Team Reading (Peer Review) of Suspicious/Positive Slides for Continuous Quality Improvement in Cervical-Vaginal Cytology: A Comparison between Methods and Indicators

Abstract

Aim: In 2013, the Local Health Unit Roma 2 ex C screening laboratory introduced a new set of indicators for quality assurance. We compare 2 sets of indicators based on routine multiple readings (peer review) for their ability to identify problems in single-reader accuracy. Methods: All suspect slides were blindly reviewed by all the cytologists of the laboratory. The standard set of indicators includes interreader Cohen's kappa, positivity rate and atypical squamous cell of undetermined significance/squamous intraepithelial lesion (ASC-US/SIL) ratio. The new set included sensitivity for cervical intraepithelial neoplasia of grade 2 or worse (CIN2+), positive predictive value (PPV) and percentage of positive high-risk (HR)-HPV cases among ASC-US. In order to estimate sensitivity and PPV, we considered all women for whom there was a consensus of negative cytology, negative HR-HPV test, negative colposcopy or negative histology true negative. Results: Kappa values ranged from 0.521 to 0.753, with narrow 95% confidence intervals (CIs). Positivity rate ranged from 73.9 to 85.7 and the ASC-US/SIL ratio from 0.61 to 0.81. Sensitivity for CIN2+ at the low-grade SIL threshold ranged from 85.8 to 94.2, PPV ranged from 14.8 to 19.4, and both had a broad 95% CI. Readers with low sensitivity did not show low kappa values. The percentage proportion of HR-HPV-positives (HR-HPV+) among ASC-US ranged from 39.9 to 43.8% with a narrow 95% CI. Conclusion: The proportion of HR-HPV+ among ASC-US cases is a powerful indicator to address in training.