Broadband IR-Fingerprinting of Human Blood as a Universal Tool for Diseases Diagnostics

Summary form only given. Many diseases cause characteristic changes in the molecular composition of biofluids such as human blood. Thus, a sufficiently sensitive and specific blood analysis could be used for disease detection. Particularly, physiological phenotypes (health as well as disease states) are reflected by minor changes in the concentration of several, possibly thousands, of different molecules in blood which cover a wide concentration dynamic range [1]. An approach that could thus quantitatively detect different molecular groups of blood simultaneously (such as e.g. proteins, metabolites, carbohydrates) would be generally very advantageous. Fourier -Transform Infrared Spectroscopy (FTIR) is one of the most versatile tools for analysis of molecularly complex, heterogeneous samples. Blood-based application of FTIR may in principle provide an efficient, less invasive, time- and cost-effective detection technique, contributing to the detection of a variety of physiological conditions. Although promising, FTIR spectroscopy has previously not been used for the analysis of blood-based samples in populational studies, nor has it been evaluated in a setting with well -powered groups of several hundreds inviduals to present a true cross-section of a given population. Towards that end we apply and evaluate the feasibility of FTIR molecular fingerprinting for populational disease screening in general. We tackle infrared molecular fingerprints in the Germanpopulation-based cohort study KORA (Cooperative Health Research in the Augsburg Region) [2]. In the KORA-FF4 study in 2013/2014 a standardized medical examination and health status interview was conducted with 2279 participants aged 39-88. The aim is to address the general paradigm: whether the biological variability of FTIR molecular fingerprints in a given adult human population facilitates efficient detection of most prevalent disease phenotypes (e.g. diabetes, cardiovascular disease). Moreover, the goal of the study is to evaluate whether blood-based FTIR molecular fingerprints could facilitate isolation of infrared spectral fingerprint markers that could be applied for unambiguous disease detection and distinction between different diseases using machine learning algorithms. Preliminary analyses suggest the feasibility of infrared molecular fingerprinting approach for phenotype detection albeit inherent biological variability of human population and provide the first overview of most common disease -specific infrared signatures.