Probability of malignancy as determined by ThyroSeq v3 genomic classifier varies according to the subtype of atypia

ThyroSeq assesses the probability of malignancy (POM) in thyroid fine-needle aspiration cytology specimens diagnosed as atypia of undetermined significance (AUS). The authors investigated whether defined AUS subcategories are associated with specific molecular alterations, the molecular-derived risk of malignancy (MDROM), and the risk of malignancy (ROM).

Fine-needle aspiration cytology reports of AUS and corresponding results from the ThyroSeq version 3 genomic classifier results were retrieved and subcategorized as follicular cells with either cytologic atypia (FC-C), architectural atypia (FC-A), both cytologic and architectural atypia (FC-CA), or a predominance of Hurthle cells (PHC). The MDROM, ROM, and frequency of molecular alterations by subcategory were computed and analyzed, and p < .05 was considered significant.

The final analysis included 541 cases subdivided into 233 with FC-A, 104 with FC-C, 116 with FC-CA, and 88 with PHC. The benign call rate and positive call rate for the AUS category were 72% and 28%, respectively, which varied between AUS subcategories. The MDROM by subcategory was 15.9% FC-A, 20.5% FC-C, 33.8% FC-CA, and 14.4% PHC. Histologic follow-up was available for 155 (28%) AUS cases with a follow-up period ≥12 months. The 95% confidence intervals of the MDROMs overlapped with the ROMs. The highest MDROM and ROM were in the FC-CA subcategory. RAS mutations were present in all subcategories. BRAF V600E mutations and papillary thyroid carcinoma were most frequent in the FC-CA subcategory. Noninvasive follicular thyroid neoplasm with papillary-like nuclear features was significantly more frequent in the FC-C subcategory.

The current results demonstrated that AUS subcategories are associated with specific genetic alterations, the MDROM, and the ROM. Molecular results and an awareness of various cancer probabilities within AUS subcategories can allow for a more tailored management.