Risk stratification
No single sonographic feature can reliably differentiate benign from malignant neoplasms. Furthermore, there is variability between the recognition and accurate reporting of these features.17 This has led to the development of standardised thyroid nodule risk-stratification systems such as the American College of Radiology Thyroid Imaging, Reporting and Data System (ACR TI-RADS). This system is superior to other major risk-stratification systems for evaluating thyroid nodules, because it reduces the number of unnecessary biopsies and has a high negative predictive value.18
The ACR TI-RADS assigns points based on five key aspects of the thyroid nodule. The total number of points categorises the nodule into five levels of increasing risk (TR1–5). Features that are suggestive of thyroid cancer (e.g. more than 90% of the nodule has a solid composition, hypoechogenicity, taller-than-wide shape on transverse view, irregular margins, microcalcifications) result in the nodule being assigned more points. In contrast, nodules with benign features (e.g. purely cystic and spongiform nodules) score fewer points. The size of the nodule in conjunction with the risk score guides the recommendation for cytological evaluation (Table 1).
There are other thyroid nodule risk-stratification systems, such as the American Thyroid Association (ATA) guideline. This also categorises nodules into five levels of risk, but smaller and lower risk nodules are recommended for biopsy, in contrast to ACR TI-RADS.2 All the major risk-stratification systems generally do not recommend biopsy for nodules that are below 1 cm in size in the absence of high-risk features, such as suspicious cervical lymphadenopathy.2,18
Monitoring
Nodules that do not meet the criteria for biopsy may require re-imaging at periodic intervals to check for new suspicious sonographic changes or growth (≥20% increase in at least two nodule diameters of ≥2 mm, or ≥50% increase in nodule volume).17 Nodules that exhibit these changes should be referred for cytological evaluation.
There is little consensus about the best interval for repeat ultrasonography or duration of follow-up. The ACR TI-RADS recommends that nodules above a certain size threshold (Table 1) be imaged again at certain intervals, assuming that there are no changes between serial ultrasound scans. TR3 nodules are recommended for repeat ultrasound at one, three and five years, TR4 nodules at one, two, three and five years. Annual scans of TR5 nodules are recommended for five years. The ACR TI-RADS does not recommend re-imaging of TR1–2 nodules or smaller TR3–5 nodules (TR3 nodule <1.5 cm, TR4 nodule <1 cm, TR5 nodule <0.5 cm) (Table 1).19 Clinicians may consider re-imaging of TR1–2 nodules at 24-month intervals and the smaller TR3–5 nodules at 12–24-month intervals.2,17 Routine follow-up of purely cystic or spongiform nodules of 1 cm or less is not recommended.2,4
The recommended duration of long-term follow-up of nodules that remain stable and continue to display low-risk features on serial sonographic assessment is five years, as these nodules have a very low risk of harbouring an undiagnosed malignancy.19,20 Nodules that have not met the criteria for biopsy after five years of ultrasound surveillance, but still have suspicious features, such as TR3–5 nodules, may be considered for further evaluation or ongoing surveillance.19 This should be judged on an individual basis, taking into consideration the characteristics of the nodule, patient’s age and comorbid conditions.
Ultrasound-guided fine-needle aspiration
Cytological evaluation is warranted for suspicious nodules (Table 2). Thyroid nodules with increased uptake on a Tc-99m pertechnetate scan (so-called ‘hot’ nodules) rarely harbour a malignancy and further evaluation with a biopsy is generally not required.7 However, the presence of suspicious sonographic features in these nodules may merit further evaluation and this should be judged on an individual basis. Suspicious cervical lymph nodes should also undergo fine-needle aspiration with a thyroglobulin washout.2