Collaborative detection for wound infections using electronic nose and FAIMS technology based on a rat wound model

•An investigation path combining two gas sensing methods-e-nose and FAIMS was proposed for the detection of wound infection.•Full-thickness skin defect model was applied on SD rats to acquire reliable odor sensing data for wound infection.•An effective information fusion framework was designed, wher...

Full description

Saved in:
Bibliographic Details
Published in:Sensors and actuators. B, Chemical Chemical, 2020-10, Vol.320, p.128595, Article 128595
Main Authors: Sun, Tong, He, Jiao, Qian, Shenyi, Zheng, Yangting, Zhang, Kun, Luo, Jing, Tian, Fengchun
Format: Article
Language:English
Subjects:
Algorithms
Bacteria
Bacterial infections
Collaboration
Data integration
Electronic nose
Electronic noses
FAIMS
Infections
Information fusion
Inoculation
Ionic mobility
Odor sensing
Wound infection detection
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:•An investigation path combining two gas sensing methods-e-nose and FAIMS was proposed for the detection of wound infection.•Full-thickness skin defect model was applied on SD rats to acquire reliable odor sensing data for wound infection.•An effective information fusion framework was designed, wherein a model fusion method named CMEP-MRE was proposed. For years, electronic nose has been extensively studied for identifying bacteria or wound bacterial infections, which was commonly based on in vitro experiments. In this paper, based on rat in vivo experiments, we integrated two odor-sensing technologies — electronic nose and FAIMS (Field Asymmetric Ion Mobility Spectrometry) — to discriminate three common wound bacterial infections and non-infection by directly sniffing the rat wound samples. With a set of dedicated information fusion algorithm, we effectively fused the information acquired by the two technologies and obtained high rat recognition rates under real independent leave-one-out cross-validation. For the wounds 24 h/48 h after inoculation, the collaborative detection of the two technologies achieved 94.63 %/88.05 % mean accuracy, 96.67 %/98.75 % sensitivity, 90.00 %/88.89 % specificity, 94.78 %/87.62 % macro F1 score, and 94.59 %/87.28 % micro F1 score, which showed better performance than either of the two technologies. The results demonstrated the potential of odor-sensing technology as a convenient screening and monitoring method for wound infections.
ISSN:0925-4005
1873-3077
DOI:10.1016/j.snb.2020.128595