Using Machine Learning to Detect Theranostic Biomarkers Predicting Respiratory Treatment Response

Area of Science:

• Biomarker discovery
• Precision medicine
• Respiratory medicine

Background:

• Theranostic approaches are emerging in precision medicine, primarily in oncology.
• Limited research exists on theranostics within respiratory medicine.
• This study addresses the gap by investigating theranostic biomarkers for respiratory treatments.

Purpose of the Study:

• To identify theranostic biomarkers associated with respiratory treatment responses.
• To advance the understanding and application of biomarkers in diagnosing respiratory diseases.
• To contribute to the development of targeted therapies for respiratory conditions.

Main Methods:

• Cross-sectional analysis of 13,102 adults from the UK Household Longitudinal Study.
• Recursive feature selection identified 16 key biomarkers.
• Machine learning algorithms predicted treatment response using biomarkers, age, sex, BMI, and lung function.

Main Results:

• Increased levels of alkaline phosphatase, glycated hemoglobin, HDL cholesterol, C-reactive protein, triglycerides, hemoglobin, and Clauss fibrinogen were associated with receiving respiratory treatments.
• These associations remained significant after adjusting for age, sex, body mass index, and lung function.

Conclusions:

• The identified biomarkers provide a diagnostic blueprint for respiratory diseases.
• Biomarker utilization can guide treatment management and personalize respiratory care.
• This research supports the integration of theranostics in respiratory medicine.