Iron molecule is of great importance in the synthesis of hemoglobin which is essential for oxygen transport. Iron levels are quantified by accurately high sensitivity tests, such as serum ferritin (SF). However, common studies to quantify SF are long and strenuous (~ 5 h), for example enzyme-linked immunosorbent assay (ELISA). In this paper, blood serum samples were analyzed by Raman spectroscopy (RS), and a computational analysis of spectra is proposed to detect differences in SF as an alternative procedure. Serum samples were obtained from 22 patients, 9 who were clinically diagnosed with anemia and 13 controls. Patients with anemia had low levels of SF (< 30 ng/ml), and a control group had levels between 30 and 500 ng/ml. The spectra obtained were conditioned with a baseline correction and smoothing, then evaluated by principal component analysis (PCA), and a predictive model was estimated by lineal discrimination analysis (LDA). The results showed a clear differentiation of the study groups by PCA, also 99.69% sensitivity and 100% specificity by LDA. This study suggest that Raman spectroscopy is a fast (~ 5 min) and a powerful tool capable to qualitative differentiate ferritin concentrations.
Keywords: Iron levels; Principal component analysis; Raman spectroscopy; Serum ferritin.