A high-precision multimodal optoelectronic-graph biosensor for home monitoring of diabetes based on lateral flow immunoassay

With the increasing prevalence of diabetes worldwide, rapid diagnosis and precise prevention strategies have become critically important. Glycated hemoglobin (HbA 1c ), recognized as the gold standard for diabetes diagnosis, provides a reliable reflection of the patient’s average glycemic status over the preceding 2–3 months and exhibits high analytical stability. However, conventional methods of HbA 1c detection, such as liquid chromatography, are limited to centralized laboratories, requiring complex procedures and extended processing times, while existing lateral flow immunoassays (LFIAs) often suffer from low precision and poor reproducibility. In this study, we present a next-generation LFIA platform designed for home diagnosis of HbA 1c . This strategy uses immunoprobes synthesized from long afterglow nanocrystals (LANs), which effectively eliminates endogenous interference, to construct a portable multimodal optoelectronic-graph biosensor, and deploys a lightweight hybrid network model (LHNM)within it. To further enhance diagnostic performance, we introduce a novel global temporal hybrid attention module (GTHAM), which significantly improves the model’s capacity to extract discriminative temporal features from afterglow temporal sequence images. Validation using whole blood samples demonstrated that the system achieves a total detection time of 5 min and 30 s, with a coefficient of variation below 3.05 %. The platform exhibits excellent specificity, signal stability, and diagnostic accuracy comparable to standard laboratory instrumentation, fully meeting clinical diagnosis requirements.