Label-free gold nanostar-based SERS with machine learning: A platform for detecting endometrial cancer-associated polyamine metabolites

Abstract

The occurrence and progression of endometrial cancer are closely associated with metabolic reprogramming, in which polyamine metabolites play a critical role in tumor cell proliferation and invasion. In this study, we developed a label free surface enhanced Raman scattering (SERS) detection platform based on gold nanostars (AuNS), integrated with machine learning algorithms, to achieve highly sensitive detection and precise identification of polyamine metabolites related to endometrial cancer. In complex biological matrices such as serum, the platform yielded stable and reproducible spectral fingerprints, with a detection limit at the nanogram level. Furthermore, by constructing a polyamine metabolite spectral database and introducing machine learning models, both the classification accuracy and AUC values exceeded 95 %, enabling effective discrimination of different metabolic states and mixed systems. Taken together, the AuNS SERS strategy combined with machine learning provides a rapid, non-invasive, and intelligent detection tool for the early diagnosis and metabolic subtyping of endometrial cancer, with significant clinical application potential.