Acid-resistant PVP/gold nanoparticles cluster for gastric juice-based metabolomics analysis

This study aims to develop a surface-enhanced Raman scattering (SERS)-based technology capable of directly analyzing metabolites in gastric juice using acid-resistant nanostructure for the diagnosis of gastric cancer. Although gastric cancer has a high incidence and mortality rate, existing endoscopic examination methods have limitations such as low sensitivity and operator dependency. In this study, we fabricated nanoparticles (NPs) cluster structure using Au NPs encapsulated in polyvinylpyrrolidone (PVP) to maintain structural stability even in strongly acidic environments. The PVP-Au NPs solution mixed with gastric juice was coated using the meniscus guided deposition technique to fabricate a uniform NPs cluster, thereby ensuring high chemical stability and reproducibility of SERS signals. The acquired SERS spectra were preprocessed and transformed into barcode-format features, which were then used to train a neural-network-based machine learning classification model. In the analysis of 121 gastric juice samples, the model accurately classified gastric cancer with a sensitivity of 91.7 % and a specificity of 91.7 % in the test set. This study demonstrates the potential of a high-precision gastric cancer diagnostic platform that integrates an acid-resistant PVP-based Au NPs cluster structure with machine learning classification and suggests its possible expansion as an auxiliary diagnostic tool for screening high-risk groups and predicting future prognosis.