Differentiation of glioblastoma tissues using spontaneous Raman scattering with dimensionality reduction and data classification

Abstract

The neurosurgery of intracranial tumors is often complicated by the difficulty of distinguishing tumor center, infiltration area, and normal tissue. The current standard for intraoperative navigation is fluorescent diagnostics with a fluorescent agent. This approach can be further enhanced by measuring the Raman spectrum of the tissue, which would provide additional information on its composition even in the absence of fluorescence. However, for the Raman spectra to be immediately helpful for a neurosurgeon, they must be additionally processed. In this work, we analyzed the Raman spectra of human brain glioblastoma multiforme tissue samples obtained during the surgery and investigated several approaches to dimensionality reduction and data classificatin to distinguish different types of tissues. In our study two approaches to Raman spectra dimensionality reduction were approbated and as a result we formulated new technique combining both of them: feature filtering based on the selection of those shifts which correspond to the biochemical components providing the statistically significant differences between groups of examined tissues (center of glioblastoma multiforme, tissues from infiltration area and normally appeared white matter) and principal component analysis. We applied the support vector machine to classify tissues after dimensionality reduction of registered Raman spectra. The accuracy of the classification of malignant tissues (tumor edge and center) and normal ones using the principal component analysis alone was 83% with sensitivity of 96% and specificity of 44%. With a combined technique of Frontiers in Oncology frontiersin.org 01

Optical Differentiation of Brain Tumors Based on Raman Spectroscopy and Cluster Analysis Methods

“…We cannot simply analyze the spectra point by point; we select characteristic regions from the spectra of all modalities, in which we calculate indices of the presence of certain components. These are included in our initial vector of features [28].…”

“…The authors of [29] found that the choice of features is fundamentally important for the outcome of the classification, and they were unable to achieve class separation when analyzing the full spectra. In the current work, we used the approach to feature filtering that we proposed earlier [28]. There, it was shown that the step of pre-filtering features before applying feature projection methods to reduce dimensionality significantly improves the classification results.…”

Machine Learning-Assisted Classification of Paraffin-Embedded Brain Tumors with Raman Spectroscopy

“…So far, the practical potential of this method in combination with supervised and unsupervised machine learning algorithms has been shown not only for the classification of tumor types and distinct histomorphological tumor areas but also for the differentiation of the tumor grade or the detection of tumor margins/infiltration zone [4][5][6][7]. Most recently, Zhang et al yielded >80% accuracy with support vector machinebased intraoperative differentiation of glioma tissue and healthy control, and Romanishkin et al reported a support vector machine-based accuracy in detecting glioblastoma tissue (regardless of the histologically proven sampling area, namely, central tumor core or tumor edge) of 83% [8,9]; by using unprocessed samples of pediatric brain tumors, even an accuracy of 86.2% was achieved when classifying between low-grade gliomas and normal brain by employing a logistic regression model [10]. Although, commonly, fresh tissue specimens are spectroscopically examined intraoperatively in vivo [11] by using a handheld Raman probe or ex vivo [4,12] by using advanced Raman imaging techniques, e.g., Stimulated Raman Histology, other approaches aim to establish RS on formalin-fixed or formalin-fixed and paraffin-embedded (FFPE) tissue.…”