Related Experiment Video
Updated: May 25, 2026
Low-Cost, Volume-Controlled Dipstick Urinalysis for Home-Testing
Published on: May 8, 2021
Smartphone-Based Microbubble-Linked Immunosorbent Assay Powered by Classification-Regression Integrated Deep Learning
Xin Tang1, Lian Xue1, Xiang Liu1
1The Key Laboratory for Biomedical Photonics of MOE at Wuhan National Laboratory for Optoelectronics - Hubei Bioinformatics & Molecular Imaging Key Laboratory, Systems Biology Theme, Department of Biomedical Engineering, College of Life Science and Technology, Huazhong University of Science and Technology, Wuhan 430074, China.
A new smartphone-based platform, CRiBDL-ELISA, enables precise and cost-effective biomarker detection. This portable diagnostic tool utilizes deep learning for rapid analysis, improving healthcare accessibility.
Area of Science:
- Biomedical Engineering
- Point-of-Care Diagnostics
- Artificial Intelligence in Healthcare
Background:
- Traditional ELISA requires expensive equipment, limiting accessibility.
- Growing demand for portable and efficient diagnostic technologies.
- Need for cost-effective and robust biomarker detection methods.
Purpose of the Study:
- To develop a smartphone-based platform for precise biomarker detection using deep learning.
- To overcome limitations of traditional ELISA systems.
- To provide a cost-effective, portable, and high-throughput diagnostic solution.
Main Methods:
- Development of CRiBDL-ELISA (Classification-Regression Integrated Deep Learning Microbubble-Based Enzyme-Linked Immunosorbent Assay) platform.
- Utilized platinum nanoprobes for bubble pattern generation upon biomarker capture.
- Integrated smartphone imaging, YOLO object detection, and deep learning for analysis.
- Quantification based on concentration-dependent bubble patterns.
Main Results:
- Achieved low detection limits (0.001 ng/mL to 0.0001 ng/mL) for biomarkers like CRP, PCT, NT-proBNP, and cTnI.
- Demonstrated high accuracy with R² values up to 0.9983.
- Clinical validation showed high accuracy for CRP (100.00%) and NT-proBNP (97.22%).
- NT-proBNP measurements showed strong concordance with commercial platforms (R² = 0.97).
Conclusions:
- CRiBDL-ELISA offers a cost-effective, robust, and portable solution for clinical diagnostics.
- The platform enables high-throughput biomarker quantification without specialized instrumentation.
- Smartphone-based deep learning analysis enhances precision and accessibility of diagnostic tests.
