Aim: Standardised reporting of thyroid cytology is recommended but not universally practiced. We compared agreement between clinicians categorising non-standardised thyroid cytology reports, and determined malignancy rates in clinician-assigned cytology categories.
Methods: We identified all thyroid cytology reports from 2008-9 and any reports prior to histology samples from 2008-9 at Middlemore Hospital. Two clinicians independently classified these cytology reports using the Bethesda System, and we assessed agreement between their classifications. We classified histology results following the cytology sample as benign or malignant according to the primary diagnosis in the histology report, and calculated malignancy rates for each cytology category.
Results: Agreement between the classifications of 259 cytology results from 227 patients was moderate (kappa=0.67, 95% confidence interval 0.61-0.74), with good agreement (>80%) for only 3 of the 6 Bethesda categories. 122 patients had subsequent thyroid histology samples. 88% had benign primary diagnoses, with 15 (12%) primary thyroid cancers and an additional 11 (9%) incidental cancers. Malignancy rates for each Bethesda category varied from published rates for the majority of clinician-assigned categories.
Conclusions: Low agreement between interpretation of non-standardised thyroid cytology reports suggests that standardised reporting should be universally adopted. Malignancy rates for thyroid cytology categories should be reported to inform local clinical practice.