Analyzing the blood compositions of different stages of obesity disease using a linear-regression model between FT-Raman spectroscopy and BMI data

Abstract

This study utilized FT-Raman spectroscopy, linear regression models, principal component analysis (PCA), and correlation to explore the link between BMI and serum chemical structure across varying obesity stages. Biochemical, hematological, and immune-inflammatory markers were analyzed in 80 patients across three groups (overweight, obese, and morbidly obese) and healthy subjects (n = 45). Only BMI significantly varied among groups, with morbidly obese patients showing higher Cystatin c levels. In the morbidly obese patient group, a correlation was observed between BMI and 5 peaks (1270 cm−1, 1344 cm−1, 1459 cm−1, 1671 cm−1, and 2922 cm−1). FT-Raman data indicated lower amide, amino acid, and lipid levels in overweight individuals. FT-Raman data were found to correlate with BMI by more than 50 %, especially in patients with the highest degree of obesity, notably 1606 cm−1 (a 53 % fit) and 2933 cm−1 (67 % fit) suggesting strong correlations between FT-Raman data and BMI, particularly in severe obesity.