Detection of lung cancer cells via high-resolution near-infrared spectra of extracellular biofluids.

Abnormal metabolism of cancer cells leads to changes in the chemical composition of extracellular fluids, which may serve as potential indicators for cancer detection. In this study, a method was developed to investigate cancer-related signatures in extracellular biofluids using near-infrared (NIR) spectroscopy combined with chemometrics. High-resolution NIR spectra were obtained through wavelet packet transform (WPT) and were applied to analyzing the culture supernatants collected from lung cancer (A549, H1975) and normal bronchial epithelial (BEAS-2B) cells. Cancerous and normal fluids were successfully distinguished with a classification accuracy of 100 % using partial least squares discriminant analysis (PLS-DA) combined with Monte Carlo uninformative variable elimination (MC-UVE). The selected informative wavelengths are associated with the absorptions from both biological components (56.8 %) and water molecules (43.2 %). Moreover, according to the spectral variations during cell cultivation, stable metabolic behaviors and a progressive increase of weakly hydrogen-bonded water structures were revealed in normal samples. In contrast, the aberrant spectral variation trends observed in the extracellular fluids of cancer cells may be attributed to the excessive release of lactate and pyruvate, resulting in an acidic microenvironment that disrupts the hydrogen-bond network of water. These findings suggest that high-resolution NIR spectroscopy, enhanced by wavelet packet transform, provides a sensitive strategy for cancer detection. Moreover, water structure with different hydrogen bonds may serve as a potential probe for sensing cancer-associated changes in biofluids.