Raman spectroscopy of blood serum and cerebrospinal fluid and multivariate data analysis for Alzheimer's disease diagnostics

The efficient and accurate diagnosis at the early stages of dementia is a key moment for effective treatment and productive research to find a new ways to combat the disease. It is especially true for Alzheimer's disease (AD) for which there is no effective cure, but several treatments are known to allow slowing down the degenerative processes. Alzheimer's disease (AD) displays only non-specific clinical symptoms of mental decline for decades after the initiation and is very challenging to differentiate even at the later stages when it becomes very aggressive. Despite the great need, current diagnostic tests are unable to diagnose different forms of dementia and other neurodegenerative diseases implicitly.;This work explored application of Raman spectroscopy for detecting changes in the biochemical composition of blood serum and CSF. Our approach is based on probing a total biochemical composition of a bio-fluid using a near-infrared (NIR) Raman spectroscopy combined with advanced statistics. High differentiation power of the method can promote the translation of Raman spectroscopy from a laboratory technique into a clinically useful tool. The focus of this study was to demonstrate that the biochemical information derived by Raman spectroscopy from cerebrospinal fluid (CSF) and human blood serum yields accurate and selective AD detection.;Combination of Raman spectroscopy with advanced multivariate statistics allowed differentiating CSF and serum samples of AD patients from normal aging samples with sensitivity and specificity close to 95%. Semi-supervised machine learning algorithms, such as PCA, GA, SVMDA and ANN, were used to analyze Raman spectra acquired from different categories of samples representing patients and healthy controls.;The outcome of this research is the demonstration of accurate discriminative method based on Raman spectroscopy for detection of AD. The findings confirm the hypothesis that the biomedical application of the method allows to distinguish AD serum and CSF. The experiments and theory presented throughout this work demonstrate that Raman spectroscopy is a viable clinical tool that can be used for an accurate diagnosis AD via a simple test of serum or CSF.