Knowledge of Patients’ Human Papillomavirus Status at the Time of Cytologic Review Significantly Affects the Performance of Cervical Cytology in the ATHENA Study

OBJECTIVES: With human papillomavirus (HPV) testing, patients' HPV status may be known when reviewing cytology specimens. METHODS: 41,955 women 25 years or older had cytology and HPV screening. Originally, cytology was reviewed blinded to HPV status. We re-reviewed unblinded to HPV status a subset of 428 cytology slides from women with cervical intraepithelial neoplasia grade 2 + (CIN2+) and 1,287 from women without CIN2+. RESULTS: Of the original interpretations of atypical squamous cells of undetermined significance (ASC-US), 33.7% were downgraded to negative after unblinded review, and 8.7% were upgraded to atypical squamous cells, cannot rule out a high-grade squamous intraepithelial lesion or high-grade squamous intraepithelial lesion. Of the original interpretations of ASC-US, 66.7% were downgraded on unblinded review in HPV-negative women and 30.2% were upgraded in HPV 16+/HPV 18+ women. Unblinding increases the sensitivity for cervical intraepithelial neoplasia grade 3+ of cotesting from 54.1% to 62.4% (P = .0015) and the sensitivity of HPV primary screening from 72.2% to 77.1% (P = .0029). With cotesting, specificity with unblinding is improved, whereas with HPV primary screening, there would be a small decrease in specificity. CONCLUSIONS: Unblinded cytology increases overall sensitivity with either cotesting or HPV primary screening; specificity is either slightly improved or is not affected by unblinding.