WoundSentry: A Revolutionary AI-Driven Wearable for Enhanced Infection Detection

T.J. Ko, L. Yeh, P. Lubet, J. Wang, A. Gnatt, X.L. Liu
United States

Keywords: wound infection, pathogen, VOCs, sensor, AI


Wound infections represent a critical aspect of healthcare-associated infections, significantly impacting patient recovery and increasing healthcare costs globally. These infections are particularly challenging in wound care, where timely and accurate management is crucial to prevent complications and ensure effective healing. In hospitals and healthcare settings, wound infections form a substantial part of the broader spectrum of healthcare-associated infections, necessitating innovative and adaptable solutions for effective management. While the overall burden of hospital-associated infections is significant, with millions occurring annually in the United States alone, the subset of wound-related infections is particularly pressing due to their direct impact on patient healing and recovery. WoundSentry, with its advanced real-time pathogen detection technology, addresses this urgent need, providing a valuable tool for patients in hospitals, long-term care facilities, and home care environments. WoundSentryTM, a novel wearable device, addresses this need by utilizing an AI-driven gas sensor array for real-time pathogen detection and infection prediction. This user-friendly, cost-effective technology enhances care and decision-making from the point of injury through the recovery process. The WoundSentry device is equipped with a nanocomposite sensor array composed of 6-12 individual sensors. The sensors are designed to efficiently adsorb volatile organic compounds from microbial, leading to a change in their electrical characteristics after chemical binding. In vitro assays were performed to evaluate the performance of the WoundSentry device on both bacterial and fungal pathogens grown on standard agar plates. An 8-class Support Vector Machine (SVM) classifier was utilized to analyze the sensor profiles. The data set was randomly split into a training set and a test set to train and validate the AI-powered pattern recognition system. The results of the study showed a high average accuracy of 97% in detecting a diverse range of microbial gases produced by bacterial and fungal pathogens. WoundSentry was able to differentiate between individual bacterial strains, mixed cultures, and empty plates within the first two hours with only a few clusters partially overlapping. The experiment on pig skin tissue showed that the sensor array was capable of effectively distinguishing between pig skin tissue samples with and without bacteria. Strong signals were recorded for pig skin tissue inoculated with Pseudomonas aeruginosa, Staphylococcus aureus, and their mixture. Also, the signal became intense as the cell density increased, suggesting WoundSentry can identify not only pathogen species but also the number of cell-based progress. These findings not only affirm WoundSentry's potential in revolutionizing wound management but also suggest broader applications in pathogen identification and antibiotic susceptibility testing in clinical laboratories. WoundSentry's innovative approach promises to enhance healthcare outcomes, particularly in challenging healthcare environments. ACKNOWLEDGEMENT CITATION: This research and development project was conducted by NanoBioFAB and is made by a contract that was awarded and administered by the U.S. Army Medical Research & Development Command (USAMRDC) and the SBIR Office - at Fort Detrick, MD under Contract Number: (DHA SBIR W81XWH22P132 and DHA SBIR W81XWH21C0087)