Metabolic Fingerprint of Dual Body Fluids Deciphers Diabetic Retinopathy.

Area of Science:

• Biomedical Science
• Analytical Chemistry
• Computational Biology

Background:

• Diabetic retinopathy (DR) is a major microvascular complication of diabetes, impacting millions globally.
• Early detection and treatment are crucial for improving DR prognosis.
• Metabolomic analysis offers a powerful method for understanding disease pathophysiology and identifying biomarkers.

Purpose of the Study:

• To establish an innovative workflow for vitreous liquid analysis.
• To develop a machine learning-based platform for DR detection using dual biofluid metabolic fingerprints.
• To identify potential biomarkers for clinical diagnosis and treatment of diabetic retinopathy.

Main Methods:

• Development of a nanoparticle-enhanced laser desorption/ionization mass spectrometry workflow for vitreous liquid and plasma metabolic fingerprinting.
• Integration of vitreous liquid metabolic fingerprint (VL-MF) and plasma metabolic fingerprint (P-MF) data.
• Application of machine learning for DR patient classification and biomarker panel construction.

Main Results:

• Direct VL-MF and P-MF were obtained with high reproducibility (CV <5%) and speed (3s/sample).
• The dual biofluid-MF platform distinguished DR patients from healthy controls with an AUC of 0.957.
• A biomarker candidate panel achieved an AUC of 0.945, with related metabolic pathways identified.

Conclusions:

• The developed multi-biofluid platform demonstrates high accuracy and speed for diabetic retinopathy detection.
• This approach facilitates the discovery of novel biomarkers and provides a robust tool for clinical applications.
• The study offers a significant advancement in leveraging metabolomics and machine learning for diagnosing diabetes complications.