Yao Ping Lu, Carolina del Real Mata, Mahsa Jalali, Marjan Khatami, Laura Montermini, Jackson McCormack-Ilersich, Janusz W. Rak, Livia Garzia, Sara Mahshid
Communication écrite (2025)
Document publié alors que les auteurs ou autrices n'étaient pas affiliés à Polytechnique Montréal
Résumé
Extracellular vesicles (EVs) are a promising biomarker for cancer hallmarks. EV molecular fingerprints can be generated at the single vesicle level with our nanostructured platform harnessing Surface-enhanced Raman Spectroscopy (SERS) readout. However, effective spectral interpretation is often hindered by the heterogeneous and nanoscale nature of EVs. To mitigate this, we integrated dimensionality reduction techniques into our SERS library of single EV spectra and subsequently evaluated Deep Learning models by binary classification accuracy of two different types of cancer and comparative healthy controls. Resampling implementation significantly improved the binary accuracy of the cancer models by 10.7% and 8.2%, respectively.
Carolina del Real Mata
Mahsa Jalali
Janusz W. Rak
Livia Garzia
Sara Mahshid
Statistiques
Dimensions
