Thyroid nodules are extremely common with a prevalence of over 50% in the general population, most of which are impalpable. Around 90–95% are benign, which comprise hyperplastic or colloid nodules and cysts (95%), follicular or Hurthle cell adenomas (5%), and focal thyroiditis (1%). Between 5 and 10% of nodules are malignant. These comprise differentiated thyroid cancers (papillary cancers (85%), follicular carcinoma (10%), medullary cell carcinomas (< 5%)), undifferentiated carcinomas (1–2%), and miscellaneous rare malignancies (e.g. lymphoma, sarcomas, metastases).
Ultrasound (US) is the initial imaging modality for suspected focal thyroid disease, as it is sensitive and permits characterization, which in turn is used to select nodules for fine needle aspiration for cytology (FNAC) or core needle biopsy. Advances in US technology over recent years have improved spatial resolution and image quality including delineation of features such as nodule margin and composition, resulting in improved detection of smaller cancers. Nevertheless, US is a dynamic and highly operator-dependent technique.
The US features suggestive of benignity and malignancy are well documented.
Benign features include cystic or predominantly cystic composition, smooth margins, spongiform appearance (Figure 1), multiple similar conglomerate nodules, isoechoic or hyperechoic echotexture, comet tail artefacts, and a thin intact capsule. Malignancy is suggested by irregular margins, hypoechoic or markedly hypoechoic echotexture, a taller-than-wide shape (Figure 2), punctate microcalcifications, extrathyroid invasion, and suspicious nodal metastases. Other sonographic features, including vascular patterns and macrocalcifications, have comparatively low discriminatory performances and thus questionable utilities. More recently, nodule stiffness on US elastography has been under intense investigation. Based on the same principles as palpation, malignant nodules are generally firmer on elastography than benign nodules, although there is some overlap. At present, the diagnostic accuracy and utility of US elastography is controversial. In fact, benign and malignant nodules show substantial variation in all sonographic features, with most showing a trend towards high specificity but poor sensitivity. Consequently, no single criterion or combination of US criteria can achieve both a high sensitivity and specificity for malignancy. Follicular neoplasms are particularly problematic as these show marked imaging overlap with follicular adenomas and hyperplastic nodules (Figure 3).
In 2009, Horvath et al. proposed the first US-based malignant stratification system for thyroid nodules, termed Thyroid Imaging Reporting and Data System (TIRADS). This was a landmark in thyroid US towards greater standardization and a systematic tiered approach to guide further investigations and treatment. Since then, several other US-based scoring systems have been proposed by different expert groups including K-TIRADS (Korean), EU-TIRADS systems, ACR-TIRADS (USA), and U classification (UK). These systems either employ pattern recognition or a cumulative score for points assigned to individual US criteria. The grade or cumulative score can be used to estimate the malignant risk for a given nodule, which can be combined with other features such as nodule size to guide work-up including the requirement for FNAC or imaging follow-up. At present, it not clear which scoring system is more accurate or simpler to use.
Fine needle cytology results are nowadays interpreted using a widely accepted cytopathological grading system (Bethesda). Cytological results are assigned to one of six categories, each of which has a different range of malignant risk. The nodule’s US appearances, cytological grade, and patient-related factors should be evaluated together to estimate the overall risk of malignancy and guide management; this is best achieved in a multidisciplinary team setting.
US is also important for thyroid cancer staging and post-treatment monitoring. US features influencing staging include tumour size, extranodal extension, and regional lymph node metastases.
The prognosis of differentiated thyroid cancer is generally excellent following treatment, equating to over 98% five-year survival. Nevertheless, treatment of thyroid cancer is not without risks and long-term morbidity, including vocal cord palsy and hypothyroidism. Furthermore, it is now widely recognized that a substantial proportion of differentiated cancers are indolent and asymptomatic during life, especially if subcentimetre, termed microcarcinomas. These facts, combined with a progressive rise in absolute number and proportion of smaller cancers being detected in recent years, has led to a paradigm shift in thyroid cancer management towards observation rather than upfront treatment of microcarcinomas with no “high-risk” features. Some of these high-risk features include US findings such as precise nodule position and thyroid capsule abutment.
This talk will outline and illustrate the aforementioned points and will include salient examples to highlight potential pitfalls in US interpretation.
The author has no competing interests to declare.