Development of AI-integrated smartphone sponge-based sensors utilizing His@Co-NC nanozymes for highly sensitive sarcosine detection

The increasing demand for point-of-care detection of low-concentration cancer biomarkers has necessitated the development of innovative nanozyme-based sensing technologies. Here, a smartphone-integrated platform is presented that utilizes artificial intelligence (AI) to detect abnormal levels of the prostate cancer marker sarcosine (Sar) in humans. The platform comprises a histidine (His)-modified single-atom cobalt nanozyme (His@Co-NC), a white absorbent sponge for color development, and an AI-powered image acquisition system utilizing Monte Carlo color analysis (MC-CA). The biomimetic coordination of His with single-atom Co significantly enhances its peroxidase-like activity, rendering it 48-fold more active than natural horseradish peroxidase (HRP), with a Michaelis constant (Km) of 0.076 mM, much lower than the 3.7 mM observed for HRP. This enhanced activity results in the generation of more intense colorimetric signals when reacting with 3,3′,5,5′-tetramethylbenzidine (TMB). The AI algorithm captures over 1000 sampling RGB points on the color-revealing sponge surface, achieving a color recognition rate of over 95 % within 5 s, ensuring rapid and accurate detection. The integration of nanozyme with AI-driven data acquisition addresses two critical challenges: the limited chromogenic intensity in low-centration biomarker detection and the potential subjectivity in color rendering results. Immobilized on a hierarchically porous sponge, the system achieves a detection limit of 0.28 μM, with less than 5 % signal variation across 20 repeated cycles. This integration of AI and nanozyme-based detection holds significant promise for advancing biosensors that provide precise data output, offering vast potential for widespread applications in the early detection of disease biomarkers and other analytes.