Abstract
BACKGROUND: Preoperative detection of RAS mutations can contribute to cancer risk assessment in indeterminate thyroid nodules, although RAS is not always associated with malignancy. METHODS: Fine-needle aspiration samples classified in 1 of 3 indeterminate cytology categories were prospectively tested for N-, H-, and K-RAS mutations using next-generation sequencing assay. RESULTS: In the study, 93 patients with 94 nodules had preoperative RAS detected, of whom 86 patients had an operation (69% total thyroidectomy, 29% lobectomy). In total, 76% of RAS-positive nodules were malignant and follicular variant papillary thyroid cancer was the most common cancer type (83%). HRAS mutations had the greatest risk of cancer (92%) followed by NRAS (74%) and KRAS (64%; P = .05). No preoperative variables were associated with malignancy including age (P = .07), sex (P = .49), RAS isoform (P = .05), mutational allelic frequency (P = .49), nodule size (P = .14), cytology category (P = .63), or ultrasound bilaterality (P = .24), multifocality (P = .23), or presence of ≥1 suspicious feature (P = .86). Only 60% of patients with a unifocal nodule on ultrasound had single focus low-risk encapsulated follicular variant papillary thyroid cancer or benign disease. CONCLUSION: Preoperative RAS mutation detection in thyroid nodules carries a substantial risk of cancer with a greater risk associated with HRAS and NRAS. Most RAS malignancies are follicular variant papillary thyroid cancer, which may inform the extent of operation.