Subscribe to RSS
DOI: 10.1055/s-0043-1767040
Raman difference spectroscopy as a non-invasive method for the diagnosis of squamous cell carcinoma in the hypopharynx – a feasibility study
Background Squamous cell carcinoma of the hypopharynx (HNSCC) is often diagnosed when it is locally advanced. The aim of this study was to use Raman spectroscopy for the bio-optical characterization of HNSCC and thus to test the potential of real-time diagnostics in-vivo and chair-side.
Methods As an optical method, Raman spectroscopy is based on inelastic scattering of monochromatic light (Raman effect). The SERDS technique (shifted-excitation Raman difference spectroscopy) was used to reduce fluorescent interference in biological tissues. The processing and evaluation of the spectral data was carried out using specially established methods using neural networks (U-Net), principal component analysis (PCA) and linear discriminant function (LDA) after referencing with the histopathological work-up.
Results Physiological and diseased (HNSCC) ex-vivo tissue samples from 9 patients (location: hypopharynx) were compared. A total of 39 measurement points were evaluated (18 measurement points = healthy, 21 measurement points = HNSCC). After processing the spectral data, characteristic differences in the protein, phosphate, lipid and nucleic acid signatures between physiological and diseased tissues were revealed. The tumor samples could mostly be identified accurately in this collective (sensitivity 71.4%, specificity 72.2%, overall accuracy 71.8%).
Conclusion With the method presented, a “chemical fingerprint” of the examined tissue could be developed in a reproducible and valid manner on the available patient population. In order to increase the diagnostic accuracy, further investigations with refinement of the measurement technology and an increase in the collective are planned.
Wilhelm Sander-Stiftung für Krebsforschung
Publication History
Article published online:
12 May 2023
Georg Thieme Verlag
Rüdigerstraße 14, 70469 Stuttgart, Germany