
Colorectal cancer biomarkers including methylated SDC2 (mSDC2), methylated SEPT9 (mSEPT9), and carcinoembryonic antigen (CEA), are effective biomarkers for the early detection and monitoring of colorectal cancer. In this study, a high-performance surface plasmon resonance (SPR) biosensor incorporating MgF2 prism/CaTiO3/Cu/Al2O3/BP multilayers is designed to enhance light-matter interaction and improve sensing performance. The brute force algorithm is employed to optimize the sensor parameters. Also, design is systematically optimized and investigated by using TMM, FEM and FDTD simulations. The accuracy of ANN, ANFIS and RBFNN system models in predicting sensor responses is thoroughly validated. The optimized biosensor achieves a sensitivity of 479.58 deg/RIU, figure of merit of 152.01, and quality factor of 167.86 RIU-1 at a wavelength of 633 nm, demonstrating the potential of the optimized architecture for sensitive refractive-index sensing relevant to future colorectal cancer biomarker detection. (c) 2026 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement