Dual-substrate SERS strategy combined with 3D AgNPs substrate for minimally invasive screening and staging of liver cancer.
Liver cancer (LC) is one of the most prevalent and deadly malignancies worldwide, and early diagnosis together with accurate staging is critical for improving patient prognosis. Although surface-enhanced Raman spectroscopy (SERS) has shown strong potential for minimally invasive tumor detection, its broader clinical application remains limited by insufficient substrate stability, incomplete spectral coverage, and inadequate discriminative ability in complex classification and staging tasks.
In this study, we developed a silver nanoparticle-composited three-dimensional porous SERS substrate (Ag@BTM) by combining the breath-figure method with a Marangoni effect-driven liquid-liquid interfacial self-assembly strategy, enabling uniform AgNP distribution within the porous scaffold and improving SERS sensitivity, uniformity, and reproducibility. To further expand biochemical information, we proposed a dual-substrate SERS detection strategy using two Ag@BTM substrates with complementary localized surface plasmon resonance characteristics, namely S-Ag@BTM and L-Ag@BTM, and fused their spectra into a dual-substrate SERS super-fingerprint. Combined with the AutoGluon automated machine learning framework and SHAP analysis, this strategy achieved 98.6% accuracy for classifying Normal, HBV, and LC samples and 80.4% accuracy for LC staging, representing improvements of 12.6 and 14.7 percentage points over the corresponding single-substrate models.
This dual-substrate fusion framework provides a sensitive, low-cost, and scalable route for minimally invasive LC screening and precise staging, and highlights the potential of substrate-complementary SERS analysis for complex clinical classification tasks.
In this study, we developed a silver nanoparticle-composited three-dimensional porous SERS substrate (Ag@BTM) by combining the breath-figure method with a Marangoni effect-driven liquid-liquid interfacial self-assembly strategy, enabling uniform AgNP distribution within the porous scaffold and improving SERS sensitivity, uniformity, and reproducibility. To further expand biochemical information, we proposed a dual-substrate SERS detection strategy using two Ag@BTM substrates with complementary localized surface plasmon resonance characteristics, namely S-Ag@BTM and L-Ag@BTM, and fused their spectra into a dual-substrate SERS super-fingerprint. Combined with the AutoGluon automated machine learning framework and SHAP analysis, this strategy achieved 98.6% accuracy for classifying Normal, HBV, and LC samples and 80.4% accuracy for LC staging, representing improvements of 12.6 and 14.7 percentage points over the corresponding single-substrate models.
This dual-substrate fusion framework provides a sensitive, low-cost, and scalable route for minimally invasive LC screening and precise staging, and highlights the potential of substrate-complementary SERS analysis for complex clinical classification tasks.
Authors
Li Li, Wang Wang, Xu Xu, Wu Wu, Gong Gong, Qiao Qiao, Lin Lin, Zhang Zhang, Lin Lin
View on Pubmed