Papanicolaou and thin-layer cervical cytology with colposcopic biopsy control. A comparison.

OBJECTIVE: To compare the accuracy of conventional Papanicolaou and fluid-based, thin-layer cervical cytology. STUDY DESIGN: Cervical cytology was performed in duplicate on women who presented for cervical screening. Papanicolaou and thin-layer (ThinPrep, Cytyc Corp., Boxborough, Massachusetts) cytologic samples were collected simultaneously using a split-sample method. Cytologic slides were read and reported independently. Clinical follow-up was based on the most abnormal result. Colposcopy was performed as clinically indicated, and biopsy results were compared with cytologic diagnoses. RESULTS: Three thousand samples were compared. Papanicolaou and thin-layer results were significantly different (P = .0001), with identical diagnoses in 1,844 (61%) of patients. Eighty thin-layer (2.7%) and 177 Papanicolaou (5.9%) samples were read as limited or unsatisfactory (P < .0001). The rates of atypical squamous cells of undetermined significance (ASCUS) were not statistically different (P = .06). Thin-layer cytology was read more often as low grade squamous intraepithelial lesion (P = .001) or high grade squamous intraepithelial lesion (P = .006). Colposcopy with biopsy was performed on 291 patients. With ASCUS considered an abnormal result, thin-layer cytology was more sensitive (91% vs. 85%) but had lower positive predictive value (69% vs. 74%) than Papanicolaou cytology for the presence of cervical neoplasia. CONCLUSION: Papanicolaou and thin-layer cervical cytology yielded significantly different information. Thin-layer cytology yielded significantly fewer unsatisfactory results and was more sensitive for identifying cervical intraepithelial neoplasia.