Fast and label-free optical detection of dysplastic and tumour brain tissues

Surgery is the usual treatment for removing malformations and tumours in brain; however, the lack of contrast between diseased tissues and normal brain is a major problem. Magnetic Resonance Imaging (MRI) can be used to detect them, but brain shifts may severely reduce the accuracy of surgical removal procedures. In this framework, optical spectroscopy - being a fast and label-free method for analysing tissue composition - has the potential for improving detection and diagnosis of diseased areas. In this study, we used a quadrifurcated optical fibre-probe system combining multiple spectroscopic techniques for analysing ex vivo human brain freshly excised biopsies taken from both tumour and dysplastic tissues. All spectral recordings were done immediately after surgical resection, requiring less than 2 minutes for each sample. The recorded data were analysed using Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) for obtaining an automated classification of the examined samples based on the intrinsic spectral information provided by all three techniques. Significant differences were observed between dysplastic and tumour spectra, resulting in high sensitivity (83%) and specificity (73%). In particular, diffuse reflectance and UV-excited fluorescence spectroscopies provided the highest accuracies in discriminating different tissue types (78% and 75%, respectively) in good agreement with the corresponding histopathological examination; moreover, their combination with Raman spectroscopy resulted in a further improve of the classification capability up to 85%. The presented method demonstrates the huge potential of multimodal spectroscopy for the examination of brain tissues and opens the way for possible applications in surgical environment.