Plasma Metabolic Profile with Machine Learning Reveals Distinct Diagnostic and Biological Signatures for Pathologic Myopia.

Abstract

Pathologic myopia (PM), characterized by serious myopic macular degeneration (MMD), is a detrimental subtype of high myopia (HM) and has become one of the leading causes of blindness worldwide. In this concern, precise and high-throughput molecular diagnosis and further pathologic insights are urgently needed. Here, through the combined strategy of nanoparticle-enhanced laser desorption/ionization mass spectrometry-based rapid metabolic analysis (<30 s) and machine learning, a precise molecular diagnostic approach of PM (HM with MMD grade ≥ 2) is proposed, which achieves areas under the curve of 0.874 and 0.889 for diagnosing PM and early-stage PM, respectively. Further, the biomarkers indicate the PM-associated systemic metabolic reprogramming of amino acid and lipid metabolism, which may mediate dysfunctional oxidative stress, inflammation, hormone/neurotransmitter systems, and energy metabolism. Notably, MMD grade 4, featuring characteristic macula atrophy, exhibits specificity in this metabolic reprogramming. Of these biomarkers, azelaic acid shows a significant protective effect in the ARPE-19 cells under abnormal oxidative stress, which may be involved in PM development as a key antioxidative active metabolite. This work will contribute to PM molecular diagnosis and pathology exploration.