Fluorescence Analysis of Circulating Exosomes for Breast Cancer Diagnosis Using a Sensor Array and Deep Learning

Emerging liquid biopsy methods for investigatingbiomarkers in bodilyfluids such as blood, saliva, or urine can beused to perform noninvasive cancer detection. However, thecomplexity and heterogeneity of exosomes require improvedmethods to achieve the desired sensitivity and accuracy. Herein,we report our study on developing a breast cancer liquid biopsysystem, including afluorescence sensor array and deep learning(DL) tool AggMapNet. In particular, we used a 12-unit sensorarray composed of conjugated polyelectrolytes,fluorophore-labeledpeptides, and monosaccharides or glycans to collectfluorescencesignals from cells and exosomes. Linear discriminant analysis(LDA) processed thefluorescence spectral data of cells and cell-derived exosomes, demonstrating successful discrimination be-tween normal and different cancerous cells and 100% accurate classification of different BC cells. For heterogeneous plasma-derivedexosome analysis, CNN-based DL tool AggMapNet was applied to transform the unorderedfluorescence spectra into feature maps(Fmaps), which gave a straightforward visual demonstration of the difference between healthy donors and BC patients with 100%prediction accuracy. Our work indicates that ourfluorescent sensor array and DL model can be used as a promising noninvasivemethod for BC diagnosis.

Automated summary

Emerging liquid biopsy methods, such as the breast cancer liquid biopsy system developed in this study, provide a promising noninvasive approach for detecting and classifying cancer cells based on fluorescence signals collected from cells and exosomes. The integration of a fluorescence sensor array and deep learning tool AggMapNet allows for accurate differentiation between normal and cancerous cells, as well as accurate prediction of breast cancer in plasma-derived exosomes.