Comprehensive analysis of THz metasurface imaging for glioblastoma in orthotopic xenograft mouse model

Abstract Terahertz optical detection and imaging technologies demonstrate great potential and challenges in various bio-applications and medical tools owing to low-energy non-ionizing characteristics and ultra-broadband spectral information, including many molecular fingerprints. Furthermore, recent advances in the terahertz sensing platform have accelerated the applicability with the help of various metasurfaces and enhanced the performance of optical detection and imaging technologies in terms of sensitivity and selectivity. This study demonstrates a highly advanced label-free terahertz imaging technique based on the enhanced signal via terahertz metasurface and further treatment exploiting an artificial neural network for mouse brain imaging beyond the absorption cross-section limit. The criteria for extracting the glioma part using the algorithm with more than 99% accuracy was verified using the fluorescence-labeled images as training sets. An accurate complex refractive index for brain cancer and normal tissues to expand the capability for the potential clinical applications of terahertz radiation was determined using a reliable numerical solution based on the finite element method simulation.