Hyperspectral imaging as a new diagnostic tool for cervical intraepithelial neoplasia

Abstract

PurposeCervical cancer screening by visual inspection with acetic acid (VIA) during colposcopy can be challenging and is highly dependent on the clinical experience of the examiner. Health-care systems lack qualified physicians able to perform the examination in both industrialized and low- and middle-income countries. Previous work has shown the general potential of hyperspectral imaging (HSI) to discriminate CIN from normal tissue, but clinical translation has been limited due to the lack of medically approved HSI systems.MethodsIn this study, we evaluate the feasibility of a commercially available HSI system for CIN detection in a prospective monocentric clinical trial.ResultsBy obtaining spectral fingerprints of 41 patients with CIN 1–3 we show that HSI-based differentiation between CIN and normal tissue is possible with high statistical significance. Major spectral differences were seen in the 555–585 wavelength area.ConclusionHSI advances tissue differentiation by associating each pixel with high-dimensional spectra and thereby obtains morphological and biochemical information of the observed tissue. Currently available and medically approved HSI systems may represent a contact- and marker-free examiner-independent method for the diagnosis of CIN.