Background: The prevalence and severity of pediatricMycoplasma pneumoniae pneumonia(MPP) poses a significant threat to the health and lives of children. In this study, we aim to systematically evaluate the value of routine blood parameters in predicting MPP and develop a robust and generalizable ensemble artificial intelligence (AI) model to assist in identifying patients with MPP.Methods: We collected 27 features, including routine blood parameters and hs-CRP levels, from patients admitted to The Affiliated Dazu’s Hospital of Chongqing Medical University with or without MPP between January, 2023 and January, 2024. A classification model was built using seven machine learning (ML) algorithms to develop an integrated prediction tool for diagnosing MPP. It was evaluated on both an internal validation set (982 individuals) and an external validation set (195 individuals). The primary outcome measured the accuracy of the model in predicting MPP.Results: The GBDT is state-of-the-art based on 27 features. Following inter-laboratory cohort testing, the GBDT demonstrated an AUC, accuracy, specificity, sensitivity, PPV, NPV, and F1-score of 0.980 (0.938–0.995), 0.928 (0.796–0.970), 0.929 (0.717-1.000), 0.926 (0.889–0.956), 0.922 (0.727-1.000), 0.937 (0.884–0.963), and 0.923 (0.800-0.966) in stratified 10-fold cross-validation. A GBDT-based AI Lab was developed to facilitate the healthcare providers in remote and impoverished areas.Conclusions: The GBDT-based AI Lab tool, with high sensitivity and specificity, could help discriminate between pediatric MPP infection and non-MPP infection based on routine blood parameters. Moreover, a user-friendly webpage tool for AI Lab could facilitate healthcare providers in remote and impoverished areas where advanced technologies are not accessible.

Background: Influenza viruses are major pathogens responsible for acute respiratory infections in humans, which present with symptoms such as fever, cough, sore throat, muscle pain, and fatigue. While molecular diagnostics remain the gold standard, their limited accessibility in resource-poor settings underscores the need for rapid, cost-effective alternatives. Routine blood parameters offer promising predictive value but lack integration into intelligent diagnostic systems for influenza subtyping. Objective: This study aimed to develop a machine learning model using 24 routine blood parameters to predict influenza A and B infections and validate a deployable diagnostic tool for low-resource clinical settings. Methods: In this multicenter retrospective study, 6628 adult patients (internal cohort: n=2951; external validation: n=3677) diagnosed with influenza A virus infection (A+ group), influenza B virus infection (B+ group), or those presenting with influenza-like symptoms but testing negative for both viruses (A–/B– group) were enrolled from 3 hospitals between January 2023 and May 2024. The CatBoost (CATB) algorithm was optimized via 5-fold cross-validation and random grid search using 24 routine blood parameters. Model performance was evaluated using metrics such as the area under the curve (AUC), accuracy, specificity, sensitivity, positive predictive value, negative predictive value, and F1-score across internal testing and external validation cohorts, with Shapley additive explanations analysis identifying key biomarkers. The Artificial Intelligence Prediction of Influenza A and B (AI-Lab) tool was subsequently developed on the basis of the best-performing model. Results: In the internal testing cohort, 7 models (K-nearest neighbors, naïve Bayes, decision tree, random forest, extreme gradient boosting, gradient-boosting decision tree, and CatBoost) were evaluated. The AUC values for diagnosing influenza A ranged from 0.788 to 0.923, and those for influenza B from 0.672 to 0.863. The CATB-based AI-Lab model achieved superior performance in influenza A detection (AUC 0.923, 95% CI 0.897-0.947) and influenza B (AUC 0.863, 95% CI 0.814-0.911), significantly outperforming conventional models (K-nearest neighbors, RF, and XGBoost; all P<.001). During external validation, AI-Lab demonstrated high performance, achieving an accuracy of 0.913 for differentiating the A+ group from the A–/B– group and 0.939 for distinguishing the B+ group from the A–/B– group. Conclusions: The CATB-based AI-Lab tool demonstrated high diagnostic accuracy for influenza A and B subtyping using routine laboratory data, achieving performance comparable to rapid antigen testing. By enabling timely subtype differentiation without specialized equipment, this system addresses critical gaps in managing influenza outbreaks, particularly in resource-constrained regions.

Streptococcus pneumoniae is a significant pathogen causing infectious diseases, including pneumonia, otitis media, septicemia, and meningitis. The introduction of multivalent vaccines has coincided with a remarkable decrease in the number of pneumococcal-related deaths. Despite this, pneumococcal infection remains a significant cause of death among children under 5 years old and adults aged 65 or older at a global level. Therefore, early detection of S. pneumoniae infection is crucial for prognosis of pneumococcal infection patients. In this study, we evaluated the utility of a real-time multienzyme isothermal rapid amplification (MIRA) assay for detecting S. pneumoniae and other non-S. pneumoniae bacterial species. A primer–probe set targeting the S. pneumoniae lytA gene was designed, followed by optimization of parameters for the MIRA assay. At the same time, we validated the real-time MIRA assay for detecting S. pneumoniae using 79 clinical isolates identified by VITEK MS. The results showed a detection sensitivity and specificity of 100%. These results demonstrate that the designed real-time MIRA assay is a promising, rapid, simple, and reliable method for detecting S. pneumoniae infection in resource-limited areas. It has great potential for application in detecting not only S. pneumoniae but also other non-S. pneumoniae bacterial species.

Objectives: Thalassemia trait (TT) screening in resource-limited settings is hampered by reliance on expensive and complex tests. This study aimed to develop and validate a highly accessible machine learning-based tool using only routine blood parameters to accurately differentiate TT from non-TT and its major subtypes. Methods: The retrospective study included 987 individuals (221 α-TT, 211 β-TT and 555 non-TT) from two medical centers. Seven machine learning methods—Logistic Regression, Gaussian Naive Bayes, Decision Tree, Random Forest, Multilayer Perceptron, XGBoost, and CatBoost—were employed to develop diagnostic models, which were evaluated using accuracy, sensitivity, specificity, AUC, PPV, NPV, and F1 score. Results: The CatBoost model emerged as superior for differentiating TT from non-TT, achieving an AUC of 0.976, accuracy of 0.940, and specificity of 0.981. It also outperformed other models in distinguishing α-TT from β-TT (AUC = 0.842). Critically, this high-performance model was successfully deployed as a user-friendly WeChat mini-program AI Lab, for real-world clinical application. Conclusion: The deployed ML-based AI Lab represents a robust, interpretable, and scalable tool poised to enhance TT screening efficiency and accessibility, particularly in underserved healthcare environments.

Antimicrobial resistance (AMR) poses a critical global health threat. Conventional antimicrobial susceptibility testing (AST) requires 12–24 h, delaying targeted treatment. Although rapid AST approaches exist, their reliance on static measurements and single-colony analysis limits diagnostic accuracy by overlooking intrasample heterogeneity, which can produce misleading results when resistant subpopulations are present. Here, we present a novel analytical framework that integrates dynamic MALDI-TOF MS spectral analysis, the robust dynamic relative growth (RBD-RG) algorithm, and machine learning. This framework decodes the mass spectral evolution of Escherichia coli exposed to levofloxacin, enabling the rapid determination of levofloxacin susceptibility. AI MedLab MS was developed and validated using 60 clinical E. coli isolates, achieving robust classification performance across resistance phenotypes. Prospective validation with 50 independent isolates confirmed clinical utility: AI MedLab MS delivered AST results within 2 h, with an overall accuracy of 84.0% and a precision of 95.8% for resistant isolates. By shifting the analytical paradigm from single-timepoint detection to dynamic trajectory modeling, this study demonstrates a precise and interpretable methodology for rapid AST. The platform addresses key limitations of conventional methods by analyzing the overall bacterial population rather than individual colonies, thereby improving classification accuracy for complex resistance phenotypes, and has the potential to provide actionable clinical decision-making within a 2-h window. IMPORTANCE: Rapid determination of antimicrobial susceptibility is critical for the management of severe bacterial infections, yet conventional culture-based assays require up to 24 h. While emerging rapid diagnostic tools offer shorter turnaround times, their accuracy is frequently limited by a reliance on static phenotypic endpoint measurements. Here, we report AI MedLab MS, a machine-learning-enabled diagnostic platform that captures the real-time, dynamic response profiles of Escherichia coli exposed to levofloxacin. By longitudinally tracking these continuous phenotypic transitions, our method differentiates drug-resistant strains within 2 h. This rapid profiling capability may provide actionable diagnostic insights to guide targeted therapy, with the potential to optimize clinical decision-making and support antimicrobial stewardship.