Ultrasound Systems for Risk Stratification of Thyroid Nodules Prompt Inappropriate Biopsy in Autonomously Functioning Thyroid Nodules

Marco Castellana; Camilla Virili; Gaetano Paone; Lorenzo Scappaticcio; Arnoldo Piccardo; Luca Giovanella; Pierpaolo Trimboli


Clin Endocrinol. 2020;93(1):67-75. 

In This Article

Abstract and Introduction


Objective: Ultrasound (US) risk stratification systems (RSSs) have been developed to reduce the number of unnecessary fine-needle aspiration (FNA) biopsies in patients with thyroid nodules. Autonomously functioning thyroid nodules (AFTN) account for 5%-10% of palpable lesions and are very rarely malignant. The present study was undertaken to investigate how RSSs classify AFTNs and whether RSSs are able to avoid unnecessary FNA biopsies in such cases.

Methods: Patients with AFTN who had undergone US, scintigraphy and thyroid function evaluation from December 2016 to December 2017 were selected. US images were retrospectively reviewed and AFTN reclassified according to AACE/ACE/AME, ACR-TIRADS, ATA, BTA, EU-TIRADS, K-TIRADS and TIRADS. Risk class and indication for FNA were assessed.

Results: A number of 87 AFTNs from 85 consecutive patients were enrolled. A median diameter of 22 mm (range 10–59) was found, with an ovoid isoechoic nodule being the most frequent US presentation. When AFTNs were reclassified according to US RSSs, the most common categories were low and intermediate risk. AFTNs were assessed as being at high risk/high suspicion/malignant in 1%-9%, with good agreement among AACE/ACE/AME, ATA, EU-TIRADS, K-TIRADS and TIRADS. Remarkably, FNA was indicated in 27%-90% of AFTNs. A statistically significant difference among the systems was found; 8% of cases were nonclassifiable by one or more US RSS.

Conclusions: Ultrasound RSSs prompt inappropriate FNA in a significant number of patients with AFTN.


Thyroid nodules are common, and ultrasound (US) is the first-line imaging tool to stratify the risk of malignancy. Specific US features, such as hypoechogenicity, taller-than-wide shape on transverse view, irregular margins, microcalcifications and extrathyroidal extension, are recognized to be associated with cancer.[1] However, many factors (ie suboptimal sensitivity and specificity, intra- and interassessment variability, operator dependence) hamper the safe and reliable use of these factors to predict or exclude malignancy in clinical practice.[2] Several US risk stratification systems (RSSs) have therefore been developed in an attempt to improve the performance of US in selecting nodules for fine-needle aspiration (FNA) and, consequently, patient management. Three of these systems are included in clinical guidelines for the diagnosis and clinical management of thyroid nodules and carcinoma,[3–5] while the remaining ones provide specific recommendations for selecting nodules for FNA;[6–9] three to six classes of risk of malignancy are described. Notably, all information on US reliability is based on studies mainly involving papillary thyroid carcinoma (PTC), as this cancer is the most prevalent thyroid malignancy. Moreover, in the available literature, a cytological diagnosis has generally been adopted as the gold standard, thus introducing a significant bias. Indeed, follicular thyroid cancer (FTC) may not be detected on FNA and, additionally, medullary thyroid cancer (MTC) is missed by cytology in up to 50% of cases. Furthermore, FTC and MTC may have a heterogeneous/unsuspicious US presentation.[10,11]

Autonomously functioning thyroid nodules (AFTNs) can also complicate this situation. The prevalence of malignancy among AFTNs is low, and indeed, AFTN is considered a benign entity in clinical practice.[4,5,12] AFTN accounts for 5%-10% of palpable nodules, with a prevalence of about 20% in iodine-deficient regions.[12,13] Thyroid scintigraphy (TS), with either 99mTc-pertechnetate or I-123-natrium iodide (I-123), is the only method of diagnosing AFTN and is recommended by the 2015 American Thyroid Association guidelines in patients with subnormal serum TSH.[5] The relationship between thyroid autonomy and TSH levels, however, is largely influenced by iodine intake, and normal TSH levels are detected in 50% of patients with AFTNs.[14] Consequently, variable indications are reported in the different clinical guidelines regarding the scintigraphic evaluation of thyroid nodules in the light of TSH.[4,15] It must also be considered that the sonographic features of AFTNs are heterogeneous.[16,17] In the presence of normal TSH levels, the likelihood of selecting these nodules for FNA might be high when TS is not performed and, even more disturbing, a significant number of indeterminate cytological reports might be expected in these cases.[18]

As most of the currently available US RSSs do not mention the scintigraphic pattern of thyroid nodules, the present study was undertaken to investigate (a) how the US RSSs classify AFTNs, and (b) whether the US RSSs are able to properly exclude AFTNs from FNA. To this end, we retrospectively reviewed all consecutive AFTNs diagnosed at our institution during a specific period and rigorously reclassified them according to all seven most commonly used US RSSs.