Aakash Desai, Assistant Professor and Associate Director of Phase 1 and Precision Oncology Program at the UAB O’Neal Comprehensive Cancer Center, posted on LinkedIn:
“New meta-analysis explores how well AI can predict EGFR mutations from digital pathology in lung cancer patients.
Across 16 studies (4 eligible for meta-analysis), AI models reached:
• AUC: 0.756 – indicating moderate predictive accuracy
• Best results with in-house datasets, surgical resections, tumor-focused regions, and ResNet algorithms
While promising, the findings highlight key limitations:
- Current accuracy isn’t sufficient for standalone use.
- Need for algorithm refinement plus large-scale validation.
- Integration into clinical workflows will be critical for impact.
As EGFR status guides therapy in NSCLC, AI-based prediction tools could play a transformative role in improving access to timely, personalized treatment—especially in low-resource or high-volume settings.”
Title: Artificial intelligence in predicting EGFR mutations from whole slide images in lung Cancer: A systematic review and Meta-Analysis
Authors: Mai Hanh Nguyen, Minh Huu Nhat Le, Anh Tuan Bui, Nguyen Quoc Khanh Le
Read The Full Article at Lung Cancer.
More posts featuring Aakash Desai.