AI-powered smartphone platform for robust detection of alpha-fetoprotein via nanozyme-mediated dual-readout lateral flow immunoassay

Traditional lateral flow immunoassay (LFIA) relies on subjective visual interpretation and shows limited quantitative capability. Existing smartphone-assisted LFIA image analysis methods enable image-based quantification, yet they suffer from result variability caused by inconsistent imaging distance, alignment and lighting, require manual processing that introduces subjective bias, and often rely on customized devices or controlled conditions, restricting practical deployment. Herein, a S martphone- Y ou Only Look Once (YOLO)- N anozyme E nhanced R egion of Interest (ROI) Grayscale assa y (SYNERGY)-LFIA platform was developed for sensitive, intelligent, and portable alpha-fetoprotein (AFP) quantification. Cu-TCPP metal-organic framework nanosheets, constructed from Cu 2 + and the tetrakis(4-carboxyphenyl)porphyrin (TCPP) ligand, sequentially load Au and Pt nanoparticles (NPs), yielding Cu‑TCPP/Au/Pt nanozymes with superior peroxidase‑like activity for signal labeling. The nanozyme‑labeled LFIA produces both intrinsic colorimetric and catalytic signals, enabling dual‑signal output. For intelligent quantification, an integrated workflow combining YOLO11 model and the ROI grayscale value analysis algorithm module was designed to convert LFIA images into standardized grayscale signals, enabling automated signal analysis with reduced dependence on controlled imaging conditions and manual processing. The full workflow is cloud-deployed and accessible via a smartphone application for convenient result interpretation. The platform achieved a wide linear range of 0.5–500 ng/mL and a calculated limit of detection of 0.25 ng/mL, approximately 62-fold lower than that of AuNP-labeled LFIA. It was successfully applied to serum samples, providing accurate and reliable AFP quantification and showing good consistency with electrochemiluminescence immunoassay. Overall, this work provides an artificial intelligence-powered LFIA signal analysis method with nanozyme amplification for intelligent and convenient AFP detection.