Hurthle cell predominance impacts results of Afirma gene expression classifier and ThyroSeq molecular panel performance in indeterminate thyroid nodules.

ABSTRACT

BACKGROUND: Molecular tests such as the Afirma gene expression classifier (GEC) and mutational panels (such as ThyroSeq) have been introduced to help risk stratify cytologically indeterminate thyroid nodules with the aim to reduce the number of unnecessary thyroidectomies. Some reports have suggested that samples with Hurthle cell predominance have higher false-positive rates on GEC testing, but data are limited. METHODS: We reviewed thyroid nodules with indeterminate (Bethesda III/IV) cytology at our institution. Patient demographics, cytologic and histologic diagnoses (where available), and molecular test results were collected. RESULTS: GEC was performed on 202 nodules, and ThyroSeq was performed on 81 nodules. In the GEC cohort, 66% of nodules with Hurthle cell predominance yielded "suspicious" result vs 46% of nodules without Hurthle cell predominance, with risk of malignancy (ROM) for surgically resected nodules of 16% and 33%, respectively. In ThyroSeq cohort, 8% of nodules with Hurthle cell predominance yielded a high-risk mutation vs 19% of nodules without Hurthle cell predominance, with ROM of 50% and 33%, respectively. CONCLUSIONS: For ThyroSeq molecular panel, while it did not appear that there was an increase in rate of high-risk mutations detected in the samples with Hurthle cell predominance, small numbers limit the generalizability of these results. For the GEC cohort, indeterminate thyroid nodules with predominance of Hurthle cells showed an increased rate of "suspicious" results compared to samples without Hurthle cell predominance. The ROM for GEC "suspicious" nodules with Hurthle cell predominance on surgical resection was lower in our study. Repeat FNA may be of use in patients with these types of nodules. In the context of a Hurthle cell predominant lesion, positive results on molecular testing may not carry a high rate of malignancy.