Hyperspectral Fluorescence and Reflectance Imaging of Bacteria and Oral Mucosa - a feasibility study

Abstract

Dental health is an area with many challenges, in which non-invasive techniques for diagnostics is of great interest. Spatially resolved spectroscopy using hyperspectral imaging may provide such a technique. This thesis focuses on two problems within the oral cavity. The first part is an evaluation of the feasibility of bacterial characterisation and differentiation using fluorescence and reflectance hyperspectral imaging for characterisation of periodontitis. The second part evaluates the hyperspectral images from a study of effects of an adhesive dental material in an animal model. The adhesive is used for tooth fillings e.g. for treatment of caries attacks.

Hyperspectral fluorescence and reflectance images were obtained of Staphylococcus epidermidis, Streptococcus mutans, Enterococcus faecalis, Pseudomonas aeruginosa, Escherichia coli strains ATCC 700926, O103:H2 and BW25113, Klebsiella pneumoniae and Staphylococcus aureus for a proof-of-concept study on the general characterisation ad differentiation of bacteria. The fluorescence images were obtained using an excitation wavelength of 365 nm. Similar images were also recorded and analysed of the fungus C. albicans, known to reside in the oral cavity. As fluorescence images have already proven to be promising for differentiating bacteria, a method for characterisation of fluorescent spots in bacteria is suggested, using a combination of RX anomaly detection together with Sequential Maximum Angle Convex Cone (SMACC) and Spectral Angle Mapping (SAM). This technique seems be successful in extracting spectra of anomalous appearance in the bacterial colonies, and some of the peaks from these spectra were matched with emission maxima of common porphyrins. A satisfactorily differentiation of the bacteria was not possible using the reflectance data set. However, a repetition of the imaging is required as an artefact from the camera was found in the spectral region 665-755 nm.

The second experiment tested whether physiological and immunological changes could be detected after exposure to adhesive, adhesive and blue light, blue light and included a control group. Images were obtained both in the VNIR range, 400-1000 nm and the SWIR, 1-2.5 um. Oxygenation was calculated using the images in the visible range for both for a shorter penetration depth and a larger penetration depth. An estimate of the relative change in hydration was found using the images in near-infrared, utilising the water absorption peaks at 1200 nm and 1450 nm. No statistical significance was obtained for the change in either parameter using Kruskal-Wallis testing.