Diagnostic and prognostic perspectives of fabry disease via fiber evanescent wave spectroscopy advanced by machine learning.

Fabry disease (FD) is a rare disorder resulting from a genetic mutation characterized by the accumulation of sphingolipids in various cells throughout the human body, leading to progressive and irreversible organ damage, particularly in males. Genetically-determined deficiency or reduced activity of the enzyme (alpha - Galactosidase; α-Gal) leads to the accumulation of sphingolipids in the lysosomes of various cell types, including the heart, kidneys, skin, eyes, central nervous system, and digestive system, triggering damage, leading to the failure of vital organs, and resulting in progressive disability and premature death. FD diagnostics currently depend on costly and time-intensive genetic tests and enzymatic analysis, often leading to delayed or inaccurate diagnoses, which contribute to rapid disease progression. In this research, mid-infrared Fiber Evanescent Wave Spectroscopy (FEWS) supported by statistical analysis and Machine Learning (ML) algorithms is shown to be an innovative and reliable method to detect globotriaosylsphingosine (Lyso-Gb3) FD biomarker in urine and serum samples by monitoring infrared spectra alone. ML showed a high selectivity for FD in the spectral range of Amide A and Amide I in blood serum, and α-D-galactosyl residues of glycosphingolipids in urine. The developed approach offers a promising, cost-effective express diagnostic tool sensitive enough for early FD detection and monitoring.