single outcome. This matters because doctors could test treatment strategies computationally before prescribing them, potentially catching better options earlier and avoiding ineffective ones.
Why it matters
For the first time, an AI system claims to model disease progression as a continuous trajectory that responds to treatment choices — not just classify static snapshots. If this works reliably in real clinical settings, it shifts cancer treatment from trial-and-error to something closer to physics: predict the system's behavior under different interventions before you act.