REDMOD Detects Pancreatic Cancer 475 Days Early, Beating Radiologists 73% to 39%

3 articles · Updated · ScienceDaily · Oct 26

REDMOD flagged hidden pancreatic cancer signatures an average of 475 days before clinical diagnosis in CT scans that radiologists had originally read as normal.
In 219 later-diagnosed patients, the AI reached 73% sensitivity versus 39% for experienced radiologists; on scans taken more than two years before diagnosis, it detected 68% versus 23%.
The model analyzes subtle pancreatic tissue patterns invisible on conventional imaging and automates pancreas outlining, reducing the need for manual tracing.
In external checks, REDMOD correctly classified just over 81% of 539 cancer-free scans and 87.5% of 80 NIH-PCT scans, with 90% to 92% consistency on repeat scans.
Researchers said prospective testing is still needed—especially in high-risk patients and more diverse populations—but earlier detection could shift more cases to treatable stage 0 disease.

Sources

ScienceDaily4h ago

AI System REDMOD Detects Pancreatic Cancer Signs Years Before Diagnosis

Indiatimes4h ago

AI Tool Detects Hidden Signs of Pancreatic Cancer 475 Days Before Diagnosis

superpowerdaily.com5h ago

Mayo Clinic AI spotted pancreatic cancer signals over a year before diagnosis in study | Superpower Daily

Could combining AI imaging with new blood tests finally make universal pancreatic cancer screening possible?

Could an invisible pattern in your routine scan predict a deadly cancer years before symptoms appear?

REDMOD AI Doubles Early Pancreatic Cancer Detection Sensitivity: Mayo Clinic’s 73% Breakthrough and Clinical Integration Roadmap

Overview

Mayo Clinic’s REDMOD is an AI model that analyzes 3D CT scan data using advanced wavelet filters to detect early pancreatic cancer. By uncovering microscopic tissue changes invisible to human eyes, REDMOD nearly doubles early detection sensitivity compared to radiologists, offering a median lead time of over a year before diagnosis. Integrated into hospital systems, REDMOD uses EHR algorithms to target high-risk patients and generates continuous risk scores, allowing clinicians to adjust thresholds for different clinical needs. This approach gives doctors a crucial window for early intervention, potentially improving survival rates while minimizing unnecessary testing and costs through targeted risk stratification.

...