Serum Exosome SERS Assay Based on TiN‐Ag@Ag Sol Composite Substrate and Its Application in the Diagnosis of Gastric Cancer

• Published in: Journal of Raman spectroscopy Vol. 56; no. 4; pp. 307 - 315
• Main Authors: Wang, Huan, Wu, Zhengang, Wei, Yingna, Chen, Ying, An, Xiao jie, Li, Jingwu, Wang, Zhiwu, Liu, Yankun, Wei, Hengyong
• Format: Journal Article
• Language: English
• Published: Bognor Regis Wiley Subscription Services, Inc 01.04.2025
• Subjects: Algorithms; Cancer; Cancer screening; Diagnosis; Discriminant analysis; Exosomes; Gastric cancer; K-nearest neighbors algorithm; Learning algorithms; Machine learning; Malignancy; Medical screening; Patients; Principal components analysis; Raman spectroscopy; Spectrum analysis; Support vector machines; surface‐enhanced Raman scattering
• ISSN: 0377-0486; 1097-4555
• DOI: 10.1002/jrs.6767

ABSTRACT Gastric cancer (GC) is a highly lethal malignancy, seriously threatening people's physical health. Accurate screening of gastric cancer could improve the survival rate of patients. Therefore, exploring noninvasive and efficient cancer screening methods for gastric cancer is of great significance. In the past few years, exosomes have received much attention for their potential in disease diagnosis and treatment. Here, the aim of this study was to explore the detection of serum exosomes via surface‐enhanced Raman spectroscopy (SERS) technique based on TiN‐Ag@Ag sol composite substrate, and its potential application in gastric cancer diagnosis is evaluated. Exosomes were extracted from the serum of 31 GC patients and 31 healthy controls (HC) using an exosome kit. This study used various machine learning algorithms such as principal component analysis linear discriminant analysis (PCA‐LDA), partial least squares discriminant analysis (PLS‐DA), support vector machine (SVM), and k‐nearest neighbor (KNN) algorithm to analyze SERS spectra, in order to distinguish between HC and GC. The results show that the k‐nearest neighbor algorithm performs the best in HC and GC classification. These results indicate that the combination of SERS and machine learning methods provides a new technological approach for gastric cancer screening. This study offers a new proposal for the universal applicability of analysis and identification with SERS of serum exosomes samples in clinical diagnosis. A rapid and label‐free cancer detection method using SERS and machine learning was developed for differentiating serum exosomes.