Detection of conditions related to human papillomavirus. Comparison of cytology, colposcopy, histology and hybridization.

The diagnosis of lesions associated with human papillomavirus infection can be difficult because the results of the tests used can be contradictory. Our goal was to compare some of these tests and to evaluate their comparative strengths and weaknesses as clinically useful tools in confirming the diagnosis, especially in borderline cases. Twenty-one consecutive patients from our colposcopy clinic were screened with cytology and colposcopy. Biopsies were taken from representative areas on the cervix and vulva and divided. One-half was evaluated with Southern blot hybridization and the other half with histology and with RNA and DNA in situ hybridization. Cytology and histology were interpreted as either "positive" (showing definite evidence of human papillomavirus infection or cervical intraepithelial neoplasia [CIN]), "negative" (showing no evidence of human papillomavirus infection or CIN) or "equivocal" (atypical [class II] Papanicolaou smears or histology suggestive but not diagnostic of condyloma). In order to determine the clinical significance of equivocal results the sensitivity and specificity of these tests were calculated, with the equivocal results reclassified as either positive or negative. Colposcopy was the most sensitive technique but was not very specific. Cytology was a very sensitive screening tool when the atypical (class II) smears were considered positive but not when they were considered negative. The specificity of the histologic diagnosis was doubled with the equivocal results considered negative when compared to the specificity of the histologic diagnosis with the equivocal results considered positive, with no loss of sensitivity. Each technique has drawbacks, and therefore no one should be used to diagnose and treat these lesions to the exclusion of all others.