AI‐Assisted Ultrasensitive Biosensing using Metal‐Electrolyte‐Metal‐Insulator‐Silicon Structure Based on Oxygen‐Tunable Iridium Oxide Nanonets for Lysyl‐Oxidase‐Like‐2 Breast Cancer Detection

ABSTRACT Early detection of breast cancer metastasis is critical for improving survival rates, yet identifying specific biomarkers such as Lysyl‐Oxidase‐Like‐2 (LOXL2) at low concentrations remains challenging. A highly sensitive, label‐free sensor utilizing a Metal‐Electrolyte‐Metal‐Insulator‐Silicon (MEMIS) configuration based on oxygen‐controlled iridium oxide (IrO x ) nano‐nets is reported for LOXL2 detection. By optimizing the reactive sputtering oxygen content (50%) and film thickness (2 nm), the IrOx nano‐net structure achieves a super‐Nernstian pH sensitivity of 150.4 mV/pH with minimal drift (1.95 mV/h). This exceptional sensitivity is attributed to high porosity and abundant redox‐active sites of the crystalline nano‐nets, which facilitate volumetric potential accumulation and significantly enhance the Limit of Detection (LoD) to 0.05 nM. To address subtle signal variations often obscured by noise, a dual‐branch 1D Convolutional Neural Network (1D‐CNN) was integrated to simulate and analyze sensor performance. The Artificial Intelligence (AI) model successfully identifies characteristic sub‐threshold voltage shifts associated with low‐abundance biomarkers, achieving diagnostic accuracy of 94.8%. This MEMIS‐AI platform serves as a promising bridge toward future ultrasensitive clinical screening.