Using infrared spectroscopy of serum and chemometrics for diagnosis of paracoccidioidomycosis

• Autor(a) principal: Koehler, Alessandra
• Data de Publicação: 2022
• Outros Autores: Scroferneker, Maria Lúcia, Pereira, Beatriz Aparecida Soares [UNESP], Pereira de Souza, Nikolas Mateus, de Souza Cavalcante, Ricardo [UNESP], Mendes, Rinaldo Pôncio [UNESP], Corbellini, Valeriano Antonio
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Institucional da UNESP
• Texto Completo: http://dx.doi.org/10.1016/j.jpba.2022.115021; http://hdl.handle.net/11449/249140
• topic: Fourier transform infrared spectroscopy; Multivariate analysis; Orthogonal partial least squares discriminant analysis; Paracoccidioidomycosis; Photodiagnosis; Systemic mycosis

Paracoccidioidomycosis (PCM) is a systemic granulomatous mycosis endemic to Latin America, whose etiologic agents are fungi of the genus Paracoccidioides. PCM is usually diagnosed by microscopic observation of the fungus in biological samples, combined or not with other techniques such as serological methods. However, all currently used diagnostic methods have limitations. The objective of this study was to develop a method based on Fourier transform infrared spectroscopy (FTIR) and chemometric analysis for PCM diagnosis. We included 224 serum samples: 132 PCM sera, 24 aspergillosis sera, 10 cryptococcosis sera, 8 histoplasmosis sera, and 50 sera from healthy blood donors. Samples were analyzed by attenuated total reflection (ATR), and chemometric analyses including exploratory analysis through principal component analysis (PCA) and a classification method (PCM and non-PCM) through orthogonal partial least squares discriminant analysis (OPLS-DA). The spectra were similar, with the main bands up to approximately 1652 cm–1 and 1543 cm–1 (amide I and amide II bands). This same region was mainly responsible for the partial separation of the samples in PCA. The OPLS-DA model correctly classified all serum samples with only one latent variable, with a determination coefficient (R²) higher than 0.999 for both the calibration set and prediction set. Sensitivity and specificity were 100% for both sets, showing better performance than the reference diagnostic methods. Therefore, the use of FTIR/ATR together with OPLS-DA modeling proved to be a promising method for PCM diagnosis.