Microfluidic Biochip-Based Multiplexed Profiling of Small Extracellular Vesicles Proteins Integrated with Machine...

Area of Science:

• Biomarker Discovery
• Proteomics
• Machine Learning in Diagnostics

Background:

• Early diagnosis is critical for improving cancer survival rates.
• Lack of reliable biomarkers hinders early detection of many diseases, including esophageal squamous cell carcinoma (ESCC).
• Current diagnostic methods for ESCC have limited accuracy.

Purpose of the Study:

• To develop and validate a novel platform for early diagnosis of ESCC.
• To integrate extracellular vesicle (EV) proteomic analysis with machine learning for biomarker discovery.
• To establish a precise diagnostic model for ESCC using multi-protein biomarkers.

Main Methods:

• Development of a barcode immunoassay biochip for capturing and lysing small extracellular vesicles (EVs) from serum.
• Quantification of multiple EV proteins, including membrane and internal proteins.
• Application of machine learning algorithms to analyze proteomic data from 273 clinical samples.
• Identification of nine diagnostic protein biomarkers through mass spectrometry.

Main Results:

• A precise diagnostic model for ESCC was developed using machine learning.
• The model achieved 91.0% accuracy in external validation for ESCC detection.
• Early-stage ESCC detection accuracy was 90.8%, significantly outperforming current biomarkers (14.4%).

Conclusions:

• Integrating extracellular vesicle protein analysis with machine learning enables accurate identification of ESCC patients.
• The developed EV analysis platform is a promising tool for clinical application of multi-biomarker detection.
• This approach significantly advances the early diagnosis of esophageal squamous cell carcinoma.