208 Performance Assessment of Electronic Nose Device for Detection of COVID-19 in Breath Samples

Abstract

Novel methods of rapid and large-scale testing for COVID-19 infection are necessary during the ongoing pandemic. Although common test samples are nasopharyngeal and oral specimens, breath analysis has also been proposed for COVID-19 detection. Our study objective was to evaluate the performance of the electronic nose (eNOSE), a portable breath testing device developed by the National Aeronautics and Space Administration (NASA) for the diagnosis of COVID-19 infection. The eNOSE sensor is approximately 6 inches by 3 inches by 3 inches and analyzes breath samples to detect combinations of volatile organic compounds (VOCs) diagnostic of SARS-CoV-2 virus infection. Following Institutional Review Board and Biosafety Committee approvals, we recruited adults with a previous positive COVID-19 nasopharyngeal reverse transcription polymerase chain reaction (RT-PCR) test and volunteers with unknown infection status. During the study visit, we collected breath samples and introduced it to the eNOSE device for VOC sensing. Concurrently, participants provided anterior nares samples for RT-PCR testing for COVID-19. Cases were those who had a positive RT-PCR and the controls were those who had a negative RT-PCR at the time of breath sample collection. The sensitivity and specificity of the eNOSE device were calculated using the concurrent RT-PCR test as the gold standard. There were 64 participants enrolled, with a mean age of 42 years (SD +13 years) and of whom 44 (69%) were female. We recruited 54 previously COVID-19 positive participants and 10 participants with unknown infection status. At the time of breath collection, there were 32 RT-PCR positives, 31 RT-PCR negatives, and one untestable sample. Of those with a RT-PCR positive result, 21 had a positive eNOSE result and 11 had a negative eNOSE result. Of those with RT-PCR negative result, 21 had a negative eNOSE result and 10 had a positive eNOSE result. The eNOSE device was 66% sensitive and 68% specific for the detection of COVID- 19 infection. The mean time from recruitment to enrollment was 7 days (SD +7 days). From the limited data set collected to date, the eNOSE device had moderate sensitivity and specificity for the diagnosis of COVID-19 infection. Both parameters are expected to improve as more samples are analyzed. Our next step is to include cycle threshold (Ct) values in the analysis to learn if the eNOSE response is correlated with vira