The utility of combined mutation analysis and microRNA classification in reclassifying cancer risk of cytologically indeterminate thyroid nodules.

ABSTRACT

OBJECTIVES: Real-world clinical results of (1) Bethesda categorization, (2) mutation analysis, and (3) a microRNA classifier were correlated to show the utility of molecular analysis in assessing malignancy risk of indeterminate thyroid nodules. METHODS: Cytology and molecular results of clinically tested thyroid nodules were compared. An additional microRNA threshold was determined based on nodules with known disease status, establishing a 3-tiered microRNA approach to clinical risk assessments. Expected rate of malignancy given mutation panel and 3-tiered microRNA approach was validated in an independent cohort of atypia of undetermined significance or follicular lesion of undetermined significance (AUS/FLUS) and follicular neoplasm or suspicious for follicular neoplasm (FN/SFN) nodules with surgically derived outcomes. RESULTS: In 2685 patients clinically tested, PIK3CA, PAX8/PPARγ, and RET/PTC mutations occurred in less than 1%. Of note, 2% had BRAFV600E mutation and 82% lacked mutations. The maximum expected risk of malignancy in nodules lacking mutations was 9% and 17% for AUS/FLUS and FN/SFN nodules, respectively. Positive microRNA status further increased risk, with the most worrisome status (level-3) elevating risk to 36% and 54%, respectively. RAS mutations occurred in 15% of nodules tested clinically, including in 8% of those that were cytologically benign. The maximum expected risk of malignancy in nodules with RAS or PAX8/PPARγ mutations was 49% and 65% for AUS/FLUS and FN/SFN nodules, respectively. Positive microRNA status further increased risk, with the most worrisome microRNA status (level-3) elevating risk to 85% and 91%, respectively. CONCLUSIONS: Mutation panels alone do not sufficiently risk stratify thyroid nodular disease. microRNA classification complements cytology and mutation analysis with the capacity to better differentiate nodules at high risk of malignancy.