ITO-based multi-band metasurface absorber for early detection of hepatic cancer through deep learning.

This study introduces a multi-band metasurface absorber, optimized for high-sensitivity detection of hepatic (liver) cancer. The unit cell comprises a top ITO sheet with several "+" - shaped slots, in which a graphene layer is incorporated into the central slot. This layer is attached to a quartz glass substrate to provide optical clarity, while a continuous ITO ground layer is used to mitigate electromagnetic (EM) transmission. Simulation results indicate that the structure attains exceptional absorption of 95%, 99%, 96%, and 94% at 0.594 (f1), 0.988 (f2), 1.196 (f3), and 1.518 (f4) THz, respectively. Electric field distribution and an equivalent electric circuit (EEC) analysis are used to theoretically analyze and confirm the resonance properties and physical operating mechanism. Furthermore, the proposed multi-band metasurface absorber demonstrates a sensitivity of 36 GHz/RIU, a Figure of Merit (FOM) of 2045 RIU-1, and a quality factor (Q-factor) of 70, thereby enabling reliable and precise detection of minor refractive index fluctuations associated with hepatic cancer biomarkers. To improve diagnostic accuracy, a deep learning (DL) analysis framework is used to distinguish cancerous hepatic tissues. The model demonstrates 99.9% accuracy on the training data, with a validation accuracy of 98.7%. Owing to its multi-band adaptability and DL-driven enhanced detection accuracy, the proposed work provides a revolutionary foundation for next-generation visually accessible oncological diagnostics and advanced biomedical imaging.