It identifies subtle signs of disease before tumors are visible, when curative treatment may still be possible. The findings, published in Gut, mark a milestone in Mayo Clinic’s multiyear research effort to enable earlier detection of one of the deadliest cancers. The study validates this next-generation AI model using data and workflows that mirror clinical practice, including CT scans from multiple institutions, imaging systems and protocols. Pancreatic cancer remains one of the deadliest cancers because it rarely causes detectable signs in its earliest stages. The Radiomics-based Early Detection Model (REDMOD) measures hundreds of quantitative imaging features that describe tissue texture and structure, capturing faint biological changes as cancer begins to develop. The model runs automatically without time-intensive manual preparation and its predictions remained stable over time. Researchers are advancing this work into clinical testing.
   —Please visit https://newsnetwork.mayoclinic.org/discussion/mayo-clinic-ai-detects-pancreatic-cancer-up-to-3-years-before-diagnosis-in-landmark-validation-study/ for more information.