OBJECTIVE

Molecular testing is recommended for risk stratification of atypia of undetermined significance (AUS) nodules in the USA; however, it is not routinely performed in some countries owing to limited availability and affordability. Here, we propose a risk stratification algorithm for AUS nodules when molecular testing is unavailable.

METHODS

We examined 304 (4.3%) AUS nodules among 7073 thyroid fine-needle aspiration cytology specimens examined at Kuma Hospital from January 2020 to December 2020. Clinical data were obtained from the medical records of Kuma Hospital.

RESULTS

AUS with nuclear atypia and AUS-other each accounted for half of the total AUS nodules. The repeat aspiration rate was 19.7%; 61.7% of the nodules were reclassified as benign or malignant upon repeat aspiration. Resection rate and overall risk of malignancy (ROM) were 32.6% and 12.8%, respectively. Architectural atypia showed the lowest (1.1%) overall ROM in the AUS nodules. For AUS with nuclear atypia, nodules ≤ 10 mm in size showed significantly lower overall ROM than those of > 10 mm, and nodules with ultrasonographically low suspicion showed significantly lower overall ROM than those with intermediate to high suspicion. AUS nodules with atypical lymphoid cells, possible medullary thyroid carcinoma, or possible parathyroid lesion were confirmed using flow cytometry, biochemical testing using needle washout fluid or immunocytochemistry, respectively.

CONCLUSIONS

Our proposed clinical management algorithm for each subdivision according to cytological findings, based on repeat aspiration rates, ROM, ultrasound findings and results of ancillary tests except for molecular testing, should be useful for the clinical management of AUS nodules.