Researchers combined patient-derived pancreatic tumor organoids with a deep learning tracker, OrganoIDNet, to automatically monitor in real time how CD318-targeting CAR T cells destroy living tumor tissue. The approach aims to translate cell-killing dynamics into quantifiable, continuously observable readouts. Organoid models preserve aspects of patient tumor heterogeneity that are often lost in simplified in vitro systems. Coupling that biology with automated tracking can reduce manual interpretation and enable more standardized comparisons across experimental conditions. For biotech teams, the significance is practical: real-time metrics can improve preclinical iteration cycles for CAR constructs, dosing schedules, and combination strategies by providing faster feedback on efficacy dynamics. This is an example of how organoid and computational imaging pipelines are converging into mechanistic platforms for next-generation cell therapies.