Lyophilized nasal swabs for COVID-19 detection by ATR-FTIR spectroscopy: Machine learning-based approach

Area of Science:

• Biomedical Diagnostics
• Spectroscopy
• Infectious Disease Research

Background:

• The COVID-19 pandemic necessitates advanced diagnostic tools beyond standard PCR.
• Spectroscopy-based diagnostics present a promising alternative for rapid and accessible disease detection.
• Lyophilized nasal swabs offer a stable and suitable matrix for spectroscopic analysis.

Purpose of the Study:

• To evaluate the efficacy of Attenuated Total Reflectance-Fourier Transform Infrared (ATR-FTIR) spectroscopy in distinguishing between COVID-19 positive and negative individuals using lyophilized nasal swab samples.
• To identify spectral biomarkers indicative of COVID-19 infection in nasal swab samples.
• To assess the diagnostic accuracy of statistical models applied to spectroscopic data.

Main Methods:

• Analysis of 55 lyophilized COVID-19 positive and 55 negative nasal swab samples using ATR-FTIR spectroscopy.
• Spectral data pre-processing and analysis in the 600-4000 cm⁻¹ range.
• Application of statistical methods including Principal Component Analysis (PCA), Linear Discriminant Analysis (LDA), and Support Vector Machine (SVM) for classification.

Main Results:

• Significant spectral variations were observed between healthy and infected samples.
• Distinct spectral peaks were identified, with 13 observed in COVID-19 samples and 4 in negative samples.
• LDA and SVM models achieved 100% classification accuracy, demonstrating high diagnostic performance with both internal and external validation.

Conclusions:

• Lyophilized nasal swab samples are optimal for FTIR-based COVID-19 diagnostics.
• ATR-FTIR spectroscopy, combined with statistical analysis, provides a robust, accessible, and highly accurate method for COVID-19 detection.
• This approach can serve as a valuable tool for post-pandemic surveillance and rapid diagnostics.