MobiHealthNews (HIMSS Media)

National Taiwan University Hospital Develops AI for Pancreatic Cancer Metabolic Profiling

Taiwan’s Academia Sinica and National Taiwan University Hospital have launched PanMETAI, an AI‑integrated metabolomics platform that analyzes roughly 260,000 metabolic signals from a 500‑microliter blood‑serum sample. Using deep‑learning on nuclear magnetic resonance data, the model achieved a 99% area‑under‑curve for distinguishing non‑cancer from pancreatic cancer and 93% for early‑stage disease. Independent blind testing in Taiwan and external validation with a Lithuanian cohort confirmed its high reproducibility. The technology promises a scalable, liquid‑biopsy‑based screening tool that could shift pancreatic cancer detection to earlier, more treatable stages.