Pseudomonas aeruginosa (PA) is a significant pathogen of clinical concern that is frequently associated with multidrug resistance, leading to respiratory tract, wound, and hospital-acquired infections. To enable rapid and accurate detection, we developed a fluorescence-based loop-mediated isothermal amplification (LAMP) method, targeting the PA-specific ecfX gene. Among ten primer sets designed, the optimal set (EC2) was identified, and reaction conditions were optimized (Bst polymerase 320 U/mL, Mg²⁺ 8 mM, dNTP 1.4 mM, inner/outer primer ratio 1:8, 64 °C, 20 min). The assay demonstrated a detection limit that was comparable to a real-time polymerase chain reaction and immunochromatographic assays, but with a markedly reduced turnaround time. No cross-reactivity was observed with non-PA pathogens, and reproducibility tests confirmed high stability. In addition, the reliability of the results was further verified using 60 standard bacterial strains, and the feasibility of the assay was validated with 2 real soil samples and 1 water sample. This LAMP method offers a simple, rapid, and sensitive tool for on-site detection of PA, with potential applications in clinical diagnostics and public health surveillance.

Rapid and accurate pathogen detection serves as a core component in infectious disease prevention and control, clinical diagnosis and treatment, and public health surveillance systems. Although traditional detection methods have been widely adopted in clinical practice, they still exhibit significant limitations in terms of detection speed, throughput, automation levels, and adaptability to complex samples. In recent years, artificial intelligence (AI) technology has provided novel technical pathways for pathogen detection by leveraging its strengths in feature learning, pattern recognition, and multidimensional data modeling. The core contribution of this review lies in providing a novel, integrated analytical framework that overcomes the limitations of existing reviews, which often focus on a single modality (such as imaging alone or molecular diagnostics alone). Based on this framework, this paper systematically reviews AI research progress in pathogen detection, focusing on typical applications of machine learning and deep learning algorithms in analyzing imaging data, molecular diagnostic data, sensor signals, microscopic images, and multimodal data. It summarizes AI’s enabling value in enhancing detection sensitivity, specificity, automation, and point-of-care capabilities. Concurrently, this paper delves into key challenges facing AI-assisted pathogen detection, including data standardization, model generalization, interpretability, and clinical translation. It also outlines future trends toward intelligent, integrated, and clinically deployable applications. This paper aims to provide researchers and clinicians in the interdisciplinary field of artificial intelligence, biosensing, and clinical medicine with a comprehensive reference and roadmap for future development.

Helicobacter pylori (H. pylori) infection is closely associated with the development of chronic gastritis, peptic ulcers, and gastric cancer, highlighting the importance of rapid and accurate detection for disease prevention and clinical management. In this study, a one-pot LAMP-CRISPR/Cas12b assay targeting the CagA gene was developed for H. pylori detection. First, the LAMP system was optimized by systematically screening key reaction components. Subsequently, a one-step LAMP-CRISPR/Cas12b detection platform was established through optimization of the ratio between the LAMP premix and CRISPR buffer, reaction temperature, Cas12b concentration, and ssDNA reporter concentration. Under optimal conditions, the assay achieved a detection limit of 3.14 × 10¹ copies/µL, representing a tenfold improvement in sensitivity compared with conventional LAMP and PCR assays (3.14 × 10² copies/µL). In addition, the entire detection process could be completed within 1 h. Validation using 17 culture-positive and 17 culture-negative samples demonstrated complete concordance with culture-based results, with no false-positive or false-negative detections observed. These findings indicate that the proposed platform possesses high sensitivity, excellent specificity, rapid turnaround, and operational simplicity, demonstrating strong potential for point-of-care testing and applications in resource-limited settings.