Machine Learning–Assisted Optimized Ag/Fe ₂ O ₃ /CNT/SrTiO ₃ Surface Plasmon Resonance Biosensor for Refractive Index–Based Breast Cancer Detection

ABSTRACT Sensitive and reliable differentiation of refractive index variations associated with normal and malignant breast tissues requires highly responsive optical sensing platforms. In this study, we propose an optimized Kretschmann‐configured surface plasmon resonance (SPR) biosensor for refractive index–based breast cancer detection, employing a multilayer heterostructure composed of a BK7 prism, silver (Ag), iron oxide (Fe 2 O 3 ), and carbon nanotubes (CNT), complemented by functional overlayers of strontium titanate (SrTiO 3 ) and platinum (Pt). Comprehensive parametric simulations were carried out using COMSOL Multiphysics to optimize the thicknesses of Ag, CNT, and Fe 2 O 3 layers and to evaluate their influence on reflectance characteristics and plasmonic coupling efficiency. The optimized BK7/Ag/Fe 2 O 3 /CNT/SrTiO 3 configuration exhibits excellent sensing performance, achieving a maximum angular sensitivity of 900°/RIU, a high figure of merit (FOM) of 456.6 RIU −1 , a low detection limit (DL) of 0.014 RIU, and a signal‐to‐noise ratio (SNR) of 5.48 across a refractive index range of 1.385–1.401, corresponding to reported optical properties of normal and malignant breast tissues. While the Pt‐overlaid configuration delivers a slightly higher peak sensitivity of 908.33°/RIU, it exhibits reduced stability at higher refractive index values. Furthermore, machine learning–based predictive models accurately capture the sensor response to Ag thickness variations, achieving coefficients of determination ( R 2 ) exceeding 0.99, thereby validating the robustness and predictability of the proposed design. The obtained results demonstrate the potential of the proposed multilayer SPR architecture as a high‐performance platform for refractive index–based breast cancer sensing and future biomarker‐oriented investigations.