Simultaneous CEA and CDH-17 Detection from Cancer Extracellular Vesicles Via Machine Learning Enabled Bio Silent Azide-Infrared Spectroscopy

Extracellular vesicles (EVs) provide a rich source of cancer biomarkers that reflect the molecular state of their parent cells and offer strong potential for non-invasive diagnosis.Here, we report a bio-silent azide-based Fourier transform infrared (FTIR) spectroscopic strategy for the simultaneous detection of two colorectal cancer (CRC) EV surface glycoproteins, carcinoembryonic antigen (CEA) and cadherin-17 (CDH-17).EVs isolated from multiple CRC cell lines and a non-cancerous control were comprehensively characterized and analyzed by FT-IR spectroscopy.Distinct biochemical signatures in lipid, protein, carbohydrate, and nucleic acid regions enabled robust differentiation between CRC-derived and control EVs, supported by multivariate and machine-learning analyses with high classification accuracy.To achieve biomarker-specific detection, CEA and CDH-17 were selectively labeled via biotinylated antibodies followed by streptavidin-azide, generating a characteristic azide vibrational signal in the bio-silent infrared region (2000-2300 cm⁻¹) without perturbing native EV spectra.Quantitative analysis of the azide signal revealed significantly elevated CEA and CDH-17 expression in CRC EVs, consistent with flow cytometry validation.Density-of-states overlap, cosine correlation, and non-negative least squares regression enabled deconvolution and simultaneous quantification of dual biomarkers from mixed samples, revealing cell line-specific expression patterns and enabling EV subtyping.This work establishes azide-FTIR spectroscopy as a rapid, label-efficient, and scalable platform for multiplexed EV biomarker detection, highlighting its promise for future liquid biopsy-based cancer diagnostics.