Background Effective management of patients with infected wounds is a crucial concern. A delay in prescribing the appropriate antibacterial agent can lead to life-threatening clinical complications. Thus, the electronic nose technique (eNose) can provide a diagnostic aid tool that allows rapid and accurate identification of pathogens. Results This study examines the effectiveness of using eNoses to aid in the diagnosis of bacterially infected wounds. The systematic search in the literature

more »

... eved 3,326 publications, of which 97 were for a complete review, and of these, 09 comprised the sample of this study. These studies involved the analysis of seven types of wounds, the most common being the infected skin wound. The most frequent bacteria were P. aeruginosa, E. coli and methicillin-susceptible Staphylococcus aureus (MSSA). The average accuracy of the eNoses in identifying these microorganisms was 95.13% for the training set and 91.5% for the test set, including the ability to differentiate between bacteria of the same genus but sensitive or resistant to antibiotics. Among the Artificial Intelligence techniques used to classify the models, the Support Vector Machine (SVM) was the most commonly used in the experiments. Conclusion The eNoses devices observed may have broad applicability in aiding diagnosis of wound infection through their high efficacy values. However, further research needs to explore the reduction of interferences in the accuracy of the application of Machine Learning algorithms.