Background The Bethesda system for reporting thyroid cytopathology was proposed to provide a clinically relevant framework for interpretations to improve interobserver agreement. Limited data is available regarding the level of interobserver agreement between groups of cytotechnologists (CTs) and cytopathologists (CPs) examining the same thyroid fine needle aspirate (FNA) samples. Methods Retrospective review of 1,229 thyroid FNAs from 976 patients between 03/2010 and 08/2012 was performed. FNAs received preliminary evaluation by a CT followed by final interpretation by a CP. We calculated Cohen's Kappa coefficient to measure agreement between CTs and CPs, and analyzed levels of discrepancy using delta analysis. Results Overall Kappa between CTs and CPs was 0.79 (95%CI: 0.76-0.83). Kappa values were higher for the nondiagnostic (0.89), benign (0.83), and malignant (0.91) categories than for other categories. Overall Kappa did not show a trend over time, and inversely correlated with the percentage of intermediate grade lesions (coefficient of -0.8; P < 0.01). CTs overcalled more cases (n = 71) than undercalled (n = 29) (P < 0.001), as compared to CPs, with a Δ1 ratio of 2.2 and Δ2 ratio of 3.5. Most two-level discrepancies were related to follicular lesions (19/21) (P = 0.0002). Differences in sample adequacy assessment occurred in 2% of cases. Conclusion Overall, there was a high level of interpretative agreement between CTs and CPs, which remained stable over time, including judgments regarding specimen adequacy. Agreement was most robust for the benign and malignant categories. Our data supports the current practice of allowing CTs to perform on-site adequacy evaluation of thyroid FNAs. Diagn. Cytopathol. 2016;44:113-118.
All Science Journal Classification (ASJC) codes
- Pathology and Forensic Medicine