Related Experiment Video
Updated: Jun 19, 2026
13:01
Industrialized, Artificial Intelligence-guided Laser Microdissection for Microscaled Proteomic Analysis of the Tumor Microenvironment
Published on: June 3, 2022
Automated HER2 Scoring with Uncertainty Quantification Using Lensfree Holography and Deep Learning.
Che-Yung Shen1,2,3, Xilin Yang1,2,3, Yuzhu Li1,2,3
1Electrical and Computer Engineering Department, University of California, Los Angeles, Los Angeles, CA 90095, USA.
BME Frontiers
|June 18, 2026
Summary
We developed an automated system for scoring human epidermal growth factor receptor 2 (HER2) in breast cancer using lensfree holography and deep learning. This cost-effective method offers rapid and reliable HER2 assessment, especially in resource-limited settings.
Area of Science:
- Biomedical Optics
- Digital Pathology
- Machine Learning in Medicine
Background:
- Accurate human epidermal growth factor receptor 2 (HER2) scoring is crucial for breast cancer management.
- Current digital HER2 scoring methods often require expensive and bulky optical systems.
- There is a need for cost-effective and accessible HER2 assessment tools.
Purpose of the Study:
- To develop an automated HER2 scoring system utilizing lensfree holography.
- To integrate deep learning for rapid and accurate HER2 classification.
- To create a cost-effective alternative to traditional digital pathology systems.
Main Methods:
- Lensfree holography device capturing diffraction patterns of stained HER2 tissue sections.
- Deep learning model incorporating Bayesian Monte Carlo dropout for uncertainty quantification.
- High-throughput data acquisition with a sample area of ~1,250 mm² and throughput of ~84 mm²/minute.
Main Results:
- Achieved 84.9% accuracy for 4-class HER2 classification (0, 1+, 2+, 3+).
- Attained 94.8% accuracy for binary HER2 scoring (0/1+ vs. 2+/3+) with uncertainty quantification.
- Demonstrated robust and reliable scoring through autonomous uncertainty estimates.
Conclusions:
- Lensfree holography with deep learning offers a practical solution for automated HER2 scoring.
- The system is highly suitable for resource-limited settings lacking traditional digital pathology infrastructure.
- This approach facilitates high-throughput and cost-effective breast cancer diagnostics.
