Molecular Testing for Indeterminate Thyroid Nodules: Association of Negative Predictive Value With Nodule Size

Background: Molecular testing helps stratify risk of malignancy in indeterminate thyroid nodules, but it may be limited in its ability to rule out malignancy in large nodules. Methods: We compared small thyroid nodules (<4 cm) to large thyroid nodules (4 cm or greater) in a retrospective, single center, cohort study of indeterminate thyroid nodules analyzed with ThyroSeq from 10/2015 through 4/2020. Our primary outcome was negative predictive value. Results: A total of 204 thyroid nodules were analyzed with ThyroSeq and 62 underwent resection allowing comparison to final pathology for 48 small nodules and 14 large nodules. A greater proportion of large nodules were observed in men (50.0% vs 18.8% in small nodules, P = .02), but median age was similar between the groups (49.0 vs 52.5 years, P = .95). False negative results comprised a higher proportion of tests for large nodules (14.3%) than small nodules (0%, P = .01). The negative predictive value of ThyroSeq in small nodules was 1 indicating a negative test reliably predicted a benign nodule, compared to .5 in large nodules indicating a negative test in this cohort was reliable in predicting benignity. Applying the same negative predictive value to the cohort of large nodules that did not undergo resection after negative molecular testing (n = 16), it is possible that 8 large thyroid malignancies were missed at our institution during the study period. Conclusions: Large thyroid nodules are associated with a higher rate of false negative results and a lower negative predictive value during molecular analysis of indeterminate thyroid nodules.