Mayo Clinic AI detects pancreatic cancer up to 3 years before diagnosis (newsnetwork.mayoclinic.org)

Mayo Clinic has announced a breakthrough AI model, the Radiomics-based Early Detection Model (REDMOD), capable of detecting pancreatic cancer up to three years before a clinical diagnosis. This landmark validation study, published in Gut, revealed that the AI can identify subtle signs of cancer in routine abdominal CT scans, even when interpreted as normal by specialists. By analyzing around 2,000 scans, the model was able to detect 73% of prediagnostic cancers approximately 16 months prior to diagnosis, nearly doubling the detection rate of human reviewers, especially notable at two years before diagnosis when it identified nearly three times as many early-stage cancers. This development is significant for the AI/ML community as it demonstrates the potential of advanced imaging technologies to transform early cancer detection, especially for a disease like pancreatic cancer, which has a dismal five-year survival rate below 15% due to late diagnosis. REDMOD utilizes hundreds of quantitative imaging features to capture early biological changes, allowing for reliable, automatic analysis across diverse clinical settings. As researchers move toward clinical testing through the AI-PACED initiative, this technology may vastly improve patient outcomes by enabling timely interventions in high-risk individuals, thus advancing predictive healthcare methods aimed at saving lives.