Interobserver variability in inconclusive diagnostic categories of thyroid fine needle aspiration cytology: An urban-based tertiary hospital experience.

Introduction: Thyroid nodules are typical lesions, usually non-malignant, and surgery is unnecessary in most patients. However, distinguishing between benign and malignant is challenging. Fine needle aspiration cytology (FNAC) is considered a primary diagnostic and prognostic tool with an effective cost for evaluating thyroid enlargement. Unfortunately, using FNAC to diagnose inconclusive lesions in the category III-Atypia of Undetermined Significance/Follicular Lesion of Undetermined Significance (AUS/FLUS) based on TBSRTC classification is a problematic issue. This study aimed to evaluate the interobserver variability of AUS/FLUS interpretation among pathologists.

Methods: A retro-observational study: previous 127 AUS/FLUS cases were enrolled. Seventy-two cases met inclusion criteria and were then reclassified by different anatomical pathologists under blinded-design assignments. The concordance among pathologists and the percent alteration of the risk of malignancy (ROM) were compared to the original reports and histological diagnosis.

Results: About 72 % of AUS/FLUS cases were changed after the reclassification. Approximately 46 % were changed to benign while 12.5 % were reclassified as carcinoma. Moreover, 30 % of those original AUS/FLUS were histologically diagnosed as malignant or carcinoma lesions. The concordances among consensus diagnosis and results from each pathologist are acceptable, Kappa(s) were 0.674 to 0.898 (p < 0.001) and Spearman correlations were 0.820 to 0.957 (p < 0.0001).

Conclusion: There are substantial interobserver differences and changes in cytological diagnosis when re-evaluation is performed by multiple pathologists using TBSRTC. A second or third opinion should be sought routinely to establish a consensus diagnosis as a supplement to the initial diagnosis of AUS/FLUS. The reclassification reduces medical expenses and the rate of unnecessary surgery, especially in patients with cytologically confirmed benign thyroid nodules. Preoperative molecular evaluation is a promising method for assisting in the diagnosis of thyroid nodules, but additional research is necessary.