Cytologic/histologic correlation for quality control in cervicovaginal cytology. Experience with 1,582 paired cases.

For quality-control purposes, federal regulations require cytology laboratories to compare Papanicolaou smear and cervical biopsy reports, if available, and determine the cause of any discrepancies. The authors reviewed 56,497 cervicovaginal smears, of which 1,582 (2.8%) had a subsequent cervical biopsy or endocervical curettage within 2 months. A total of 175 discrepancies (11%) were identifed, and biopsies and smears from these cases were reviewed at a weekly conference. In the majority of cases, the diagnosis of the smear and biopsy was confirmed on review, and the discrepancy was attributed to sampling error (n = 159; 91%). Six cases (3.4% of all discrepant cases) were errors in cytologic diagnosis. Five of these were interpretation errors, and one case was a screening error. There were 10 errors in the evaluation of cervical biopsies (5.7% of all discrepant cases): five biopsies were undercalled, and five were overdiagnosed as a squamous intraepithelial lesion. The results of testing for human papillomavirus DNA by in situ hybridization were helpful in arbitrating some diagnositic disagreements.