Objective: This study aimed to highlight that colposcopic vaginal patterns are not specific, unlike cervical colposcopic patterns, and to provide a simpler classification of vaginal colposcopic patterns.
Materials and methods: A total of 223 patients who underwent colposcopy with Schiller test were assessed (hierarchical log-linear model) retrospectively.
Results: The greatest predictability for histologically confirmed warts and cancers is represented by colposcopic patterns of wart and cancer. Lugol-negative area is strongly predictive of koilocytosis, even if it is found in other vaginal lesions. Thickened epithelium seems to better predict a severe vaginal lesion, whereas thin white epithelium better suggests a mild vaginal lesion. Colposcopic patterns were simplified as follows: Lugol-negative area, white epithelia (thin white epithelium and white thickened epithelium), vascular lesions (regular and irregular mosaicisms and punctations), wart, and cancer. Thus, koilocytosis is predicted by the Lugol-negative area, whereas white epithelia patterns and vascular patterns are not specific, suggesting overall vaginal intraepithelial neoplasias. Wart and cancer patterns are pathognomonic for histologically confirmed warts and cancers.
Conclusions: Vaginal colposcopy poorly predicts the severity of vaginal lesions. By including each type of white epithelium within a new category called "white epithelia patterns" and each type of vascular pattern within a new category called "vascular patterns," it is possible to simplify vaginal colposcopy without compromising its accuracy.