Back to all news
AI Model Predicts Microsatellite Instability and Immunotherapy Response from Histology
Yonsei University researchers introduced MSI-SEER, an AI model for MSI and immunotherapy response prediction from histology images of gastric and colorectal cancers.
Key Details
- 1MSI-SEER uses deep Gaussian process modeling to analyze H&E-stained whole-slide images.
- 2The model integrates uncertainty quantification, providing a Bayesian Confidence Score for each prediction.
- 3MSI-SEER flags uncertain cases for human review to enhance reliability and safety.
- 4Validated on large, racially diverse datasets, it achieved state-of-the-art MSI prediction accuracy.
- 5The model also predicts immune checkpoint inhibitor (ICI) response, integrating tumor MSI status and stroma-to-tumor ratio.
- 6Published in npj Digital Medicine on May 19, 2025.
Why It Matters
This innovation demonstrates the power of AI in pathology to provide clinically actionable predictions and to integrate uncertainty for safer AI-human collaboration in cancer diagnosis and therapy selection. The approach may serve as a model for broader use of AI in precision oncology workflows.
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.