Macroscopic prediction of cervical dysplasia at colposcopy for see-and-treat patient selection: a systematic review

In see-and-treat protocols, women referred to colposcopy are treated with a loop electrosurgical excision procedure at the first visit when findings suggest high-grade cervical intraepithelial neoplasia (CIN), making accurate colposcopic diagnosis essential. This systematic review evaluates colposcopy guidelines and supporting evidence to identify macroscopic indicators that should guide patient selection for see-and-treat management to minimize over- and undertreatment.

TRIP and Google Scholar were searched in January 2021 to identify guidelines describing colposcopic indicators. In April 2025, MEDLINE and Embase were searched for primary studies reporting diagnostic accuracy or interrater agreement of these indicators. Embase was also searched for prediction models for colposcopic impression, and Scopus and Web of Science for studies citing the Reid Index and Swede score. Two reviewers independently selected studies. One reviewer extracted study characteristics, quality (adapted QUADAS-2), and results, which were verified by a second reviewer. Findings were summarized narratively.

Six guidelines were included. Eight indicators were mentioned in at least three guidelines: visibility of the squamocolumnar junction, colposcopic impression, acetowhite lesion, lesion location, lesion size, Lugol's staining, vascular patterns, lesion border and surface. Thirty-nine primary studies were included. Combinations of indicators to reach a colposcopic impression showed overtreatment rates of 6.3-56.5% for CIN2/moderate dysplasia and 9.9-72.5% for CIN3. Undertreatment occurred in 7.6% of women with normal colposcopic impression but CIN3 or cervical cancer on histopathology.

There is substantial heterogeneity in guideline-recommended colposcopic indicators with limited supporting evidence. See-and-treat protocols may lead to considerable overtreatment and a risk of clinically relevant undertreatment.