Mayo Clinic AI Model Detects Pancreatic Cancer Up to Three Years Before Diagnosis

Mayo Clinic researchers have developed an AI model called REDMOD that can detect pancreatic cancer on routine CT scans up to three years before conventional diagnosis, identifying 73% of prediagnostic cancers at a median of 16 months early—nearly double the detection rate of specialists without AI assistance. The model analyzes hundreds of quantitative imaging features to identify subtle tissue changes before tumors become visible, and was validated across nearly 2,000 CT scans from multiple institutions. This breakthrough addresses a critical need, as more than 85% of pancreatic cancer patients are diagnosed after the disease has spread, contributing to five-year survival rates below 15%. The technology is designed to analyze existing CT scans obtained for other reasons, particularly in high-risk patients with new-onset diabetes, and is advancing to clinical testing through the AI-PACED prospective study.

AI Model's Early Detection Capabilities and Performance

• ▪The Mayo Clinic Radiomics-based Early Detection Model (REDMOD) measures hundreds of quantitative imaging features that describe tissue texture and structure.
• ▪The Mayo Clinic Radiomics-based Early Detection Model (REDMOD) can detect pancreatic cancer on routine abdominal CT scans up to three years before clinical diagnosis.
• ▪The Mayo Clinic Radiomics-based Early Detection Model (REDMOD) identified 73% of prediagnostic pancreatic cancers at a median of about 16 months before diagnosis.
• ▪The Mayo Clinic Radiomics-based Early Detection Model (REDMOD) runs automatically without time-intensive manual preparation.
• ▪The Mayo Clinic Radiomics-based Early Detection Model (REDMOD) identifies subtle signs of pancreatic cancer before tumors are visible.
• ▪The Mayo Clinic Radiomics-based Early Detection Model (REDMOD) detection rate was nearly double the detection rate of specialists reviewing the same scans without AI assistance.

Clinical Validation and Real-World Application

• ▪Mayo Clinic researchers used the Radiomics-based Early Detection Model (REDMOD) to analyze nearly 2,000 CT scans, including scans from patients later diagnosed with pancreatic cancer that were originally interpreted as normal.
• ▪In patients with multiple scans, the Mayo Clinic Radiomics-based Early Detection Model (REDMOD) produced consistent results months apart.
• ▪The Mayo Clinic Radiomics-based Early Detection Model (REDMOD) was validated across CT scans from multiple institutions, imaging systems and protocols.
• ▪The Mayo Clinic Radiomics-based Early Detection Model (REDMOD) study findings were published in the journal Gut.
• ▪Dr. Ajit Goenka is a Mayo Clinic radiologist and nuclear medicine specialist who served as senior author of the pancreatic cancer AI study.

Pancreatic Cancer's Deadly Nature and Detection Challenges

• ▪More than 85% of pancreatic cancer patients receive a diagnosis after the disease has already spread.
• ▪Pancreatic cancer five-year survival rates remain below 15% according to the National Cancer Institute.
• ▪Pancreatic cancer is projected to become the second-leading cause of cancer-related death in the United States by 2030.

Perspective of Mayo Clinic researchers

• ▪The Mayo Clinic Radiomics-based Early Detection Model (REDMOD) could enable opportunistic screening by analyzing CT scans already performed for other medical reasons.
• ▪The Artificial Intelligence for Pancreatic Cancer Early Detection (AI-PACED) prospective study will test Mayo Clinic Radiomics-based Early Detection Model (REDMOD) in real clinical settings.
• ▪High-risk patients such as those with new-onset diabetes represent a priority population for Mayo Clinic Radiomics-based Early Detection Model (REDMOD) screening.

Perspective of Pancreatic cancer patients and advocacy groups

• ▪Earlier detection of pancreatic cancer through AI models like REDMOD could significantly improve the current below-15% five-year survival rate.
• ▪The ability to detect pancreatic cancer before it spreads addresses the critical problem that over 85% of patients are diagnosed after metastasis.

3 sources