Researchers developed and validated an AI model that simultaneously detects multiple genetic markers in colorectal cancer tissue slides.
Key Details
- 1Study led by EKFZ for Digital Health at TU Dresden analyzed nearly 2,000 digitized pathology slides across seven independent cohorts in Europe and the US.
- 2The AI 'multi-target transformer' model predicts a wide range of genetic alterations, including microsatellite instability and BRAF/RNF43 mutations, from routine histological sections.
- 3Model performance matched or exceeded single-target models for key biomarkers and revealed the ability to identify shared morphological patterns tied to multiple mutations.
- 4Findings published in The Lancet Digital Health, with plans to extend methods to other cancers.
- 5Collaboration involved multiple academic centers, highlighting interdisciplinary and international partnership.
Why It Matters
This work represents a substantial advance for digital pathology and AI-driven precision diagnostics, potentially enabling faster, more comprehensive, and scalable assessment of genetic biomarkers directly from standard tissue images. Faster and lower-cost pre-screening for key mutations could streamline patient selection for molecular testing and personalize colorectal cancer care.
Source
EurekAlert
Related News
•EurekAlert
AI-Powered OCT Enables Rapid 'Optical Biopsy' for Early Endometrial Cancer Detection
A team at Washington University has developed a catheter-based 3D OCT system with AI to quickly and noninvasively detect early endometrial cancers.
•EurekAlert
AI Clinical Reasoning in Diagnostics and Digital Fatigue in Healthcare
Recent JMIR features explore large language models in clinical diagnostics and digital fatigue among healthcare professionals.
•EurekAlert
KAIST, MIT, Microsoft Develop Efficient AI Image Upsampling for Robotics
KAIST, MIT, and Microsoft have created 'Upsample Anything,' a training-free AI method to restore high-resolution visual data from compressed images with up to 16x improved GPU memory efficiency.