Graphene and Vanadium Dioxide-Based THz Biosensor for Breast Cancer Detection: Machine Learning-Driven Performance Optimization
Abstract
Early and accurate detection of breast tissue abnormalities is crucial for reducing breast cancer rates. This work proposes a biosensor based on 2D materials operating in the terahertz (THz) frequency range, integrating graphene and
vanadium dioxide to enhance sensitivity. The Wave Concept Iterative Process
(WCIP) method is employed to model the interaction of healthy fat, fibrous,
and tumor breast tissues with the biosensor. A novel implementation of the
WCIP algorithm enables an extensive simulation study, comparing theoretical
and simulated results for 2D materials. The findings demonstrate excellent
agreement between analytical predictions and WCIP simulations. To optimize
the sensor’s performance, a deep neural network (DNN) model is developed to
predict sensitivity values based on key material parameters, including vanadium
dioxide conductivity, graphene relaxation time, temperature, and chemical po-