The performance of VCS(volume, conductivity, light scatter) parameters in distinguishing latent tuberculosis and...

Area of Science:

• Medical diagnostics
• Infectious diseases
• Computational biology

Background:

• Tuberculosis (TB) remains a significant global health threat, causing millions of deaths annually.
• Distinguishing between active TB and latent TB infection (LTBI) is crucial for effective patient management and treatment.
• Current diagnostic methods face challenges in differentiating these two states.

Purpose of the Study:

• To develop and evaluate machine learning models for differentiating active TB from LTBI.
• To assess the utility of leukocyte Volume, Conductivity, and Light Scatter (VCS) parameters in this diagnostic task.

Main Methods:

• Collected data from 220 subjects (97 active TB, 113 latent TB).
• Analyzed 46 features including blood routine indicators and leukocyte VCS parameters (neutrophils, monocytes, lymphocytes).
• Employed four machine learning algorithms: logistic regression (LR), random forest (RF), support vector machine (SVM), and k-nearest neighbor (KNN).
• Evaluated model performance using Area Under the Precision-Recall Curve (AUPRC) and Area Under the Receiver Operating Characteristic Curve (AUROC).

Main Results:

• In the training set, LR and RF achieved perfect performance (AUROC=1, AUPRC=1).
• SVM and KNN also demonstrated high performance in the training set.
• In the testing set, LR showed the best performance (AUROC=0.977, AUPRC=0.957), followed by SVM, RF, and KNN.
• All models showed robust predictive capabilities in distinguishing active TB from LTBI.

Conclusions:

• Machine learning classifiers utilizing leukocyte VCS parameters are highly valuable for differentiating active TB from latent TB infection.
• This approach offers a promising tool for improving the accuracy of TB diagnosis.
• The findings support the potential for non-invasive, data-driven methods in managing tuberculosis.