Smartphone-based fluorescence imaging combined with a convolutional neural network-assisted intelligent multichannel immunosensor for simultaneous identification of four Salmonella serotypes based on multicolor luminescent iridium complexes.

BACKGROUND: Salmonella Typhi, Salmonella Paratyphi A, Salmonella Paratyphi B, and Salmonella Paratyphi C are known to cause enteric fever and often present similar clinical symptoms, which may lead to misdiagnosis and pose a significant threat to public health. Therefore, there is an urgent need for a portable, cost-effective, and time-saving method for the simultaneous identification and detection of these four Salmonella serotypes.

RESULTS: In this work, an intelligent multichannel fluorescent immunosensor was developed for the simultaneous identification and detection of four Salmonella serotypes by using multicolor iridium-doped silica nanoparticles, combined with smartphone-based fluorescence imaging and convolutional neural network (CNN)-assisted analysis. In addition to conventional fluorescence spectral analysis, a portable fluorescence imaging system was successfully constructed using a 96-well plate, a smartphone, and a custom-built black box. The captured fluorescence images were further processed by a ResNet-based CNN model, enabling rapid classification of the four Salmonella serotypes. The proposed immunosensor achieved detection limits below 100 CFU/mL for all four targets and showed lower relative standard deviations in real sample analysis.

SIGNIFICANCE AND NOVELTY: These results demonstrate that the proposed portable intelligent immunosensor provides a promising strategy for multiplexed pathogen detection.