Paracoccidioidomycosis screening diagnosis by FTIR spectroscopy and multivariate analysis

Area of Science:

• Mycology
• Infectious Diseases
• Biotechnology
• Spectroscopy

Background:

• Paracoccidioidomycosis (PCM) is a significant systemic fungal infection endemic to Latin America, particularly Brazil, where it ranks as a major cause of mortality from chronic infectious diseases.
• Current diagnostic methods for PCM, including microbiological, immunological, and histopathological techniques, often rely on direct fungal observation, which is considered the gold standard but can be time-consuming.
• There is a need for rapid, accurate, and accessible diagnostic tools to improve patient outcomes and disease management for Paracoccidioidomycosis.

Purpose of the Study:

• To evaluate the efficacy of Fourier Transform Infrared (FTIR) spectroscopy coupled with machine learning as a novel diagnostic strategy for Paracoccidioidomycosis (PCM).
• To assess the ability of FTIR-based analysis of blood serum to differentiate between individuals with PCM and healthy controls.
• To determine the diagnostic accuracy of machine learning algorithms applied to FTIR spectral data for PCM detection.

Main Methods:

• Blood serum samples were collected from 20 patients diagnosed with Paracoccidioidomycosis (PCM) and 20 healthy individuals.
• Fourier Transform Infrared (FTIR) spectra were acquired from the blood serum samples.
• Principal Component Analysis (PCA) was employed for data dimensionality reduction, followed by the application of a Cubic Support Vector Machine (Cubic SVM) machine learning algorithm for classification.

Main Results:

• The FTIR spectra analysis, processed using PCA and classified with Cubic SVM, demonstrated a high diagnostic potential for Paracoccidioidomycosis (PCM).
• The combined approach of FTIR spectroscopy and machine learning achieved an overall accuracy of 91.67% in distinguishing PCM patients from healthy individuals.
• This study highlights the effectiveness of spectral data analysis for identifying biomarkers associated with PCM in blood serum.

Conclusions:

• Fourier Transform Infrared (FTIR) spectroscopy, when integrated with machine learning algorithms like Cubic SVM, presents a promising, non-invasive, and rapid method for the diagnosis of Paracoccidioidomycosis (PCM).
• This innovative strategy offers a high degree of accuracy and could serve as a valuable complementary tool to existing diagnostic methods for PCM.
• Further research and validation are warranted to establish FTIR-based spectral analysis as a routine diagnostic technique for PCM in endemic regions.