Hyperspectral imaging method for noninvasive quantitative detection of vulvar lichen sclerosis: a pilot study

Abstract

Vulvar lichen sclerosis (VLS) is a chronic mucocutaneous disorder, the pathogenesis of which is poorly understood. Few methods are available for noninvasive quantitative detection of VLS. Here, we proposed a hyper-spectral imaging method for in-vivo quantitative detection of VLS. An IRB approved clinical trial was carried out on three VLS patients and one patient with normal vulva as a control group. We acquired a hyper-spectral data cube of each patient with wavelengths ranging from 450 to 650 nm at an internal of 2 nm, followed by quantitative analysis of the spectral contribution of three known ingredients in skin tissues: Oxy-hemoglobin(HbO2), deoxy-hemoglobin(Hb) and melanin using an Independent Component Correlation Algorithm (ICA) method. Our analysis results suggested that the proportion of melanin and the total hemoglobin in VLS subjects were different from normal subjects. The segmentation of VLS regions based on the index of melanin and total hemoglobin are close to the VLS regions marked by a clinician. Our pilot study has firstly found the absorption spectra difference between normal vulva and VLS and demonstrated the technical potential of using this hyper-spectral imaging method for in-vivo quantitative detection of VLS.