Application of Machine Learning for the Prediction of Etiological Types of Classic Fever of Unknown Origin

Artificial intelligence applications in the diagnosis and treatment of bacterial infections

The diagnosis and treatment of bacterial infections in the medical and public health field in the 21st century remain significantly challenging. Artificial Intelligence (AI) has emerged as a powerful new tool in diagnosing and treating bacterial infections. AI is rapidly revolutionizing epidemiological studies of infectious diseases, providing effective early warning, prevention, and control of outbreaks. Machine learning models provide a highly flexible way to simulate and predict the complex mechanisms of pathogen-host interactions, which is crucial for a comprehensive understanding of the nature of diseases. Machine learning-based pathogen identification technology and antimicrobial drug susceptibility testing break through the limitations of traditional methods, significantly shorten the time from sample collection to the determination of result, and greatly improve the speed and accuracy of laboratory testing. In addition, AI technology application in treating bacterial infections, particularly in the research and development of drugs and vaccines, and the application of innovative therapies such as bacteriophage, provides new strategies for improving therapy and curbing bacterial resistance. Although AI has a broad application prospect in diagnosing and treating bacterial infections, significant challenges remain in data quality and quantity, model interpretability, clinical integration, and patient privacy protection. To overcome these challenges and, realize widespread application in clinical practice, interdisciplinary cooperation, technology innovation, and policy support are essential components of the joint efforts required. In summary, with continuous advancements and in-depth application of AI technology, AI will enable doctors to more effectivelyaddress the challenge of bacterial infection, promoting the development of medical practice toward precision, efficiency, and personalization; optimizing the best nursing and treatment plans for patients; and providing strong support for public health safety.

A score-based method of immune status evaluation for healthy individuals with complete blood cell counts

BACKGROUND

With the COVID-19 outbreak, an increasing number of individuals are concerned about their health, particularly their immune status. However, as of now, there is no available algorithm that effectively assesses the immune status of normal, healthy individuals. In response to this, a new score-based method is proposed that utilizes complete blood cell counts (CBC) to provide early warning of disease risks, such as COVID-19.

METHODS

First, data on immune-related CBC measurements from 16,715 healthy individuals were collected. Then, a three-platform model was developed to normalize the data, and a Gaussian mixture model was optimized with expectation maximization (EM-GMM) to cluster the immune status of healthy individuals. Based on the results, Random Forest (RF), Light Gradient Boosting Machine (LightGBM) and Extreme Gradient Boosting (XGBoost) were used to determine the correlation of each CBC index with the immune status. Consequently, a weighted sum model was constructed to calculate a continuous immunity score, enabling the evaluation of immune status.

RESULTS

The results demonstrated a significant negative correlation between the immunity score and the age of healthy individuals, thereby validating the effectiveness of the proposed method. In addition, a nonlinear polynomial regression model was developed to depict this trend. By comparing an individual's immune status with the reference value corresponding to their age, their immune status can be evaluated.

CONCLUSION

In summary, this study has established a novel model for evaluating the immune status of healthy individuals, providing a good approach for early detection of abnormal immune status in healthy individuals. It is helpful in early warning of the risk of infectious diseases and of significant importance.

Diagnostic models for fever of unknown origin based on 18F-FDG PET/CT: a prospective study in China

Background

This study aims to analyze the 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) characteristics of different causes of fever of unknown origin (FUO) and identify independent predictors to develop a suitable diagnostic model for distinguishing between these causes. A total of 524 patients with classical FUO who underwent standard diagnostic procedures and PET/CT were prospectively studied. The diagnostic performance of PET/CT imaging was analyzed, and relevant clinical parameters that could improve diagnostic efficacy were identified. The model was established using the data of 369 patients and the other 155 patients comprised the validation cohort for verifying the diagnostic performance of the model.

Results

The metabolic characteristics of the “hottest” lesion, the spleen, bone marrow, and lymph nodes varied for various causes. PET/CT combined with clinical parameters achieved better discrimination in the differential diagnosis of FUO. The etiological diagnostic models included the following factors: multisite metabolic characteristics, blood cell counts, inflammatory indicators (erythrocyte sedimentation rate, C-reactive protein, serum ferritin, and lactate dehydrogenase), immunological indicators (interferon gamma release assay, antinuclear antibody, and anti-neutrophil cytoplasm antibody), specific signs (weight loss, rash, and splenomegaly), and age. In the testing cohort, the AUCs of the infection prediction model, the malignancy diagnostic model, and the noninfectious inflammatory disease prediction model were 0.89 (95% CI 0.86–0.92), 0.94 (95% CI 0.92–0.97), and 0.95 (95% CI 0.93–0.97), respectively. The corresponding AUCs for the validation cohort were 0.88 (95% CI 0.82–0.93), 0.93 (95% CI 0.89–0.98), and 0.95 (95% CI 0.92–0.99), respectively.

Conclusions

18F-FDG PET/CT has a certain level of sensitivity and accuracy in diagnosing FUO, which can be further improved by combining it with clinical parameters. Diagnostic models based on PET/CT show excellent performance and can be used as reliable tools to discriminate the cause of FUO.

Trial registration This study (a two-step method apparently improved the physicians’ level of diagnosis decision-making for adult patients with FUO) was registered on the website http://www.clinical-trials.gov on January 14, 2014, with registration number NCT02035670.