Interobserver reproducibility and agreement with original diagnosis in the categories "atypical" and "suspicious for malignancy" for bile and pancreatic duct brushings.

Abstract

The Papanicolaou Society of Cytopathology has developed a set of guidelines which include a diagnostic scheme with the categories "atypical" and "suspicious for malignancy." These intermediate categories may help stratify risk of malignancy for samples obtained from the bile and pancreatic ducts. However, the reproducibility of these intermediate categories is currently unknown. Twenty sequential brushing specimens of bile or pancreatic ducts from each of the categories "atypical" and "suspicious for malignancy" were identified and the slides retrieved. All 40 cases were reviewed independently by four cytopathologists blinded to the original diagnoses. Resulting review diagnoses were statistically analyzed for agreement and the Kappa statistic calculated. Agreement of the observers' diagnoses with original diagnoses was also evaluated. Interobserver agreement was graded as slight to fair with observers agreeing in about 50% of cases. The corresponding kappa statistic for the category "atypical" was 0.21 and 0.18 for the category "suspicious for malignancy." Reviewer agreement with the original reference diagnosis occurred in approximately one half of review diagnoses. Analysis of agreement shows that interobserver agreement was only slight to fair. Despite the categories "atypical" and "suspicious for malignancy" having distinct risks of malignancy (62% versus 74%), the reproducibility of these categories is relatively poor. A single intermediate category may improve reproducibility over the scheme proposed by the Papanicolaou Society of Cytopathology while maintaining an ability to stratify risk of malignancy.