Abstract
The evolution of thyroid cancer diagnostics has been beneficially impacted by imaging modalities such as 131I nuclear scanning and ultrasound (US), and by the introduction of fine needle aspiration (FNA) biopsy techniques in the 1980s, itself leading to a twofold decrease in diagnostic thyroidectomies with a twofold increase in histologic malignancy (1). Diagnostic advancements have also occurred with Bethesda standardization of cytologic interpretations, addition of US to guide FNA biopsy, and most recently, the use of molecular testing (MT) for further risk assessment (2). With next-generation technology, the cost, sensitivity, and reliability of nucleotide-level MT characterization have facilitated further refinement in thyroid diagnostic testing (3).
Since MT was developed to help reconcile cancer risk in cytologically indeterminate (Bethesda III/IV) nodules, initial reports focused on accuracy and diagnostic performance. As the clinical experience with MT expands, studies are now emerging that further evaluate whether and how such testing can optimize preoperative decision-making as the management of thyroid cancer becomes more personalized (2). In this context, Guan et al. report a valuable single institution experience with ThyroSeq v2/3 testing and describe results in 80 RAS-positive nodules that were heterogeneously selected for MT and for surgical management (4). Although retrospective, it was likely indicative of real-world practice, and the interesting findings suggest that in cytologically indeterminate nodules, (a) RAS is the most common mutation identified with an equivalent incidence in Bethesda III (32%) compared with Bethesda IV (36%) nodules, (b) in the 70% who received surgery for RAS-positive Bethesda III/IV nodules (n = 40), the rate of histologic malignancy or noninvasive follicular thyroid neoplasm with papillary-like nuclear features (NIFTP) was 3-fold higher than in RAS-negative nodules (30% vs. 9%), and (c) utilization of surgery for definitive diagnosis was 5-fold less likely when RAS testing was negative (70% vs. 14%) (4).
Similar to prior reports, the authors also observed that RAS testing alone has moderate sensitivity (80%) and specificity (52%) for predicting cancer (4). The American College of Radiology Thyroid Imaging Reporting and Data System features were not associated with malignancy in this study, although the US characteristics considered high risk in these classification systems typically correlate with conventional papillary thyroid cancer (PTC), which comprised only a minority of the malignancies described. However, RAS was associated with a higher rate of malignancy/NIFTP in Bethesda IV than Bethesda III nodules and MT results should be interpreted in the context of other available variables known to be associated with thyroid malignancy (4). Ultimately, any study assessing the clinical utility of MT for cytologically indeterminate nodules, which does add cost, should also account for whether results will change patient management. There is currently no documented role for MT if the decision for thyroidectomy, extent of surgery, or surveillance can be reached otherwise (5).
The characterization of histology in this series was also informative for patient counseling particularly in the post-NIFTP era (6). Although the study's benign nodules are not further described as being hyperplastic or clonal adenomas, ∼50% of the nonbenign RAS-positive Bethesda III/IV nodules were NIFTP, while the other half were malignancies (4). Furthermore, the majority of the RAS-positive malignancies were follicular variant of PTC typically associated with low recurrence risk, which is consistent with prior reports (7). Thus, in an otherwise cytologically and radiographically indeterminate nodule, RAS positivity can be productively utilized to recommend initial thyroid lobectomy as likely the definitive surgery for most patients.
Interestingly, nodules with multiple mutations had a higher risk of malignancy, and 2 of 4 nodules with concurrent TERT mutations had concerning histologic features including poorly differentiated subtype and angioinvasion. Other studies also support this finding and, particularly in conjunction with other driver mutations, TERT mutations should be considered a marker of likely aggressive thyroid cancer (8). While active surveillance of RAS-positive nodules deserves careful consideration, long-term data to confirm safety are still needed (5) as poorly differentiated and medullary thyroid cancer were present in 17% of the RAS-only histologic cancers diagnosed here (4). It is unlikely that single institution data are going to be sufficiently powered to adequately answer this question and a collaborative multicenter study is needed.
Simply put, MT results are no longer binary. Instead, diagnostic tests for thyroid nodules are symbiotic and complementary, and no single test or genetic alteration will be adequate in isolation to accurately diagnose or exclude thyroid cancer and direct ideal management. Testing for RAS is not a home run, but it is useful and part of a growing compendium. Just as US is currently used to refine thyroid nodule assessment and guide FNA biopsy, similarly, incorporating the comprehensive molecular data now available with MT (including RAS status) can enhance already available clinical, imaging, and cytologic information to additionally optimize patient care decisions (5).