Related Experiment Video
Updated: Sep 24, 2025
09:30
Using Nanoplasmon-Enhanced Scattering and Low-Magnification Microscope Imaging to Quantify Tumor-Derived Exosomes
Published on: May 24, 2019
7.5K
Fluorescence Analysis of Circulating Exosomes for Breast Cancer Diagnosis Using a Sensor Array and Deep Learning
Yuyao Jin1,2, Nan Du1, Yuanfang Huang1,2
1The State Key Laboratory of Chemical Oncogenomics, Key Laboratory of Chemical Biology, Shenzhen International Graduate School, Tsinghua University, Shenzhen 518055, P. R. China.
ACS Sensors
|May 5, 2022
Summary
This study presents a novel liquid biopsy system for breast cancer (BC) detection. Combining a fluorescence sensor array and deep learning (DL), it accurately distinguishes between healthy and cancerous cells and plasma exosomes noninvasively.
Area of Science:
- Biomedical Engineering
- Cancer Research
- Analytical Chemistry
Background:
- Liquid biopsy offers noninvasive cancer detection via biomarkers in bodily fluids.
- Exosome analysis for cancer detection faces challenges due to complexity and heterogeneity.
- Improved sensitivity and accuracy are crucial for effective liquid biopsy systems.
Purpose of the Study:
- To develop a breast cancer (BC) liquid biopsy system.
- To enhance sensitivity and accuracy in noninvasive cancer detection.
- To utilize a fluorescence sensor array and deep learning (DL) tool for BC diagnosis.
Main Methods:
- A 12-unit fluorescence sensor array using conjugated polyelectrolytes, peptides, and glycans was employed.
- Collected fluorescence signals from cells and exosomes.
- Applied Linear Discriminant Analysis (LDA) for cell-derived exosome analysis and a Convolutional Neural Network (CNN)-based DL tool (AggMapNet) for plasma-derived exosome analysis.
Main Results:
- LDA achieved successful discrimination between normal and cancerous cells, with 100% accuracy for classifying different BC cells.
- AggMapNet transformed fluorescence spectra into feature maps, visually demonstrating differences between healthy donors and BC patients.
- The system achieved 100% prediction accuracy for plasma-derived exosome analysis.
Conclusions:
- The developed fluorescent sensor array and DL model show promise as a noninvasive method for BC diagnosis.
- This approach addresses the challenges of exosome heterogeneity in liquid biopsies.
- The system offers a sensitive and accurate tool for early cancer detection.
