Metasurface Based Surface Plasmon Resonance (SPR) Biosensor for Cervical Cancer Detection with Behaviour Prediction using Machine Learning Optimization Based on Support Vector Regression
Jacob Wekalao,Mouleeswaran Singanallur Kumaresan,Srinivasan Mallan,Garapati Satyanarayana Murthy,Nagarajan Ramanathan Nagarajan,Santhanakrishnan Karthikeyan,Nithya Dorairajan,Ramachandran Thandaiah Prabu,Ahmed Nabih Zaki Rashed
DOI: https://doi.org/10.1007/s11468-024-02623-8
IF: 2.726
2024-10-30
Plasmonics
Abstract:Early detection of cervical cancer, a leading cause of morbidity and mortality among women globally, is crucial for successful treatment. Existing diagnostic methods, while effective, face limitations including cost, accessibility, and reliance on specialized personnel, especially in resource-constrained settings. Recent advancements in nanomaterials and metasurfaces offer promising avenues for developing highly sensitive, label-free biosensors capable of detecting cancer biomarkers at minute concentrations. This study presents the design and theoretical modelling of a novel metasurface-based sensor for cervical cancer detection, utilizing graphene, black phosphorus, and titanium dioxide as core sensing materials. The sensor exhibits dual-band operation (1.369–1.383 THz and 0.313–0.317 THz) with exceptional performance metrics: sensitivity reaching 400 GHzRIU −1 , a figure of merit of 5.882 RIU −1 , and quality factors ranging from 9.206 to 9.950. Furthermore,the sensor's dual-band functionality and 2-bit encoding capability suggest its potential for multi-parametric analysis and information processing, paving the way for more comprehensive diagnostic approaches. Additionally, integration of Support Vector Regression (SVR) with a Polynomial Kernel demonstrates remarkable performance, achieving an optimal R 2 score of 100%. This approach significantly reduces simulation time (80%) and resource requirements for sensor optimization.
materials science, multidisciplinary,nanoscience & nanotechnology,chemistry, physical