Interobserver variability amongst cytopathologists and histopathologists in the diagnosis of neoplastic follicular...

The difficulty in correctly identifying follicular lesions/neoplasms (FN) of thyroid is a well recognized fact. The present study was undertaken to study the degree of interobserver variability among cytopathologists and histopathologists in the diagnosis of these lesions. A total of 74 cases were chosen with a cytological diagnosis varying from FN (43 cases), hürthle cell neoplasm (18 cases), neoplasm suspicious of papillary carcinoma (7 cases) and follicular variant of papillary carcinoma (FVPTC) (4 cases) and neoplasm, NOS (2 cases). All these cases were reviewed by a cytopathologist along with surgical follow-up. In cases of cytohistological discordance, histopathological diagnoses were also reviewed by a histopathologist. The accuracy of initial and reviewed cytology in predicting neoplastic histological outcome was 100%; however 15 cases were over-diagnosed on initial cytology as FN and on reviewed cytology, 8/15 cases were under-diagnosed as non-neoplastic; however still seven cases were over-diagnosed on reviewed cytology, thereby specificity for neoplastic histological outcome ranging from 50 to 68%. The histopathology diagnosis was reviewed and changed in six cases; therefore interobserver agreement among histopathologist is 92% which reduces further in diagnosis of FVPTC, as in two cases diagnosis of follicular adenoma was changed to FVPTC. Neoplastic follicular patterned lesions (FN) pose diagnostic difficulties among pathologists. So, all cases with a cytological diagnosis of FN should be reviewed along with histopathology to minimize interobserver variability and reviewed histopathology should be considered as "gold standard".