Au@AgPt Nanozymes with Synergistic Plasmonic-Catalytic Sites for Multimodal Large Model-Empowered Dual-Mode Diagnostics of Helicobacter pylori in Saliva

Authors:
  • JMJiani Man
    Jiani Man
  • GWGuojun Weng
    Guojun Weng
  • JLJian-Jun Li
    Jian-Jun Li
  • JZJw Zhao
    Jw Zhao
  • YGYubo Guo
    Yubo Guo
  • HZHaoyu Zhang
    Haoyu Zhang

Abstract
Noninvasive detection of Helicobacter pylori (H. pylori) currently faces a fundamental trade-off between laboratory-grade sensitivity and point-of-care (POC) accessibility. Here, a trimetallic nanozyme Au@AgPt Patrick star has been architected to leverage both ultrasensitive surface-enhanced Raman scattering (SERS) and instrument-free colorimetry for noninvasive H. pylori detection in saliva. A lamellar Au nanostar core is cloaked by an Ag–Pt alloy shell, with the cavity size and Ag/Pt ratio tuned to balance plasmonic enhancement and peroxidase-like catalysis. Density functional theory (DFT) reveals that the synergistic electron transfer from Ag to Pt induces a downshift in the Pt d-band center, thermodynamically optimizing intermediate adsorption. The optimized nanozyme delivers a SERS enhancement factor of 1.95 × 105 and a peroxidase-like Michaelis constant of 91.8 μM. SERS achieves a limit of detection (LOD) of 7 CFU mL-1. For instrument-free POC settings, the nanozyme effectively catalyzes TMB-H2O2 oxidation, generating a robust colorimetric readout with an LOD of 25 CFU mL-1. A PCA-CNN classifier trained on a curated SERS spectral library of six oral microbes achieves clinically aligned H. pylori discrimination with a positive recall of 88.7%. Furthermore, a smartphone-integrated multi-vision large language model consensus architecture autonomously mitigates device-specific imaging biases for colorimetric result interpretation. By synergizing atom-level material engineering with multimodal artificial intelligence, this platform establishes a highly reliable paradigm for the rapid, noninvasive screening of H. pylori.