Machine Learning-Driven ZnO NRs@Ag SERS Platform for Rheumatoid Arthritis Liquid Biopsy in a Multicenter-Recruited Cohort

Abstract The analytical chemistry diagnosis of rheumatoid arthritis (RA) faces challenges such as the time lag of conventional serological assays and matrix masking in complex biofluid analysis. This study proposes a label-free liquid biopsy framework integrating band-regulated surface-enhanced Raman scattering (SERS) sensing with machine learning algorithms. Experimental characterizations (XPS and TRPL) and density functional theory (DFT) calculations confirm that the ZnO NRs@Ag heterojunction drives efficient spatial separation of charge carriers via stepped band alignment. This achieves a deep synergy between electromagnetic (EM) and chemical enhancements (CM), significantly improving the signal selectivity of metabolites against high-abundance backgrounds. Meanwhile, the substrate exhibits good batch-to-batch structural reproducibility and environmental storage stability. Based on this sensing platform, a patient-level internal validation was conducted on a multicenter-recruited serum cohort. By averaging replicate spectra to prevent data leakage, the model classification was strictly performed on 74 independent patient-level input features (comprising 66 RA cases and 8 healthy controls). Under a leave-one-patient-out cross-validation architecture coupled with dynamic in-fold oversampling, the linear classification model (PLS-DA) achieved an area under the curve (AUC) of 0.9337. Furthermore, utilizing the SHAP method, the model identified putative spectral features that influenced the classification decision, providing a reference for exploring related pathological metabolic changes. This study elucidates the synergistic enhancement mechanism of the heterojunction at the electronic state level and verifies the preliminary diagnostic potential of the platform in decoding high-dimensional metabolic profiles, offering a methodological foundation for applying SERS technology to autoimmune disease screening.

Authors

Institutions

Publication Details

Journal
ACS Sensors
Published
2026-11-14
DOI
https://doi.org/10.1021/acssensors.6c03967
Primary Topic
Gold and Silver Nanoparticles Synthesis and Applications
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Machine Learning-Driven ZnO NRs@Ag SERS Platform for Rheumatoid Arthritis Liquid Biopsy in a Multicenter-Recruited Cohort

Shibin Han, Xiaoyi Lv, Xiangwei Zhao, Lijun Wu et al.
ACS Sensors
Gold and Silver Nanoparticles Synthesis and Applications
article

Machine Learning-Driven ZnO NRs@Ag SERS Platform for Rheumatoid Arthritis Liquid Biopsy in a Multicenter-Recruited Cohort

Shibin Han, Xiaoyi Lv, Xiangwei Zhao, Lijun Wu, Cheng Chen, Mingcan Li, Xuguang Zhou, Chen Chen, Zhenhong Jia, Xue Wu
article en

Abstract

Abstract The analytical chemistry diagnosis of rheumatoid arthritis (RA) faces challenges such as the time lag of conventional serological assays and matrix masking in complex biofluid analysis. This study proposes a label-free liquid biopsy framework integrating band-regulated surface-enhanced Raman scattering (SERS) sensing with machine learning algorithms. Experimental characterizations (XPS and TRPL) and density functional theory (DFT) calculations confirm that the ZnO NRs@Ag heterojunction drives efficient spatial separation of charge carriers via stepped band alignment. This achieves a deep synergy between electromagnetic (EM) and chemical enhancements (CM), significantly improving the signal selectivity of metabolites against high-abundance backgrounds. Meanwhile, the substrate exhibits good batch-to-batch structural reproducibility and environmental storage stability. Based on this sensing platform, a patient-level internal validation was conducted on a multicenter-recruited serum cohort. By averaging replicate spectra to prevent data leakage, the model classification was strictly performed on 74 independent patient-level input features (comprising 66 RA cases and 8 healthy controls). Under a leave-one-patient-out cross-validation architecture coupled with dynamic in-fold oversampling, the linear classification model (PLS-DA) achieved an area under the curve (AUC) of 0.9337. Furthermore, utilizing the SHAP method, the model identified putative spectral features that influenced the classification decision, providing a reference for exploring related pathological metabolic changes. This study elucidates the synergistic enhancement mechanism of the heterojunction at the electronic state level and verifies the preliminary diagnostic potential of the platform in decoding high-dimensional metabolic profiles, offering a methodological foundation for applying SERS technology to autoimmune disease screening.

ACS Sensors
People's Hospital of Xinjiang Uygur Autonomous Region (CN), Southeast University (CN), Xinjiang University (CN), Tsinghua University (CN)
Gold and Silver Nanoparticles Synthesis and Applications
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon