Advanced image processing and pattern-matching algorithms assisted enzyme/hydrogel platform for dual-signal detection of glucose

Abstract

This study presents a dual-signal biosensing platform for glucose detection based on alginate hydrogel beads (AlgelBeads). Its primary innovation lies in the integration of enzyme co-encapsulation with advanced image processing and pattern-matching algorithms. Glucose oxidase, horseradish peroxidase, and bovine serum albumin-templated gold nanoclusters are co-encapsulated within the AlgelBeads, which are uniform spheres approximately 2.1 mm in diameter, featuring a unique vein-like surface pattern that provides sites for glucose recognition and substrate catalysis. Upon exposure to glucose, an enzymatic cascade reaction is initiated, producing a visible color change to blue and simultaneous fluorescence quenching. For high-throughput analysis, a custom array plate was designed. The interpretation of the dual signals (colorimetric and fluorescent) is automated and enhanced using a Hough circle algorithm for AlgelBeads localization and a dictionary-based spatial classification algorithm for robust pattern recognition. Under optimized parameters, these AlgelBeads enable quantitative detection of glucose from 0.0625 to 4.0 mg/mL. The limits of detection are 0.077 mg/mL (colorimetric) and 0.030 mg/mL (fluorescent), with both precision and accuracy falling within acceptable limits. The AlgelBeads were applied for determining 54 serum samples collected from 18 pregnant women undergoing the oral glucose tolerance test, yielding results consistent with Beckman Coulter Glucose Assay Kit. These findings affirm the AlgelBeads offer an accurate and effective platform for glucose detection, holding potential for assisting in the diagnosis of gestational diabetes mellitus and other diabetes types in clinical settings.

Keywords: Alginate hydrogel; Dual-mode detection; Enzyme immobilization; Gestational diabetes mellitus; Glucose.