Using the ASC:SIL ratio, human papillomavirus, and interobserver variability to assess and monitor cytopathology fellow training performance

BACKGROUND: The goal of this study was to assess objective measurements of cytopathology fellow performance during their training. METHODS: The authors examined cytopathology performance characteristics (the ratio of atypical squamous cells to squamous intraepithelial lesions [ASC:SIL], interobserver variability [IOV], high-risk human papillomavirus [hr-HPV]-positive atypical squamous cells of undetermined significance [ASC-US]) of cytopathology fellows and assessed whether they could be used as tools to further their education. RESULTS: The ASC:SIL ratio, the proportion of hr-HPV-positive ASC-US, and IOV were calculated for 5 consecutive cytopathology fellows. The average ASC:SIL ratio for the fellows was 1.15. The overall average Cohen κ-coefficient (κ-value) between fellow and cytopathologist interpretation was 0.75 (substantial agreement). The conditional κ-value for ASC-US only was higher for cases the fellows called ASCUS (0.70) than for cases the cytopathologist called ASC-US (0.60). Of the cases that were diagnosed as "negative for intraepithelial lesion or malignancy" (NILM) by the fellow and ASC-US by the pathologist, 33.2% were positive for hr-HPV. This was higher than the expected frequency of hr-HPV-positive results in the NILM population, suggesting that the fellows were over-interpreting NILM in hr-HPV-positive cases that had cytologic features sufficient for an ASC-US interpretation. CONCLUSIONS: In this study, agreement was compared between trainee and cytopathologist to determine where a fellow's interpretation differed. With the use of IOV, the ASC:SIL ratio, and the percentage of hr-HPV-positive results in the NILM, ASC-US, and low-grade SIL categories, the authors attempted to outline objective assessments and areas of improvement for fellows before they enter independent practice.