An antibody modified terahertz (THz) metamaterial biosensor is proposed to detect the concentration of carcinoembryonic antigen (CEA). The biosensor has four metal split-ring-resonators (SRR) in a unit cell. The finite integration time domain (FITD) method is used to design and optimize the structure. The simulation shows that the biosensor is insensitive to incident angles and the sensitivity reaches to 76.5 GHz/RIU (Refractive Index Unit). The sample is manufactured using a surface micromachining process and characterized by a THz time domain spectroscopy (THz-TDS) system. The experimental results indicate that the resonant frequency of the biosensor decreases with the increase of the concentration of analytes on the surface. After that, the biosensor is modified with anti-CEA and used to detect the concentration of the CEA. Interestingly, the response of the biosensor depends heavily on the concentration of the Anti-CEA modified on the surface of the biosensor. When the concentration of Anti-CEA is 20 ng/ml, the resonant frequency shift shows good linearity to the concentration of CEA and the limit of detection (LOD) reaches to 0.1 ng/ml. This study paves a new way for sensitively detection of biomolecules, cancer biomarkers and immune responses, which is important for early stage diagnosis of cancers.

We propose a simple method for fast and label-free identification of early-stage cervical cancerous tissues using a metamaterial (MM) terahertz (THz) biosensor with two resonant absorption frequencies at 0.286 THz and 0.850 THz, respectively. Because the two resonant absorption frequencies are very sensitive to the ambient change of dielectric properties, the cervical cancerous tissues can be distinguished from the normal tissues by analyzing the resonant frequency shift. The calculated sensitivity of the low resonant frequency and that of the high resonant frequency are 29 and 74 GHz/RIU, respectively. In the experiment, the cervical cancerous tissue samples were taken from different patients, and the cancerous and normal parts were determined by hematoxylin-eosin (HE) stain. Compared with the traditional method of directly detecting cancer tissue without the sensor, this method has the advantages of being simple, label free and fast. In addition, compared with biosensors which have a single resonant absorption frequency, the biosensors with double resonant absorption frequencies have higher accuracy. This method potentially can be applied for the detection and diagnosis of many other cancer tissues.

Squamous cell carcinoma antigen (SCCA) level is an auxiliary indicator for lung cancer diagnosis and predictive monitoring, and its standard reference value is less than 1.5 ng/mL. Therefore, achieving high sensitivity and specific detection of the SCCA is a challenge. In this work, we designed a double split-ring-resonator (SRR) terahertz (THz) metasurface sensor. Simulation results show that the sensitivity of the sensor can reach 103 GHz/RIU (refractive index unit). After the sensor is modified with different concentrations of halloysite-nanotubes (HNTs), the resonant frequency of the sensor exhibits different linear responses to different concentrations of SCCA. The optimum concentration of HNTs to enhance the detection performance of the sensor was obtained as 10 mg/mL. The linear range of SCCA concentration of the HNTs modified sensor was 1 pg/mL-100 ng/mL. On this basis, we modified anti-SCCA on the sensor, and other biomarkers of carcinoembryonic-antigen (CEA), neuron-specific-enolase (NSE), CA125, and CA199 were used as comparisons. The results indicated that the sensor modified with HNTs showed a more remarkable frequency shift for the SCCA, which verified that the sensor had good specificity for the SCCA. Therefore, this sensor can be used for highly sensitive and specific detection of the SCCA.

Terahertz (THz) biosensors provide a new avenue for the early diagnosis and treatment of cancers. However, conventional THz biosensors cannot achieve high sensitivity due to limited absorption peaks. In this work, we propose a cancer biomarker ultrasensitive detection method based on THz frequency-comb-like. At first, we design a basic biosensor which has three resonant absorption peaks. Then, we proportionally change the size of basic biosensor and obtain other 4 biosensors with different resonant absorption peaks. Finally, a frequency-comb-like containing 8 uniformly spaced absorption peaks in the range of 0.5 - 1.3 THz is realized by appropriately selecting the absorption peaks of these 5 biosensors. Three cancer biomarkers of carcinoembryonic antigen (CEA), neuron-specific enolase (NSE) and squamous cell carcinoma (SCC) are used as samples of experiment. Under the same concentration, the frequency shifts generated by each peak are accumulated. For samples from 0.1 pg/mL to 50 ng/mL, the cumulative frequency shifts are 122-547 GHz, 168-815 GHz, and 142-505 GHz, respectively. The sensitivity (Sc) of concentration are 64.995 log(ng/ml), 105.005 log(ng/ml), and 62.634 log(ng/ml), respectively. The limit-of-detection (LOD) is 0.1 pg/mL. Compared with other methods using single or few resonant absorption peaks, this method shows higher sensitivity and lower LOD. Therefore, it is valuable for trace molecule detection, biomedical sensing and cancer diagnosis at early stage.

MicroRNA (miRNA) is closely related to various cancers, and the change in its expression level is closely related to the development and death of tumor cells. Here, we design and manufacture a terahertz (THz) metasurface sensor to realize the concentration detection and category identification of the miRNAs related to lung cancer. We established two spectral classification datasets, one contains 9 concentrations of miRNAs, and the other contains 3 categories of lung cancer-related miRNAs. And then, we used a deep neural network (DNN) algorithm to classify these spectral datasets and obtained a LOD (Limit-of-Detection) of 100 aM for miRNAs. Moreover, this method can identify the categories of miRNAs. Compared with the other five machine learning (ML) algorithms, the proposed neural network framework achieves the best classification results. This work provides a new way for the detection and identification of trace nucleic acid and biomarkers of many cancers.