3.2.6. Head and Neck Cancer

Currently, there is a lack of well-established effective biomarkers and convenient detection methods for predicting radioresistance. In the study by Wu et al. [118], surfaceenhanced Raman spectroscopy combined with proteomics was employed to initially profile the distinct spectral patterns of the exosomes released by self-established nasopharyngeal cancer cells (NPC). They identified specific variations in protein expression during the formation of radioresistance, including collagen alpha-2 (COL1A2), which is negatively associated with DNA repair. The researchers used bioinformatics analysis and a deep learning model to accurately identify the exosomes from the radioresistance group, achieving an accuracy of 92.4%. Overall, the study provides a promising approach for identifying radioresistance-associated biomarkers in NPC.

Diagnosing cancer and other diseases using data from non-specific sensors, like electronic tongues (e-tongues), poses challenges due to the lack of selectivity and the variability of biological samples [119]. Braz et al. [119] presented an e-tongue biosensor based on microfluidic impedance flow cytometry for mouth cancer detection. Saliva samples from 27 individuals were analyzed using multidimensional projection techniques and machine learning algorithms, including the SVM with a radial basis function kernel and RF algorithms. The authors achieved an accuracy of over 80% for the binary classification of cancer vs. healthy individuals. The study suggests that the impedance data obtained with the e-tongue in saliva samples can be used for cancer diagnosis in the mouth, and the approach presented here is promising for computer-assisted diagnoses. The accuracy tended to increase when clinical information, such as alcohol consumption, was used in conjunction with the e-tongue data.
