Digital analysis of colposcopic images with acetowhite epithelium in "real-time" mode

Abstract

Digital analysis is a method of assessing qualitative data by converting them into quantitative data using a personal computer’s graphics processing unit (GPU). Visual changes on the cervical surface in human papillomavirus (HPV)-associated pathology can be evaluated by extended colposcopy with a 3–5% acetic acid test and Schiller’s test. Acetowhite epithelium (appears after a test with acetic acid) in the form of flat condyloma is one of the main clinical manifestations of HPV infection (excluding manifestations on the metaplastic epithelium). The intensity of acetowhite epithelium (AWE) is a subjective parameter during gynecological examination, and to exclude the malignant process, a histology or liquid-based cytology should be performed. Objective assessment of the AWE intensity is introduced by digital image analysis using specialized software. The aim. To improve the assessment of visual changes of acetowhite epithelium on the cervical surface using a digital approach as an additional tool when performing extended colposcopy. Materials and methods. Colposcopy images of 359 patients were estimated by digital approach. Colposcopy images were made with a screenshot function after performing extended colposcopy. Screenshots of the cervix were converted to 8-bit type. This allowed us to estimate the intensity of AWE by software ImageJ. Index grey value (IndGV) was calculated to estimate the intensity of AWE. The IndGV data were assessed by MS office 365 (Excel v16.46) and RStudio v1.2.5042 statistical packages. Results and conclusion. The quantification approach detects the severity of AWE on the cervical surface. IndGV correlates with the grades of cervical intraepithelial neoplasia (CIN) and clinical manifestations of HPV. The clinical manifestations of HPV infection in CIN have a direct dependence on digital measurement as grey value (GV). These indicators give the opportunity to evaluate clinical manifestations in digital analysis in real-time mode, which does not require a significant financial investment. IndGV can be used as a prognostic factor of the severity of HPV infection and its clinical forms (CIN I–III and cervical cancer). Key words: digital medicine, quantitative approach, acetowhite epithelium, HPV infection