Rapid Identification of Candida auris by Raman Spectroscopy Combined With Deep Learning

Abstract

Candida auris is a multidrug-resistant yeast that can lead to outbreaks in healthcare facilities, even with strict infection prevention and control measures. Candida auris detection is challenging using standard laboratory methods. Advancements in identification methods, such as matrix-assisted laser desorption/ionization time-of-flight mass spectrometry and polymerase chain reaction, have improved detection, though these methodologies can be costly and impractical in resource-limited settings. This study presents a practical, portable, and reagentless platform known as Counter-Propagating Gaussian Beam Raman Spectroscopy (CPGB-RS), integrated with deep learning spectral analysis for the rapid and accurate identification of C. auris. This method has shown a sensitivity of 96% and a specificity of 99% in differentiating C. auris from other highly prevalent pathogenic species, such as Candida albicans, Candida glabrata, and Candida tropicalis. The differentiation between species is based on unique variations in their Raman spectra, influenced by differences in cell wall composition (including β-glucan, chitin, and mannoprotein), cell membrane components (like ergosterol), and cellular energy states (mitochondrial cytochromes b and c). This platform allows for automated molecular screening, generating diagnostic results within 2 min, making it highly practical for clinical applications. Furthermore, this technology has the potential to evaluate the effectiveness of antifungal agents, which could significantly improve patient outcomes.