Application of Raman spectroscopy and multivariate analysis to detect ionizing radiation-induced changes in blood plasma

Abstract

The risk of large-scale radiological/nuclear events has notably increased in recent years. Biodosimetry is considered an essential tool for emergency management following such unplanned exposures to ionizing radiation. For example, by assessing an individual's received dose to blood, biodosimetry can support medical screening and individual health management. Current biodosimetry techniques, such as the dicentric chromosome assay, are based on the analysis of chromosomal aberrations Although highly accurate, these methods are time-consuming and labour-intensive. We recently developed a new high-throughput approach based on Raman spectroscopy of blood combined with covariate-adjusted multivariate analysis for the detection of irradiated blood. We found that the protein bands in the Raman spectra were the main sources of discrimination between unirradiated (control) and irradiated blood. In this follow up work, we explored the application of Raman spectroscopy and multivariate analysis to blood plasma to avoid dominant hemoglobin contributions. Peripheral blood drawn from a healthy volunteer was irradiated at 0 (control), 5 and 20 Gy using 250 kV Xrays. After a 4 hour incubation time, plasma centrifuged from the blood sample was immediately frozen at -80 deg C. Raman measurements were performed in triplicate on thawed blood plasma samples. Partial least squares-discriminant analysis (PLS-DA) was utilized for multi-class differentiation between Raman spectra of 0, 5 and 20 Gy irradiated plasma. Sparse PLS-DA (sPLS-DA) provided improved dose classification after combining Raman spectral data from different batches. Biomarker information related to radiation-induced changes in blood plasma was also extracted from sPLS-DA. The outcomes of these initial studies highlight the value of Raman spectroscopy to support biodosimetry.