Development and Evaluation of the New Predictive Models in Tuberculous Pleuritis

Area of Science:

• Pulmonology
• Infectious Diseases
• Biomarker Research

Background:

• Distinguishing tuberculous pleural effusion (TPE) from non-TPE is clinically important.
• Interferon gamma (IFN-γ) is a highly accurate biomarker but not always practical.
• Need for alternative diagnostic models for TPE.

Purpose of the Study:

• To develop predictive models for differentiating TPE from non-TPE.
• Models to utilize clinical data and pleural fluid biomarkers, excluding IFN-γ.
• Achieve high diagnostic accuracy comparable to existing methods.

Main Methods:

• Prospective enrollment of 242 patients with newly diagnosed pleural effusion.
• Measurement of pleural fluid biomarkers including ADA, IL-2, IL-18, IP-10, and TNF-α.
• Development of predictive models using ROC analysis and logistic regression.

Main Results:

• Two predictive models with AUC > 0.95 were developed.
• Model 1: body temperature, white blood cell count, pleural fluid ADA, and IP-10.
• Model 2: age, sex, body temperature, white blood cell count, lymphocyte percentage, and IP-10.

Conclusions:

• Developed models show very high diagnostic performance for TPE vs. non-TPE.
• These models can potentially be used in clinical practice.
• Provides valuable tools for differentiating pleural effusion types.