Outcome of thyroid gene expression classifier testing in clinical practice.

OBJECTIVES/HYPOTHESIS: Gene expression classifiers can safely reduce diagnostic thyroid surgery for fine-needle aspiration cytology (FNAC) indeterminate thyroid nodules.

STUDY DESIGN: Retrospective review, single-institution, single-practice surgeon.

METHODS: Three-year retrospective review of indeterminate FNAC that went on to gene expression classifier testing.

RESULTS: A total of 520 patients met American Thyroid Association guideline criteria for surgeon-performed ultrasound-guided FNAC for a thyroid nodule with on-site cytopathology. The indeterminate (Bethesda III or IV) FNAC rate was 9%. Prevalence of malignancy in FNAC indeterminate was 21%. Thirty-two cases went on to gene expression classifier testing. Fourteen were benign, 15 suspicious, and three with no result.

CONCLUSIONS: Benign gene expression classifier testing had an estimated negative predictive value of 100% during the study period. These patients have been observed for a mean and median duration of 14 and 7 months, respectively. In this small series, 14 of 29 patients with indeterminate FNAC were spared diagnostic surgery.

LEVEL OF EVIDENCE: 4.