DETECTION OF DIABETES MELLITUS TYPE II BY INFRARED SPECTROSCOPY AND MULTIVARIATE ANALYSIS

• Main Authors: M. V. Uspenskaya, T. N. Nosenko, V. E. Sitnikova, R. O. Olekhnovich
• Format: Article
• Language: English
• Published: Saint Petersburg National Research University of Information Technologies, Mechanics and Optics (ITMO University) 2019-03-01
• Series: Naučno-tehničeskij Vestnik Informacionnyh Tehnologij, Mehaniki i Optiki
• Subjects: diabetes mellitus; IR spectroscopy; multivariate analysis; serum; cluster analysis; PCA; PLS
• Online Access: https://ntv.ifmo.ru/file/article/18589.pdf

The paper presents detection method for one of the most common diseases. Diabetes mellitus is a group of metabolic disorders in which blood sugar level rises over a long period. Type II diabetes begins with insulin resistance, a condition when cells do not respond properly to insulin. Type II diabetes accounts for about 90 % of all diagnosed diabetes cases. Therefore, the development of new methods for diagnosing is relevant. We applied the method of IR-spectroscopy combined with multivariate analysis for blood serum analysis. The research is based on serum samples from 44 patients with diabetes and an observational group of 30 healthy donors. Cluster analysis, projection on latent structures, principal component analysis is used to differentiate groups of healthy donors and patients with diabetes. Each of the three presented methods gives the best results when the first derivative of the spectra is considered in the whole range of the spectrum. Cluster analysis, according to the procedure performed, gives the sensitivity of 100% and the specificity of 95.8 % for the considered sample. Principal Component Analysis provides a visual separation of the sample into groups of healthy donors and patients with diabetes mellitus. When analyzing this sample, using the procedure, the sensitivity was 100 % and the specificity of the analysis was 96.6 %.