Detection and visualization of intracutaneous allergic type-specific elements using long-wavelength near-infrared hyperspectral imaging
Department of Computer Science and Engineering
Toyohashi University of Technology
This study aimed to develop a method for the assessment of allergic dermatitis by using the long-wavelength near-infrared spectrum (more than 1000 nm) to detect intracutaneous allergic type-specific elements. Such a method was realized by establishing a spectral classifier for the spectra of type I and type IV allergic dermatitis reactions.
Near-infrared spectral images of histamine-induced cutaneous reaction (type I) and contact hypersensitivity erythema elicited by squaric acid dibutylester (SADBE; type IV) were obtained, and the absorption spectra of normal and inflamed skin were extracted from these spectral images. A spectral classifier was established from these training datasets, and it was then applied to two test cases, red flare by methyl nicotinate (normal) and metal allergy (type IV).
The spectral classifier established by canonical discriminant analysis (CDA) achieved very accurate detection (normal: 87.67%, type I: 87.00%, type IV: 98.5%). Furthermore, the test cases were also correctly classified: the red flare induced by methyl nicotinate was categorized as normal skin and the metal allergy was categorized as a type IV allergic reaction.
These results suggest a possible application of near-infrared spectral imaging to the assessment of allergic dermatitis.