Lists
Lists
Hamza, Musa N., Alibakhshikenari, Mohammad, Virdee, Bal Singh, Lavadiya, Sunil, Din, Iftikhar Ud, Sanches, Bruno, Koziel, Slawomir, Naqvi, Syeda Iffat, Panda, Abinash, Farmani, Ali, Mezache, Zinelabiddine, Zakeri, Hassan, Naser-Moghadasi, Mohammad and Saber, Takfarinas (2026) An AI-enhanced multiband terahertz metamaterial biosensor for intelligent leukaemia detection. IET Wireless Sensor Systems, 16(1) (e70039). pp. 1-28. ISSN 2043-6386, 2043-6394
This study presents an artificial intelligence (AI)-augmented micron-scale terahertz (THz) metamaterial biosensor for early-stage leukaemia diagnosis. The proposed biosensor employs a tri-negative (ε, μ, n) perfect absorber structure with an optimised multiband spectral response, enabling high-Q narrowband resonances across the 0.5–2 THz frequency range. These engineered electromagnetic characteristics enhance field confinement and analyte interaction, enabling the detection of refractive index variations as small as ∼0.014 RIU between healthy and leukaemia-affected blood samples. Spectral and field analyses reveal distinct resonance shifts, absorption variations and altered electric and magnetic field distributions in the presence of cancerous samples. Quantitative performance evaluation demonstrates excellent sensing characteristics, including a Q-factor of 268.34, a figure of merit (FOM) of 56722.80 RIU−1, and Euclidean sensitivity of 355.859564 THz RIU−1. These values indicate improved performance compared with previously reported THz biosensors for cancer detection. AI integration enables automated classification of healthy and cancerous samples using S-parameter processing and full-spectrum similarity metrics, achieving classification accuracy exceeding 95%. Comparative analysis with state-of-the-art biosensors demonstrates leukaemia-specific dielectric targeting, early-stage detection capability and an AI-assisted analytical framework for enhanced spectral interpretation. The results highlight the potential of the proposed platform as a noninvasive and label-free approach for blood cancer diagnostics, combining high spectral resolution, full-spectrum analysis and automated decision-making within a unified sensing framework.
Available under License Creative Commons Attribution 4.0.
Download (7MB) | Preview
 |
View Item |