Nanostructured High-Entropy Yolk-Shell Oxides Platform for Efficient Metabolic Profiling of Diabetic Retinopathy

Diabetic retinopathy (DR) is a leading cause of blindness in middle-aged and elderly populations worldwide, and its diagnosis remains challenging due to the reliance on subjective interpretation of retinal images by experienced ophthalmologists. Here, we report a nanoenabled metabolomics strategy based on laser desorption/ionization mass spectrometry (LDI MS) for accurate DR discrimination. High-entropy yolk-shell metal oxides were rationally designed as a novel inorganic LDI matrix. Owing to the pronounced lattice distortion and multicomponent "cocktail" effect, these high-entropy oxides exhibit enhanced light-matter interaction, efficient energy transfer, and superior desorption/ionization behavior compared with conventional trimetallic oxides. Leveraging this matrix, high-quality serum metabolic fingerprints were obtained from DR patients and healthy controls. Machine learning analysis using a Random Forest model achieved an area under the curve (AUC) exceeding 0.99 for group discrimination. Notably, a compact DR differentiation panel composed of five key metabolites (DR-D5) retained diagnostic performance comparable to the full fingerprint model, delivering an accuracy above 93%. Subsequent pathway analysis revealed metabolic perturbations associated with DR pathogenesis. This work establishes an integrated workflow encompassing nanostructured matrix design, high-performance metabolite detection, and data-driven modeling, offering a high-throughput and precise metabolomics platform with strong potential for noninvasive clinical screening of diabetic retinopathy.