Cyto-histological correlates in a colposcopic clinic: a 1-year prospective study

As part of the Quality Assurance Program, all surgical pathology reports of cervical colposcopic biopsies, cone biopsies, and hysterectomy specimens of 772 consecutive patients were correlated with cytology results for a period of 1 yr. The abnormality in cytology or histology ranged from HPV effect to invasive carcinoma. Cytohistological correlation within one grade of severity was obtained in 662 (86%) cases. On formal review of the remaining 110 (14%) cases, sampling error occurred in 75 cytology smears and in 21 colposcopic biopsies. Screening and interpretation error occurred in 10 smears and four biopsies. In this study the sensitivity of cervical cytology in detecting epithelial abnormalities was 90%, with a positive predictive value of 99.7%. The false-negative rate of cytology was 11% and the false-positive rate 0.26%. The majority (90%) of false negatives were due to sampling error; the rest were due to laboratory error. The sensitivity of colposcopic biopsy was 97% with a positive predictive value of 99.4%. The false-negative rate of biopsy was 2.7% and the false-positive rate 0.5%. Endocervical cells were seen in 48% (370) of smears and absent in 52% (402). However, in both instances the detection rate of squamous abnormality was 90%. Also 43% (33) of smears that showed endocervical cells failed to contain the abnormal cells. Endocervical curettage or cone biopsy in 16 of these 33 cases clearly showed a small CIN-III lesion high in the endocervical canal. We believe that the value of endocervical cells as an indicator of adequacy is questionable.(ABSTRACT TRUNCATED AT 250 WORDS)