Hematological Biomarkers of the Obstructive Sleep Apnea Syndrome: A Machine Learning-Based Diagnostic and Prognostic...

Area of Science:

• Sleep Medicine
• Immunology
• Biomarker Research

Background:

• Obstructive sleep apnea syndrome (OSAS) is a prevalent condition associated with systemic inflammation.
• Diagnostic and prognostic tools for OSAS severity and treatment response are crucial for effective patient management.

Purpose of the Study:

• To evaluate the diagnostic and prognostic value of C-reactive protein (CRP), systemic immune-inflammation index (SII), and fibrinogen in OSAS patients.
• To develop a machine learning model for stratifying OSAS severity and predicting treatment outcomes.

Main Methods:

• A prospective observational cohort study involving 195 adult OSAS patients treated with CPAP or surgery.
• Measurement of hematologic biomarkers (CRP, SII, fibrinogen) pre- and post-treatment.
• Utilized polysomnography (PSG) for staging, ROC analysis, and Random Forest machine learning for biomarker evaluation.

Main Results:

• CRP exhibited high diagnostic accuracy for severe OSAS (AUC=0.91).
• Fibrinogen strongly correlated with OSAS severity (ρ=0.81) and showed significant post-treatment reduction.
• Machine learning identified CRP, SII, and fibrinogen as key predictors; three inflammatory OSAS phenotypes were identified.

Conclusions:

• CRP, SII, and fibrinogen show potential for OSAS risk stratification and monitoring.
• Exploratory findings require larger, multicenter validation before routine clinical implementation of biomarkers or the ML model.