Detecting Bacteria on Wounds with Hyperspectral Imaging in Fluorescence Mode

Abstract

Chronic non-healing wounds represent an increasing problem. In order to enable physicians and nurses to make evidence based decisions on wound treatment, the professional societies call for supporting tools to be offered to physicians. Oxygen supply, bacteria colonization and other parameters influence the healing process. So far, these parameters cannot be monitored in an objective and routinely manner. Existing methods like the microbiological analysis of wound swabs, mean a great deal of effort and partly a long delay. In this paper 42 fluorescence images from 42 patients with diabetic foot ulcer, recorded with a hyperspectral imaging system (TIVITA®), converted for fluorescence imaging, were analysed. Beside the fluorescence images, information about the bacterial colonization is available from microbiological analysis of wound swabs. After preprocessing, principal component analysis, PCA, is used for data analysis with a 405 nm excitation wavelength, the emission wavelength range 510 - 745 nm is used for analysis. After dividing the data into a training and a test dataset it could be shown, that bacteria are detectable in the wound area. A quantification in bacterial colonization counts (BCC) was not in the focus of the research in this study stage.

Gas Plasma Exposure Alters Microcirculation and Inflammation during Wound Healing in a Diabetic Mouse Model

“…As shown in bead-based cytokine and chemokine profiling, the repeated, gas plasma-induced inflammatory response is differentially regulated on a systemic level in both genders. By analyzing the spectral signatures of bacteria or microbial metabolites, HSI can identify the presence of infection and differentiate between infected and non-infected wounds [63]. Overall, hyperspectral imaging offers a non-invasive and objective approach to assess wound healing processes in diabetic wound healing in mouse models.…”

Hyperspectral imaging for early diagnosis of diseases: A review

“…Hornberger et al, 2020 propose a method of bacteria detection on wounds (diabetic foot syndrome [DFS]) by analysing fluorescence images using PCA. Blanco et al, 2012 propose using the NIR hyperspectral imaging technique for the classification of kidney stones in a faster and more robust way for the urologist for rapid diagnosis using deep learning.…”

“…Halicek et al, 2017 and Akbari et al, 2012 use data normalization of HS images of head and neck and prostate cancer tissues. While Zherdeva et al, 2016 make use of a single‐scale retinex algorithm for normalization and Hornberger et al, 2020, Collins et al, 2021, and Baltussen et al, 2019 utilize the SNV or standard normal variate normalization for supervised ML‐based studies.…”

“…Noise removal from hyperspectral images can be done using moving average filter(M. H. Aref et al, 2020), inverse minimum noise fraction (Calin & Parasca, 2022) and 2nd and 1st derivative computation (Hornberger et al, 2020), (Hu et al, 2019). Generally, de‐noising is done simultaneously with normalization using white and dark references.…”