Stanford Medicine and Kyoto University researchers developed a machine‑learning model that predicts whether donation‑after‑circulatory‑death (DCD) donors will die within the time window suitable for liver procurement, outperforming surgeon judgment and reducing futile procurements by 60% in multicenter validation. Published in The Lancet Digital Health, the model uses pre‑withdrawal clinical variables to forecast time‑to‑death and was shown to cut futile surgical mobilizations where death occurred too late for safe organ recovery. Lead investigators highlighted efficiency gains and potential to increase transplant utilization. Model deployment could change procurement decision‑making, reduce wasted operating room resources, and expand organ availability for candidates on waiting lists, but prospective implementation studies are needed to assess clinical workflow integration and ethical considerations.