Noninvasive SARS-CoV-2 detection using a low-cost electronic nose.

Area of Science:

• Biomedical Engineering
• Infectious Disease Diagnostics
• Sensor Technology

Background:

• The COVID-19 pandemic underscored the need for improved SARS-CoV-2 detection methods.
• Existing diagnostic tests can be invasive, slow, or lack accuracy in early stages.
• Noninvasive, rapid, and accurate point-of-care diagnostics are crucial for pandemic response.

Purpose of the Study:

• To develop and evaluate a noninvasive electronic nose for rapid SARS-CoV-2 detection.
• To assess the performance of volatile organic compound (VOC) analysis in saliva and exhaled breath for identifying infection.
• To compare the efficacy of various machine learning algorithms for classifying SARS-CoV-2 positive and negative samples.

Main Methods:

• An electronic nose composed of an array of metal-oxide gas sensors was utilized.
• Volatile organic compounds (VOCs) from saliva and exhaled breath samples were analyzed.
• Machine learning algorithms, including KNN, SVM, NN, and Random Forest, were employed for sample classification.

Main Results:

• The electronic nose achieved accuracy rates of 76%-89% for exhaled breath and 75%-86% for saliva samples.
• The K-Nearest Neighbors (KNN) algorithm demonstrated optimal performance.
• KNN yielded an Area Under the Curve (AUC) of 0.895 for exhaled breath and 0.861 for saliva.

Conclusions:

• The study demonstrates the feasibility of using a low-cost electronic nose for SARS-CoV-2 detection.
• This noninvasive approach shows potential for rapid, accurate point-of-care diagnostics.
• Further validation in real-world clinical settings is supported by these proof-of-concept results.