Overview of the clinical utility of ultrasonography in thyroid disease

INTRODUCTION — Although ultrasonography yields limited diagnostic information, it provides extremely valuable and pertinent information about the thyroid gland, its diseases, and adjacent structures. This cumulative ultrasonography evidence, in context, can significantly augment diagnosis as well as clinical, surgical, radiation, ablation, and other management. High among the merits of ultrasonography is enhanced sampling of thyroid cells and tissues (in conjunction with fine-needle aspiration [FNA] biopsy) for cytologic, biochemical, molecular, and genetic investigation. Ultrasonography also offers insights into aspects of physiology and epidemiology.

This topic will review the clinical utility of thyroid ultrasonography. The technical aspects of thyroid ultrasound and the use of ultrasonography to aid in thyroid biopsy or the destruction (ablation) of benign thyroid nodules or goiters (with alcohol, high-energy ultrasound, or other modalities) are considered elsewhere.

●(See "Technical aspects of thyroid ultrasound".)

●(See "Thyroid biopsy", section on 'Ultrasound guidance'.)

●(See "Cystic thyroid nodules", section on 'Percutaneous interventional ultrasound-guided therapy'.)

●(See "Treatment of toxic adenoma and toxic multinodular goiter", section on 'Other therapies'.)

GENERAL APPLICATIONS — Ultrasound is the preferred imaging modality for evaluation of the thyroid. Current methods of ultrasonography permit "real-time" identification of structures as small as 2 mm in diameter, thereby allowing the visualization of very small tumors of the thyroid and parathyroid glands. These methods also permit estimates of overall and regional blood flow to the thyroid.

However, the results of thyroid ultrasonography do not correlate perfectly with histopathologic findings. In addition, there is imperfect concordance among experienced thyroid sonographers interpreting thyroid ultrasound images. This interobserver variability may be particularly important when using thyroid ultrasound to monitor growth of thyroid nodules over time. (See 'Monitoring thyroid nodules' below.)

Although the cost effectiveness of ultrasonography in solving patient-specific clinical problems has not been formally tested, ultrasonography, used judiciously, helps to answer important clinical questions in specific patients. Ultrasonography is considered useful in the following situations:

●To evaluate the anatomic features of thyroid nodules

●To assist in fine-needle aspiration (FNA) of thyroid nodules and cervical lymph nodes

●To monitor nodular thyroid disease

●To aid in the differential diagnosis of thyrotoxicosis and thyromegaly

●To assist in the planning of thyroid cancer surgery

●To assist in the surveillance for recurrence in patients with thyroid cancer

●To screen for the presence of thyroid cancer in high-risk groups

●To assess for fetal goiter

●To facilitate certain epidemiologic investigations

THYROID NODULES — High-resolution ultrasonography provides a detailed map of thyroid nodules and helps to characterize the nodules and adjacent structures in the neck. It should be performed in all patients with a suspected thyroid nodule or nodular goiter on physical examination or with nodules incidentally noted on other imaging studies (carotid ultrasound, computed tomography [CT], magnetic resonance imaging [MRI], or fluorodeoxyglucose [FDG]-positron emission tomography [PET] scan). Ultrasound findings can be used to select nodules for fine-needle aspiration (FNA) biopsy. (See "Diagnostic approach to and treatment of thyroid nodules", section on 'Evaluation'.)

Nodule identification and characterization — Thyroid ultrasound is used to identify the location and characteristics of thyroid nodules.

●Solid nodules should be described by using normal thyroid tissue as a reference. They are called isoechoic if their texture closely resembles that of normal thyroid tissue (image 1), hyperechoic if more echogenic (or brighter) (image 2), and hypoechoic if less echogenic (or darker) (image 3). Nodules with an echogenicity that is as dark as or darker than the surrounding strap musculature are termed "markedly hypoechoic" (image 4).

●Simple cysts are ultrasonically characterized as thin walled and spherical, without internal structure. The fluid component of a cyst is anechoic, or black (image 5). Because ultrasound travels very efficiently through fluid, there is minimal attenuation (or loss of the strength) of the ultrasound wave as it travels through the cyst. Consequently, there is an increased echogenicity (or whiteness) behind the structure, which is called posterior acoustic enhancement. The identification of this increased echogenicity posterior to a structure may be a clue that the lesion has increased fluid content. Pure (or simple) cysts are rare and benign, but "cystic" degeneration of previously solid, benign, or malignant nodules is commonly seen.

●Hemorrhage within nodules alters the sonographic appearance. Clot may appear as a hyperechoic density and, after liquefaction, may become hypoechoic (image 6). As a result, in a hemorrhagic nodule, part may be cystic and part solid; this is called a complex nodule.

●Calcifications are brightly reflective echogenic densities that reflect the ultrasound beam. When larger than 1 mm, blockade of the ultrasound signal causes the phenomenon of "shadowing," which renders the tissue deep to the calcification invisible (image 7). By altering the acoustic pathway (or changing the angle of the ultrasound probe), the ultrasonographer can avoid this pitfall and see posterior to the area of calcification.

Air and metallic artifacts (eg, surgical clips) also cause shadowing but are typically linear and fairly straightforward to distinguish from organic calcifications (image 8).

Criteria for identifying cancer — Several sonographic features raise the concern for malignancy in a given nodule. The presence of an individual suspicious feature is insufficient to diagnose a malignancy, and the absence of suspicious features does not exclude malignancy. However, ultrasound features can be used to select nodules for FNA biopsy, for example, in a patient with multiple nodules. (See "Diagnostic approach to and treatment of thyroid nodules", section on 'Fine-needle aspiration biopsy'.)

The individual sonographic features of nodules that have been evaluated for identifying cancers include the following (table 1) [1-3].

●Echogenicity (or intensity of the echoes)

●Microcalcifications (now termed punctate echogenic foci)

●Irregular or infiltrative margins

●Composition (solid or cystic)

●Shape (taller than wide in the transverse view)

●Size

The predictive value of these characteristics varies widely. In a retrospective evaluation of 849 nodules that were diagnosed at surgery or biopsy in 831 patients, the presence of at least one malignant sonographic feature had a sensitivity and specificity of 83 and 74 percent, respectively [4]. The presence of multiple suspicious ultrasound features further increases the likelihood for malignancy [4,5].

In another series of 1244 nodules in 900 patients, nodules were stratified according to ultrasound characteristics that assessed cancer risk on a scale of 1 to 5, with nodules classified as 3.5 or greater evaluated as malignant [6]. Of 233 nodules that were ultrasonographically evaluated as malignant, 76.8 percent were also malignant by cytology (FNA). Of the 142 nodules that were categorized as malignant on ultrasound and were surgically resected, 138 were confirmed by pathology to be thyroid cancer (97 percent positive predictive value for diagnosed cancer). Of 710 nodules evaluated as benign (ultrasound score ≤2.5), 96.1 percent had benign cytology. However, those nodules with borderline features, comprising 17.6 percent of the nodule population studied, had a malignancy risk of 16.8 percent. These findings confirm that nodules on either end of the sonographic spectrum are reliably diagnosed by ultrasound. Those nodules with less distinctive features are more challenging to predict the malignancy risk.

Several sonographic risk stratification systems (SRSS) have been created to categorize thyroid nodules based on their likelihood of malignancy and to select nodules for biopsy. The most widely used systems include the American Thyroid Association (ATA) sonographic system, the American College of Radiology Thyroid Imaging Reporting and Data System (ACR-TIRADS), the American Association of Clinical Endocrinologists/Associazione Medici Endocrinologi (AACE/ACE), the European Thyroid Association TIRADS (EU-TIRADS), and the Korean Society of Thyroid Radiology (K-TIRADS). SSRS are reviewed separately. (See "Diagnostic approach to and treatment of thyroid nodules", section on 'Sonographic criteria for FNA'.)

It is important for the clinician to understand that ultrasonography classification applies directly to the population investigated and only indirectly and imperfectly to a specific patient. Individual demographic features (patient age, comorbid conditions, serum thyroid-stimulating hormone [TSH], family history of thyroid cancer, and radiation exposure) may also impact malignancy risk and should be considered in the evaluation of a nodule in a given patient [7-12]. Existing SRSS do not capture these demographic features in the assessment of an individual patient.

Echogenicity — Malignant thyroid nodules often have a hypoechoic appearance on ultrasound (image 3). In a study of 202 patients with nodules, 62 percent of the cancers were hypoechoic, and few were hyperechoic [13]. In another series of 132 consecutive ultrasound-guided FNA biopsies, none of the 14 cancers were hyperechoic [14]. However, many benign nodules are also less echogenic than surrounding normal thyroid tissue. The sensitivity and specificity of a hypoechoic appearance is approximately 53 and 73 percent, respectively [3].

Nodule margins — The risk of malignancy based on a nodule's margins is difficult to quantify due to the varying terms used to describe nodule margins. Often the terms "blurred" or "ill-defined" are interchanged with "infiltrative" or "irregular" margins. It is important to note, however, that the terms "infiltrative" and "ill-defined" are not synonymous and should not be used interchangeably [15]. There is no uniform verbiage to describe the margins of a nodule that are ascribed to suspicious or benign lesions, though efforts are underway to standardize the ultrasound lexicon [16].

●Ill-defined margins – The sharpness of the border of a nodule probably has little diagnostic value:

•Benign isoechoic nodules may have margins that are ill-defined due to their similarity in echotexture to the surrounding normal thyroid parenchyma (image 9).

•Malignant hypoechoic nodules in a gland affected by Hashimoto's may be described as having poorly defined margins due to its similarity in echotexture to the surrounding gland (image 10).

Ill-defined nodule margins also may be seen in conditions of adenomatous hyperplasia. Thyroid glands with adenomatous hyperplasia have innumerable coalescent nodules with minimal or no intervening normal thyroid stroma. Individual nodules can be difficult to delineate in these settings, yet these nodules harbor a low likelihood of malignancy.

●Infiltrative or irregular margins – Infiltrative or irregular nodule margins have an increased risk of cancer. Similarly, nodules that have invaded into the surrounding perithyroidal tissues are highly specific for malignancy (image 11) [15]. Extrathyroidal invasion of a papillary cancer can be predicted preoperatively (with 75 percent accuracy) when 50 percent or more of the tumor abuts the thyroid capsule [17].

●Halo around a nodule – Some nodules are partly or completely surrounded by a sonolucent ring or halo. A halo only indicates that there is an interface between the thyroid tissue and the nodule that is less echogenic than either the nodule or the gland. The partition could be compressed or atrophic thyroid tissue, local inflammation, or edema. When seen with color-flow Doppler imaging, the halo is often vascular and may represent capsular vessels (image 12). An incomplete or absent halo has been reported as a feature of malignancy but has neither high sensitivity nor specificity (64 and 61 percent, respectively) [3] and is no longer considered in the evaluation of nodules for malignancy stratification.

Calcifications — Calcifications may be present in both benign and malignant nodules and are therefore only partially predictive of histopathology. In one report, the prevalence of cancer was significantly higher when there were calcium flecks in a nodule (29 percent) than in nodules without calcium (14 percent) [18]. Furthermore, some information about a nodule (or lymph node) can be inferred from the nature of the calcification:

●Punctate – Punctate calcifications (microcalcifications) in the solid portions of a nodule in the range of 1 mm may be seen in papillary thyroid carcinoma and, in some cases, correlate with the histologic presence of microscopic psammoma bodies (image 13). Rarely, microcalcifications in the absence of an ultrasonically distinct mass have been associated with papillary thyroid cancer [19]. This may be seen with the diffuse sclerosing variant of papillary thyroid carcinoma, which is frequently associated with underlying Hashimoto's thyroiditis (image 14).

●Eggshell-like – Peripheral or uninterrupted eggshell-like calcification is indicative of chronicity, suggesting a longstanding benign nodule (image 15). However, some cancers, probably those that are chronic and have undergone degenerative change, may demonstrate peripheral calcification, and therefore, diagnostic aspiration biopsy may be appropriate [20]. In addition, the presence of an interruption with soft tissue extrusion is concerning for malignancy [15].

●Coarse – Coarse, scattered calcifications may be seen in benign (usually) or malignant (image 16) nodules.

●Large areas – Large areas of calcification may be seen in medullary cancers of the thyroid [21].

Nodule composition — Nodule composition refers to the degree of solid versus cystic content. The uniformity of the internal structure of a nodule is not a useful indicator for diagnosis of cancer; cancers may be either entirely solid or contain some component of cystic content (this is referred to as a complex nodule). The complex patterns in both benign nodules and cancers may result from cystic or hemorrhagic degeneration or auto-infarction. In general, the larger the cystic component, the less likely the nodule is to be malignant (eg, nodules with >50 percent cystic component are rarely malignant).

There are some ultrasound findings that are useful predictors of benign nodules [22]. The most common is a layered appearance of the echo pattern described as spongiform or honeycomb, which has been defined as an aggregation of microcystic spaces involving at least 50 percent of the nodule (image 17) [4]. This finding has been observed with colloid nodules that are almost invariably benign.

A bright spot with a fading plume tail, like a comet, has also been described with small deposits of colloid (image 18) and must be differentiated from "psammoma body" punctate calcifications (microcalcifications) described above (see 'Calcifications' above). The finding of a comet-tail artifact in a cystic portion of a nodule is almost universally associated with a benign lesion. In contrast, the presence of a comet-tail sign in a solid portion of a nodule may be associated with a benign or malignant neoplasm.

Nodule shape — The natural growth plane of a nodule is in the transverse direction; a nodule that grows against this tissue plane (taller than wide) is concerning for a more aggressive neoplasm (image 19). In some reports, a taller-than-wide shape has a high specificity for malignancy, with rates ranging from 81.5 to 96.6 percent [23,24]. In one study, an anteroposterior and transverse diameter ratio (A/T) >1 was more likely than central blood flow, microcalcifications, or irregular borders to identify cancers [23]. The A/T in combination with any other suspicious ultrasonographic characteristic missed fewer cancers than hypoechogenicity combined with any other suspicious characteristic.

Nodule size — Large nodule diameter and volume predicts a higher likelihood of thyroid cancer and prognosis [1,25-27]. In one large study of over 20,000 thyroid nodules, for every centimeter increase in size of a nodule over 1 cm, there was a 15 to 30 percent escalation in malignancy risk [1].

Color Doppler flow patterns — The pattern of color Doppler flow in a nodule is typically described as intranodular, peripheral, or absent. An internal Doppler flow pattern in a nodule may increase suspicion of malignancy (image 20) [2,3]. However, color Doppler flow cannot reliably distinguish benign from malignant pathology and is not included in the assessment of malignancy risk with most modern sonographic risk stratification systems.

In one study of consecutive patients with nonpalpable thyroid nodules (8 to 15 mm), 31 of 402 thyroid nodules were thyroid malignancies [28]. At ultrasound, an intranodular vascular pattern was present in 74 percent of malignant nodules. In another study of 203 patients with thyroid nodules who were treated surgically, the addition of color-flow Doppler imaging to conventional sonography slightly increased the ability to distinguish malignant thyroid nodules [29]. However, in another study, most thyroid cancers detected by ultrasonography lacked intranodular vascularity, and most hypervascular nodules were adenomas or "adenomatoid" structures that were not tumors [30]. Hypervascularity may be more frequent in medullary and follicular thyroid cancers than in papillary thyroid cancer [31-33].

Ultrasound-guided FNA — Ultrasonography can be employed as a real-time guide to direct the path of a biopsy needle into a thyroid nodule or lymphadenopathy, thus enhancing the diagnostic value of the fine-needle aspiration (FNA) procedure. This topic is reviewed in detail separately. (See "Thyroid biopsy", section on 'Ultrasound guidance'.)

Monitoring thyroid nodules — Ultrasound is useful to monitor cytologically benign thyroid nodules and nodules that initially do not meet indications for FNA biopsy. Small changes in nodule size on serial ultrasonography do not require a repeat aspiration. However, reassessment is warranted when there is substantial growth, defined as more than a 50 percent change in volume or 20 percent increase in nodule diameter with a minimum increase in two or more dimensions of at least 2 mm. However, enlargement alone is poorly predictive of missed malignancy. Instead, the development of suspicious ultrasound findings is associated with a higher likelihood of detecting a malignancy in a thyroid nodule with initially benign cytology [34]. The criteria for repeat aspiration are reviewed in detail elsewhere. (See "Diagnostic approach to and treatment of thyroid nodules", section on 'Benign nodules (Bethesda II)'.)

One must be especially cautious when interpreting reported changes in the volume of nodules based upon ultrasonic volume assessments because it is difficult to reproduce identical two-dimensional image planes for follow-up studies. In addition, thyroid nodule volume determinations vary among experienced thyroid ultrasonographers [35,36]. The accuracy of thyroid volume assessment, interobserver variability, and repeatability appears to be improved with three-dimensional echography. (See "Technical aspects of thyroid ultrasound", section on 'Characterization of thyroid nodules'.)

Nonpalpable nodules ("incidentalomas") — Incidentalomas are nonpalpable thyroid nodules that are detected during imaging procedures for evaluation of nonthyroid pathology (image 21). Nonpalpable nodules have approximately the same risk of malignancy as palpable nodules. Thus, incidentalomas cannot be dismissed as harmless. A dedicated thyroid ultrasound should be performed in all patients with nodules incidentally noted on other imaging studies (carotid ultrasound, CT, MRI, or FDG-PET scan). This topic is reviewed in more detail elsewhere. (See "Diagnostic approach to and treatment of thyroid nodules", section on 'Thyroid incidentalomas'.)

Autonomous thyroid nodules — Autonomous thyroid nodules probably have a long evolution, often cause hyperthyroidism in mid- or later life, and, when they are over several centimeters in size, may undergo hemorrhagic and cystic degeneration [37]. Ultrasonography can detect and quantify the size of the cystic area that appears on a scintiscan as a "cold" area within the "hot" nodule, assuaging clinical concern that the "cold" zone might be a cancer. "Hot" nodules can sometimes be differentiated from "cold" nodules with Doppler scanning because of their more prominent vascular patterns and significantly higher peak systolic velocity values [38].

Ultrasonography versus CT and MRI for thyroid nodules — In general, ultrasonography rather than computed tomography (CT) is performed to evaluate thyroid nodules because of superior resolution, lack of exposure to ionizing irradiation, lower cost, and avoidance of iodinated contrast medium (which can elicit thyrotoxicosis in cases of mild nodular autonomy).

When additional imaging is required for evaluation of the presence and degree of retrosternal extension of a multinodular goiter (particularly if the inferior edge of the gland is not visible with ultrasound), noncontrast CT or magnetic resonance imaging (MRI) are useful. Additionally, cross-sectional imaging may be necessary to assess the integrity of the tracheal lumen, particularly in cases of large multinodular goiters and in the setting of dyspnea. The ability to assess for tracheal narrowing can inform the determination of whether a patient's compressive symptoms are a consequence of the goiter or some other pathology. Pulmonary function testing with flow volume loop study can also aid in the differential diagnosis of respiratory compromise. (See "Clinical presentation and evaluation of goiter in adults", section on 'Goiter with obstructive symptoms or suspected substernal goiter'.)

THYROID CANCER — Ultrasonography plays an important role in the assessment of lymph node status in patients with thyroid nodules or newly diagnosed thyroid cancer, and in the detection of recurrent disease in treated thyroid cancer patients. (See "Diagnostic approach to and treatment of thyroid nodules", section on 'Thyroid ultrasonography' and "Differentiated thyroid cancer: Overview of management".)

Lymphadenopathy

●Role of ultrasound – Ultrasonography is important to confirm or detect cervical lymphadenopathy in a patient with a thyroid nodule or newly diagnosed thyroid cancer.

•Diagnosis of thyroid cancer – Identification of suspicious lymph nodes in a patient with a thyroid nodule increases the likelihood of malignancy. In some cases, fine-needle aspiration (FNA) cytology is obtained from the abnormal cervical lymph node rather than the nodule.

Ultrasound-guided aspiration biopsy of enlarged cervical lymph nodes for cytologic and immunocytologic analysis can differentiate metastases from thyroid cancer and inflammatory lymphadenopathy [39]. It is often diagnostically helpful to rinse the needle that was used to aspirate a suspicious lymph node with a small volume of saline and to assay the washings for thyroglobulin. The presence of high levels of thyroglobulin (or calcitonin in cases of suspected medullary thyroid carcinoma) in needle washings of aspirates of lymph nodes is presumptive evidence of metastatic thyroid cancer despite negative cytology. In contrast, it is not useful to rinse a needle that was employed to aspirate a thyroid nodule, because thyroid tissues should contain thyroglobulin whether or not there is cancer. (See "Thyroid biopsy", section on 'Lymphadenopathy'.)

•Preoperative assessment – For patients requiring thyroidectomy for biopsy-proven papillary thyroid cancer, preoperative neck ultrasound is used to identify suspicious lymph nodes prior to thyroidectomy as metastatic disease is sometimes not clinically apparent to the surgeon intraoperatively. Preoperative cervical ultrasound can detect clinically nonpalpable, metastatic nodes in up to 30 percent of patients with papillary thyroid cancer, which reportedly can result in an altered surgical approach in as many as 40 percent of patients [40-42].

•Monitoring for recurrence – Ultrasonography is also essential to detect recurrence after thyroidectomy for thyroid cancer. Cervical lymph nodes are the most common site of recurrent papillary thyroid cancer.

●Ultrasound appearance – Optimal recognition and diagnosis of adenopathy require a high-resolution ultrasound system equipped with a high-energy linear probe; a 12- to 14-MHz transducer; B-mode and Doppler capability; and a dedicated, specifically skilled operator. Benign lymph nodes tend to be thin (short axis <5 mm) and oval in shape (also called fusiform) and have an echogenic hilum (image 22), whereas malignant ones may have microcalcifications or cystic regions, are "plump" or rounded, lack a defined hilum, and may be intensely vascular (image 23) [43]. Color and power Doppler demonstrate peripheral rather than central vascularity [44,45].

There are sonographic features of adenopathy that have a reasonably high specificity for malignancy but lesser sensitivity [43,46-50]. This was illustrated in a study of 56 lymph nodes (28 benign and 28 malignant) from patients who had a thyroidectomy for cancer and treatment with iodine-131 (I-131) [43]. Of eight sonographic characteristics that were examined for sensitivity and specificity, the following were determined to be major or minor ultrasound criteria of lymph node malignancy:

Major criteria:

•Cystic appearance (100 percent specific but only 11 percent sensitive)

•Bright hyperechoic spots (100 percent specific, 46 percent sensitive)

•Loss of a fatty hilum

•Peripheral vascularization

Minor criteria:

•Round shape

•Hypoechogenicity

•Loss of hyperechoic hilum

In another analysis, central neck location (odds ratio [OR] 4.07, 95% CI 1.64-10.10) and size of lymph node (OR 5.14, 95% CI 1.64-16.06) were significantly associated with the presence of metastatic involvement [50]. The receiver operating characteristic curve for the size of lymph nodes showed a large area under the curve of 0.77 (95% CI 0.68-0.85), and a size of 7.5 mm showed the highest sensitivity and specificity. It is best for the clinician to be guided by several characteristics of a lymph node rather than the single feature.

In older individuals with diabetes and/or obesity, fatty involution of lymph nodes (called lipoplastic lymphadenopathy) may enlarge lymph nodes and mimic a palpable thyroid metastasis, which may confuse ultrasonic diagnosis (image 24) [51]. In children and adolescents, inflammatory lymph nodes are common, whereas malignant lymph nodes are uncommon, so nodes must be conservatively interpreted in this age group even when a thyroid nodule is present.

Thyroid cancer follow-up — Ultrasonography is the most frequently used imaging procedure for following patients with thyroid cancer [52,53]. The procedure can be performed without discontinuing levothyroxine therapy, thereby avoiding the inconvenience and risks of hypothyroidism, and without administering recombinant thyrotropin, either of which are needed for radionuclide imaging. (See "Differentiated thyroid cancer: Overview of management", section on 'Imaging'.)

●Detection of recurrent thyroid cancer – Thyroid ultrasound can detect recurrent cancer in the thyroid bed after total thyroidectomy or lobectomy, or in the contralateral lobe after lobectomy.

Ultrasound examination enhances the sensitivity for detecting recurrent thyroid cancer, which is achieved by the tumor marker thyroglobulin. For example, 340 consecutive patients with differentiated thyroid cancer who had been treated with near-total thyroidectomy, 131-I thyroid ablation, and TSH suppressive doses of levothyroxine, were given recombinant human TSH (rhTSH) prior to whole body scan. The diagnostic whole body scan had a cancer-diagnostic sensitivity of 85 percent and a negative predictive value of 98.2 percent. After adding the results of neck ultrasound, the sensitivity increased to 96.3 percent and the negative predictive value to 99.5 percent [54].

●Identification of lymph node metastases – In addition, ultrasonography of the neck may identify lymph node metastases from differentiated thyroid cancer before they have grown sufficiently large to become palpable (image 23), even in patients with low or undetectable serum thyroglobulin levels and negative I-131 whole body scintiscan performed after rhTSH stimulation [55]. Furthermore, in a five-year observational study of 80 consecutive cases of papillary thyroid microcarcinoma, ultrasonography was highly sensitive in detecting nodal metastases in patients who had a near-total thyroidectomy but no I-131 therapy [56].

●Important caveats – Sonography during the initial several months after surgery for thyroid cancer may give misleading results. During this time, there may be abundant noncancerous, enlarged lymph nodes and edematous or inflammatory postoperative changes that appear as heterogeneous and ultrasound-dense focal structures. These findings should not be confused with tumor and can be avoided by delaying the postoperative examination for three or more months after the surgery.

There are other sources of sonographic false-positive evidence of recurrence. In one report, a suture granuloma (confirmed by ultrasound-guided FNA biopsy) mimicked cancer on thyroid ultrasound and 2-[fluorine-18] fluoro-2-deoxy-D-glucose positron emission tomography (PET) [57]. On ultrasound, the suture granuloma demonstrates central or paracentral clustering of the internal echogenic foci and a paired appearance of echogenic foci that usually exceeds a diameter of 1 mm (image 25) [58].

Traumatic neuroma may also be confused with a malignant lymph node. These neuromas may develop in patients who have had a lateral neck dissection. They are typically fusiform in shape, ill-defined, heterogeneous in echogenicity, and lack Doppler flow [59]. In some lesions, the nerve may be visualized entering into the mass (image 26). Insertion of a needle into the lesion for FNA typically elicits sharp, radiating pain.

Screening for thyroid cancer — Routine screening of a population of individuals at low risk for thyroid cancer does not appear to be cost effective. In one study, the detection rate for thyroid cancer was 2.6 percent for all subjects among 1401 women who were scheduled to undergo either a breast examination or a follow-up examination for breast cancer and who had thyroid ultrasonography [60].

Furthermore, screening ultrasounds of the neck during health maintenance examinations in South Korea resulted in thyroid cancer becoming the most common malignancy in females [61]. The survival from thyroid cancer during this campaign did not change, even with the increased detection of small cancers. These findings, coupled with a high rate of occult thyroid malignancies in autopsy series [62,63], has resulted in a call to end the practice of screening neck ultrasound examinations in South Korea. Several years after the cessation of the screening ultrasound program, the rates of thyroid cancer detection have declined and thyroid cancer related mortality remains stable [64].

Ultrasound screening for thyroid cancer may be useful in a population with a higher than average risk of cancer (eg, individuals with a childhood history of therapeutic radiation exposure to the neck region or environmental radiation exposure). Screening in these populations is controversial since it has not been shown to improve outcomes [65]. (See "Radiation-induced thyroid disease", section on 'Surveillance for structural thyroid abnormalities'.)

GOITER — In healthy adults without iodine deficiency, a normal thyroid lobe is approximately 4 to 4.8 by 1 to 1.8 by 0.8 to 1.6 cm in size, with a mean total volume for both lobes of 8 to 10 mL (range 3 to 20 mL) [66,67]. Thyroid enlargement above these approximate normal measurements is considered a goiter.

Thyroid enlargement, like thyroid nodules, is detected much more often by ultrasonography than physical examination. In an individual patient with a palpable diffuse goiter, ultrasonography may add several types of information. (See "Clinical presentation and evaluation of goiter in adults", section on 'Approach to evaluation'.)

●It may identify distinctive, nonpalpable thyroid nodules within a nodular or diffuse goiter. Nodules with suspicious ultrasound features should be considered for biopsy because the prevalence of cancer in an individual nodule in a goiter is independent of the number of sonographically identified nodules [68,69].

●It may be useful in evaluating a region of a goiter with unusual characteristics on palpation, such as hard consistency or tenderness, findings suggestive of possible coexisting adenoma, cancer, or lymphoma [70]. Screening of goiters for ultrasonic characteristics, such as enhanced posterior echoes, has been reported to contribute to early diagnosis of lymphoma, thereby facilitating early initiation of therapy [71].

●In a situation where a medical treatment may be associated with goiter, ultrasonography is superior to palpation in detecting enlargement of the thyroid. For instance, in a cross-sectional study of patients undergoing long-term treatment with lithium for bipolar, major depressive, and schizoaffective disease, total thyroid volume was significantly greater (23.7 versus 13.6 mL) in the lithium-treated group (n = 96) than among controls (96 sex- and age-matched control subjects). Ultrasonography detected enlargement significantly more often than palpation [72]. (See "Lithium and the thyroid".)

●The ultrasonic texture of a goiter and Doppler dynamics may yield information about diagnosis, function, and possibly even prognosis and medical management. (See 'Autoimmune thyroid disease' below.)

OTHER

Autoimmune thyroid disease

●Low echogenicity – Thyroid glands in patients with chronic autoimmune thyroiditis (Hashimoto's disease) tend to be diffusely enlarged and have low echogenicity (image 27) [73-75]. Diffusely low echogenicity may also be seen in patients with Graves' hyperthyroidism (image 28) and in patients with various forms of thyroiditis [76,77]. In an investigation of 485 patients with diffuse thyroid hypoechogenicity compared with 100 patients with normal echogenicity, the positive and negative predictive values of hypoechogenicity as an indicator of autoimmune thyroid disease were 88 and 95 percent, respectively [76].

●Color-flow Doppler – Color-flow Doppler imaging may offer additional clinical information [78-80]. Among 55 patients with hyperthyroidism (Graves' disease or toxic adenoma), 24 with Hashimoto's thyroiditis, and 39 euthyroid controls, color-flow Doppler examination differentiated untreated Graves' disease (high blood flow) from Hashimoto's thyroiditis (low blood flow) (image 29), whereas conventional gray-scale findings were similar [38]. In a similar study of 114 patients with hyperthyroidism (56 Graves' disease, 28 painless thyroiditis, 30 subacute thyroiditis), color-flow Doppler could distinguish Graves' disease (high blood flow) from thyroiditis (normal blood flow) and total blood flow values significantly correlated with radioactive iodine uptake [81]. Additionally, the restoration of euthyroidism with a thionamide or radioiodine may be associated with a reduction in the degree of color Doppler flow and thyroid volume as measured by ultrasound [82,83].

Although these ultrasound findings are interesting, they are less clinically useful in the diagnosis of Graves' disease or Hashimoto's thyroiditis in comparison with checking thyroid function tests, thyroid peroxidase antibodies (to diagnose Hashimoto's thyroiditis), thyrotropin receptor antibodies (thyrotropin-stimulating immunoglobulin [TSI] or TRAb, to diagnose Graves' disease), or a radioiodine uptake and scan (to distinguish Graves' from painless thyroiditis in nonpregnant individuals). In circumstances where radioiodine studies are contraindicated (pregnancy, breastfeeding) and Graves' disease and thyroiditis cannot be differentiated by clinical findings, ultrasound with Doppler imaging may be clinically useful. (See "Diagnosis of and screening for hypothyroidism in nonpregnant adults" and "Postpartum thyroiditis", section on 'Differential diagnosis' and "Diagnosis of hyperthyroidism", section on 'Radioiodine uptake' and "Diagnosis of hyperthyroidism", section on 'Determining the etiology'.)

●Nodular Hashimoto's thyroiditis – A micronodular pattern on ultrasound is often noted in the thyroid glands of patients with Hashimoto's thyroiditis (image 30). Discrete nodules may also be present (image 31). The sonographic features of nodular Hashimoto's thyroiditis are variable, although nodules are more often solid and often hypoechoic [84]. On Doppler analysis, nodules may be hyper-, hypo-, or isovascular. In an analysis of 78 patients with 82 nodules within glands with diffuse Hashimoto's thyroiditis, 84 percent were benign and 16 percent were malignant (12 papillary carcinoma and one lymphoma) [85]. The ultrasound characteristics of malignant and benign nodules within diffuse Hashimoto's thyroiditis were similar to the characteristics of nodules in the general population (see 'Criteria for identifying cancer' above). Thus, the decision to perform fine-needle aspiration (FNA) biopsy in a patient with Hashimoto's thyroiditis is based upon the same criteria as those used in the general population with thyroid nodules.

Amiodarone-induced thyrotoxicosis — Doppler flow sonography has been reported to differentiate the two types of amiodarone-induced hyperthyroidism. In several studies, patients with hyperthyroidism who had underlying thyroid disease (eg, nodular goiter) had increased vascularity (amiodarone-induced thyrotoxicosis type 1), while patients with thyrotoxicosis due to drug-related thyroiditis who had no apparent underlying thyroid disease showed low or absent vascular flow (amiodarone-induced thyrotoxicosis type 2) [86-89]. This topic is reviewed separately. (See "Amiodarone and thyroid dysfunction", section on 'Differentiating the two types'.)

Thyroid lymphoma — Thyroid lymphoma has a hypoechogenic appearance that is difficult to differentiate from chronic thyroiditis (image 32 and image 33). Three patterns have been reported according to internal echoes within the suspected lesion, the border of the lesion, and the intensity of the echoes behind the lesion (due to enhanced transmission of the signal through uniform tissue that minimally reflects ultrasound) [90].

●Nodular lymphoma reportedly has a very uniform and hypoechoic pattern that may be sufficiently hypoechoic to be pseudo-cystic. The border between lymphoma and non-lymphomatous tissue is well defined, and the borderline may be "broccoli-like" or "coastline-like" irregular.

●The diffuse type of lymphoma has an indistinct border between lymphoma and non-lymphomatous thyroid gland and otherwise looked like goiter, but the internal echoes were of exceedingly low intensity.

●The mixed-type lymphoma shows multiple, patchy hypoechoic lesions, each with enhanced posterior echoes.

Subacute granulomatous thyroiditis — Though the clinical presentation makes the diagnosis of subacute granulomatous (deQuervain's) thyroiditis fairly straightforward, sonography can aid in the diagnosis. In the acute phase of subacute granulomatous thyroiditis, ultrasound shows a pattern of ill-defined, migratory hypoechogenicity with minimal vascularity. The gray-scale image returns to a normal pattern with resolution of the illness (image 34) [91-95]. (See "Subacute thyroiditis".)

Nonthyroid and ectopic thyroid cervical lesions — Ultrasonography can distinguish and sometimes identify disorders in the neck that are not thyroid in origin. Examples include parathyroid adenomas or cysts, subcutaneous lesions, schwannomas, paragangliomas, vascular abnormalities, or branchial cleft cysts. It can also be useful in diagnosing and postoperatively evaluating patients with thyroid cancer of a thyroglossal duct [96]. Esophageal tumors and diverticula may be misidentified as a thyroid mass based on sonography (image 35) [97]. This lesion may be better localized by asking the patient to swallow during sonography. The identification of a "flash" of saliva within the lesion with swallowing is indicative of esophageal tissue.

Prior to parathyroid surgery, thyroid ultrasonography is useful and may be cost effective to evaluate the thyroid gland for lesions that may need surgical intervention during parathyroidectomy. Thyroid nodules are also a potential source of a false-positive sestamibi study for the preoperative localization of parathyroid adenomas. In a retrospective study of 1200 consecutive patients who were treated surgically for primary hyperparathyroidism with bilateral neck exploration, routine preoperative neck ultrasound identified coexisting thyroid disease in 43 percent (150 of 350 patients) [98]. Thyroid cancer was found in 4.6 percent of the patients and was treated with total thyroidectomy at the time of the parathyroid surgery, avoiding the need for a second operation. Preoperative fine-needle biopsy of sonographically detected thyroid nodules was performed in 20 percent, which was reportedly cost effective in limiting concomitant thyroid surgery to fewer patients (6 percent, 21 of 350 patients). Nearly one-half of thyroid pathology was colloid nodules/goiters, followed by follicular adenomas, papillary cancer, thyroiditis, and intrathyroidal parathyroid.

Tuberculous thyroiditis — Primary tuberculosis of the thyroid is extremely rare, estimated to occur in 0.1 to 0.4 percent of all tuberculous cases [99]. In a retrospective review of all described cases in the literature, sonographic findings were nonspecific [100]. Findings may include a heterogeneous hypoechoic lesion in the thyroid or, in the setting of abscess formation, the area may be anechoic. Locoregional adenopathy or an enlarging, and fluctuant thyroid may be noted. In one study, FNA cytology revealed a small number of epithelioid histiocytes in a necrotic background, which was suggestive of tuberculosis [101]. Sonography after three months of antituberculosis chemotherapy showed that the thyroid lesions had resolved.

Intraoperative thyroid ultrasonography — Intraoperative ultrasound of the neck performed by the surgeon influences the operative approach [102]. In one study, surgeons identified suspicious lesions in the central and lateral neck that were not identified by the radiologist in nearly 30 percent of cases [103]. While most of these ultrasound examinations are performed in the clinic setting, some surgeons also perform a sonographic evaluation before and after surgery to carefully examine for missed metastatic nodes.

ULTRASONOGRAPHY OF THE THYROID IN THE FETUS AND NEONATE

Fetal thyroid gland — Ultrasonography is useful to assess the fetal thyroid gland, diagnose fetal goiter or thyroid dysfunction, and to facilitate therapy. It can reduce obstetric complications and contribute to neonatal health. (See "Overview of thyroid disease and pregnancy" and "Hyperthyroidism during pregnancy: Treatment", section on 'Fetal monitoring'.)

●Normative data – Normative data (gestational age-dependent and age-independent) were collected in a study of sonographic images of the fetal thyroid, obtained from 200 fetuses between 16 and 37 weeks of pregnancy [104].

●Maternal Graves' disease – In mothers with Graves' disease, ultrasonography of the fetal thyroid gland by an experienced ultrasonographer is an excellent diagnostic adjunct that can facilitate assessment of fetal thyroid function [105]. Used in conjunction with clinical features, such as fetal tachycardia, intrauterine growth retardation, and occasionally cord blood sampling, findings of fetal goiter may lead to the diagnosis of fetal thyrotoxicosis. Conversely, fetal goiter without clinical manifestations of fetal thyrotoxicosis may suggest overtreatment of the mother with antithyroid drugs and result in a dose reduction [106].

●Fetal goitrous hypothyroidism – Ultrasonography can also be used to identify and facilitate treatment of fetal goitrous hypothyroidism. In one case report, a diagnosis of fetal goitrous hypothyroidism associated with high-output cardiac failure was made at 32 weeks of gestation using maternal ultrasound examination, amniocentesis, and cordocentesis. The fetus was treated by injection of levothyroxine sodium into the amniotic fluid at 33 weeks of gestation and the goiter decreased in size, the cardiac failure improved, and fetal TSH was reduced [107]. In a second case report, routine ultrasound examination of a mother revealed a 29-week fetus with a goiter [108]. Fetal blood sampling showed a high TSH, indicative of fetal hypothyroidism. Treatment was given between 31 and 36 weeks with intra-amniotic injections of triiodothyronine and subsequently with thyroxine. Thereafter, the fetal goiter became smaller, neck flexion increased, and polyhydramnios resolved. Following birth, neonatal serum TSH was within the normal range [108].

Neonates and infants — Ultrasonography has been applied to assess the thyroid in neonates and infants. Normative data were collected in a study of 100 (49 male) healthy term Scottish neonates to ascertain the length, breadth, and depth of each thyroid lobe and calculated the volume of each lobe using the formula for a prolate ellipsoid. There was no difference in mean volume between the right and left lobes, but there was considerable variation between the two lobes in individual infants. A wider geographic sample with attention to ambient iodine intake will be required before these data can be applied globally [109].

Another useful method for assessing the size of the thyroid in infants and small children is to determine the ratio of the maximum width of both thyroid lobes to the width of the trachea [110].

Sonography can be used as the initial imaging tool in neonates and infants with congenital hypothyroidism, but scintigraphy should be used to distinguish agenesis from ectopia. In one study, ultrasonography identified 66 instances where the thyroid gland was not located in the usual anatomical position and one case of hemiagenesis. The diagnoses were confirmed by scintigraphy [111]. (See "Clinical features and detection of congenital hypothyroidism", section on 'Thyroid imaging'.)

EPIDEMIOLOGIC USE — Thyroid anatomy and size in iodine-deficient or radiation-exposed populations can be evaluated by ultrasonography. While interobserver agreement on estimates of thyroid volume has been good in epidemiologic studies, agreement on echogenicity has been poor [112].

In children, when thyroid volumes are to be compared with reference values, assessment based upon age is reportedly the most reliable method, providing there is normal somatic development [113]. Furthermore, standards for thyroid volume and Doppler data that are specific to a geographic region, iodine status, sex, and pubertal stage may be more appropriate than a single age-specific international reference [114,115].

Systematic screening with ultrasound has been reported as useful for early detection of thyroid cancer in the population of Belarus that were exposed to radiation due to the Chernobyl accident [116].

Other examples of epidemiologic applications of thyroid sonography include detection of thyroid cancers in populations with endemic goiter [117] and the demonstration that the prevalence of thyroid cancer has not increased in the population exposed to the accidental release of iodine-131 (I-131) in Hanford, Washington during 1944 to 1957 [118].

SUMMARY AND RECOMMENDATIONS

●General applications – Ultrasound is the preferred imaging modality for evaluation of the thyroid. Ultrasonography permits "real-time" identification of structures as small as 2 mm in diameter, thereby allowing the visualization of very small tumors of the thyroid and parathyroid glands. (See 'General applications' above.)

●Thyroid nodules – Thyroid ultrasound should be performed in all patients with a suspected thyroid nodule or nodular goiter on physical examination or with nodules incidentally noted on other imaging studies (carotid ultrasound, computed tomography [CT], magnetic resonance imaging [MRI], or fluorodeoxyglucose [FDG]-positron emission tomography [PET] scan). (See 'Thyroid nodules' above.)

●Nodule characterization

•Echogenicity – Solid nodules should be described by using normal thyroid tissue as a reference. They are called isoechoic if their texture closely resembles that of normal thyroid tissue (image 1), hyperechoic if more echogenic (or brighter) (image 2), and hypoechoic if less echogenic (or darker) (image 4). (See 'Nodule identification and characterization' above.)

•Suspicious findings – Several sonographic features raise the concern for malignancy in a given nodule (table 1); those with the highest clinical utility include hypoechogenicity (image 3), punctate echogenic foci (microcalcifications) (image 13), taller-than-wide shape (image 19), and infiltrative margins (especially when they are focal) (image 11). The absence of any suspicious features does not exclude malignancy. (See 'Criteria for identifying cancer' above.)

•Selecting nodules for aspiration – Ultrasound cannot be used to definitively diagnose thyroid cancer or select patients for thyroid surgery. However, ultrasound findings offer important clues and can be used to select nodules for fine-needle aspiration (FNA) biopsy, which offers the best preoperative identification of thyroid cancer. (See 'Criteria for identifying cancer' above.)

●FNA biopsy – Ultrasonography can be employed as a real-time guide to direct the path of a biopsy needle into a thyroid nodule or lymphadenopathy, thus enhancing the diagnostic value of the procedure. (See "Thyroid biopsy", section on 'Ultrasound guidance'.)

●Thyroid cancer – Ultrasonography plays an important role in the assessment of lymph node status in patients with thyroid nodules or newly diagnosed thyroid cancer and in the detection of recurrent disease in treated thyroid cancer patients (image 23). (See 'Thyroid cancer' above and "Diagnostic approach to and treatment of thyroid nodules", section on 'Thyroid ultrasonography' and "Differentiated thyroid cancer: Overview of management".)

●Evaluation of goiter – In healthy adults without iodine deficiency, a normal thyroid lobe is approximately 4 to 4.8 by 1 to 1.8 by 0.8 to 1.6 cm in size, with a mean total volume for both lobes of 8 to 10 mL (range 3 to 20 mL). Thyroid enlargement above these approximate normal measurements is considered a goiter. In an individual patient with a palpable diffuse goiter, ultrasonography may add several pieces of information, including quantification of the goiter volume and the identification of distinctive, nonpalpable thyroid nodules within a nodular or diffuse goiter, which may require FNA. (See 'Goiter' above.)

●Fetal and neonatal thyroid evaluation – Ultrasonography can be used to identify and facilitate treatment of fetal goitrous hypothyroidism and fetal thyrotoxicosis. It can be used as the initial imaging tool in neonates and infants with congenital hypothyroidism. (See 'Ultrasonography of the thyroid in the fetus and neonate' above and "Clinical features and detection of congenital hypothyroidism", section on 'Thyroid imaging'.)

ACKNOWLEDGMENT — The UpToDate editorial staff acknowledges Manfred Blum, MD, FACP (deceased), who contributed to earlier versions of this topic review.