The performance of a new scoring system for colposcopy in detecting high-grade dysplasia in the uterine cervix

Abstract

Objective: To construct a simple scoring system for colposcopic examination that can facilitate education of colposcopists and increase the accuracy of evaluation.

Design: Prospective clinical study.

Setting and population: Two hundred ninety-seven examinations of women referred for colposcopy in western Sweden.

Methods: Five variables were scored: acetowhiteness, margins and surface, vessels, lesion size, and iodine staining. Each variable could be assigned one of three ordered values. Multiple logistic regression was used in order to assess the ability of each single score to predict high-grade lesions (HGL) in histology (cone or biopsy).

Main outcome measures: Histopathology.

Results: All five variables independently predicted for HGL. The analysis resulted in an 'ideal' weighted scoring system, which showed good sensitivity and specificity. Rounding off of each weight gave a more useful and simpler scoring system with values of 0, 1, or 2 without any significant change in performance. The possible total score was then 0-10. A score of > or =5 points identified all HGL and > or =8 points had a specificity of 90%.

Conclusions: The scoring system safely identified a group of patients with low-grade lesions or normal findings, thus allowing 17% to be followed only by colposcopy or cytology. Furthermore, it could select women for see-and-treat with only 10% of cases having less than HGL. With this strategy, only approximately 50% of the cases would have needed biopsy in the evaluation.