Integrated colorimetric immunoassay for PSA using confined Pt nanozymes and smartphone-assisted machine learning analysis

Journal article, 2026

Prostate-specific antigen (PSA) is the most widely used tumor marker for prostate cancer (PCa), and its accurate detection is critical for early diagnosis and intelligent identification of the disease. Here, we present a highly sensitive magnetic colorimetric immunosensor for PSA detection, constructed by integrating mesoporous silica nanoparticles (MSN) loaded with confined platinum (Pt) nanozymes (MSN-Pt) and Fe3O4@SiO2 magnetic microspheres. The MSN-Pt nanozymes exhibit exceptional catalytic activity in the TMB reaction, producing a visible colorimetric signal that enables quantification of PSA across a broad dynamic range (0.05 pg·mL⁻¹ to 50 ng·mL⁻¹) with a detection limit as low as 1.1 fg·mL⁻¹. To enable intelligent diagnosis, colorimetric responses captured by smartphones were analyzed through machine learning algorithms, including linear discriminant analysis (LDA), support vector machine (SVM), random forest (RF), k-nearest neighbors (KNN), naïve Bayes (NB), neural network (NN), and decision tree (DT). Among these, LDA achieved 100 % accurate differentiation of PCa status based on RGB value thresholds (R > 132 and B < 210 indicating healthy samples; R < 124 indicating PCa). This integrated sensing and data-processing platform provides a robust, rapid, and precise tool for PSA detection and PCa identification. Our study not only advances the design of nanozyme-based sensors with enhanced catalytic activity but also demonstrates the clinical potential of combining colorimetric biosensing with machine learning and smartphone technology for intelligent biomarker detection in complex biological samples.