Lung cancer is the most frequently life-threatening disease and the prominent cause of cancer-related mortality among human beings worldwide, where poor early diagnosis and expensive detection costs are considered as significant reasons. Here, we try to tackle this issue by proposing a novel label-free and low-cost strategy for rapid detection and distinction of lung cancer cells relying on plasmonic toroidal metasurfaces at terahertz frequencies. Three disjoint regions are displayed in identifiable intensity-frequency diagram, which could directly help doctors determine the type of lung cancer cells for clinical treatment. The metasurface is generated by two mirrored gold split ring resonators with subwavelength sizes. When placing analytes on the metasurface, apparent shifts of both the resonance frequency and the resonance depth can be observed in the terahertz transmission spectra. The theoretical sensitivity of the biosensor over the reflective index reaches as high as 485.3 GHz/RIU. Moreover, the proposed metasurface shows high angular stability for oblique incident angle from 0 to 30°, where the maximum resonance frequency shift is less than 0.66% and the maximum transmittance variation keeps below 1.33%. To experimentally verify the sensing strategy, three types of non-small cell lung cancer cells (Calu-1, A427, and 95D) are cultured with different concentrations and their terahertz transmission spectra are measured with the proposed metasurface biosensor. The two-dimensional fingerprint diagram considering both the frequency and transmittance variations of the toroidal resonance dip is obtained, where the curves for different cells are completely separated with each other. This implies that we can directly distinguish the type of the analyte cells and its concentration by only single spectral measurement. We envisage that the proposed strategy has potential for clinical diagnosis and significantly expands the capabilities of plasmonic metamaterials in biological detection.

Identification of novel non-invasive biomarkers is critical for the early diagnosis of lung adenocarcinoma (LUAD), especially for the accurate classification of pulmonary nodule. Here, a multiplexed assay is developed on an optimized nanoparticle-based laser desorption/ionization mass spectrometry platform for the sensitive and selective detection of serum metabolic fingerprints (SMFs). Integrative SMFs based multi-modal platforms are constructed for the early detection of LUAD and the classification of pulmonary nodule. The dual modal model, metabolic fingerprints with protein tumor marker neural network (MP-NN), integrating SMFs with protein tumor marker carcinoembryonic antigen (CEA) via deep learning, shows superior performance compared with the single modal model Met-NN (p < 0.001). Based on MP-NN, the tri modal model MPI-RF integrating SMFs, tumor marker CEA, and image features via random forest demonstrates significantly higher performance than the clinical models (Mayo Clinic and Veterans Affairs) and the image artificial intelligence in pulmonary nodule classification (p < 0.001). The developed platforms would be promising tools for LUAD screening and pulmonary nodule management, paving the conceptual and practical foundation for the clinical application of omics tools.

Background: The intravoxel incoherent motion (IVIM) method of magnetic resonance imaging (MRI) analysis can provide information regarding many physiological and pathological processes. This study aimed to investigate whether IVIM-derived parameters and the apparent diffusion coefficient (ADC) can act as imaging biomarkers for predicting non-small cell lung cancer (NSCLC) response to anti-tumor therapy and compare their performances.Methods: This prospective study included 45 patients with NSCLC treated with chemotherapy (29 men and 16 women, mean age 57.9±9.7 years). Diffusion-weighted imaging was performed with 13 b-values before and 2-4 weeks after treatment. The IVIM parameter pseudo-diffusion coefficient (D*), perfusion fraction (f), diffusion coefficient (D), and ADC from a mono-exponential model were obtained. Responses 2 months after chemotherapy were assessed. The diagnostic performance was evaluated, and optimal cutoff values were determined by receiver operating characteristic (ROC) curve analysis, and the differences of progression-free survival (PFS) in groups of responders and non-responders were tested by Cox regression and Kaplan-Meier survival analyses.Results: Of 45 patients, 30 (66.7%) were categorized as responders, and 15 as non-responders. Differences in the diffusion coefficient D and ADC between responders and non-responders were statistically significant (all P<0.05). Conversely, differences in f and D* between responders and non-responders were both not statistically significance (all P>0.05). The ROC analyses showed the change in D value (ΔD) was the best predictor of early response to anti-tumor therapy [area under the ROC curve (AUC), 0.764]. The Coxregression model showed that all ADC and D parameters were independent predictors of PFS, with a range of reduction in risk from 56.2% to 82.7%, and ΔD criteria responders had the highest reduction (82.7%).Conclusions: ADC and D derived from IVIM are potentially useful for the prediction of NSCLC treatment response to anti-tumor therapy. Although ΔD is best at predicting response to treatment, ΔADC