Plasmonic biosensors based on X-shaped resonators for cancer cell detection with machine learning-driven transmittance prediction

The measurement of the refractive index (RI) of biological tissues plays a fundamental role in bio-optical sensing systems. In this work, 5 plasmonic biosensors are proposed based on metal–insulator–metal (MIM) waveguides coupled with 1 to 5 X-shaped resonators (XSRs). By increasing the number of coupled resonators, the sharpness of the photonic band gap (PBG) edges is significantly improved, resulting in a substantial enhancement of the figure of merit (FoM) from 5.82 to 80.76. The distinct spectral separation between the two PBG edges also ensures stable sensing behavior without overlapping under changes in analyte RI, enabling broadband sensing capability. The proposed biosensors are further evaluated for cancer cell detection, demonstrating high sensitivity and reliable performance. In addition, transmission responses are predicted using an Extreme Randomized Tree Regression model, which reduces computational time by nearly 90% compared with finite-difference time-domain (FDTD) simulations. Overall, the combination of a simple structure, improved FoM, high sensitivity, and machine learning-based prediction highlights the potential of these biosensors for advanced biomedical sensing applications.