The Thyroid Nodule

The terms cold and hot are commonly used to describe the functional activity of thyroid nodules as revealed by scintigraphy. These descriptors refer to the apparent amount of radionuclide in the lesion relative to that in surrounding normal thyroid tissue. Until recently, this distinction was diagnostically important to identify thyroid cancer. Nearly all malignant tumors in the thyroid concentrate less radioiodine or ^99mTc than the normal gland and therefore appear cold (hypofunctional). However, since many benign tumors and nontumorous nodules are also cold, that characteristic is now of limited clinical use because of the advent of fine-needle biopsy (FNB). Now “hot versus cold” is important in the choice not to do an FNB in patients with a nodule when TSH is low and the nodule is hot; hot nodules are very rarely malignant, and the biopsy may be misleading by incorrectly suggesting malignancy (Fig. 7-2).

The term warm nodule is ambiguous and not helpful in assessing the risk of cancer. The term should not be used. A nodule that is not clearly delineated on a scintiscan may in fact be a cold nodule that is too small or too close to normal activity to be distinguishable from surrounding normal thyroid tissue. In general, a cold nodule must be almost 1 cm or larger in diameter to be detected by pinhole imaging.

Because the great majority of nodules are cold on radioisotope scintiscanning, and such cold nodules may be either benign or malignant, it is not cost effective to obtain a scintiscan as the initial diagnostic test in a euthyroid (normal TSH) patient with a nodule. A fine-needle aspiration (FNA) biopsy is usually performed first. However, when the FNA result indicates sheets of follicular cells that do not look malignant but could be a tumor, scintigraphy may be helpful to identify a hot nodule, thus avoiding surgery.

There are two kinds of hot nodules. The majority of them are reactive to elevated TSH in an otherwise failing thyroid gland and thus are compensatory. These are called hyperplastic nodules. The hot nodules associated with low TSH are autonomous nodules. They function autonomously because the TSH receptor of the thyroid cell has mutated and does not require TSH to stimulate cell activities such as the production of thyroxine or cell replication and nodule growth. The clone of the mutated cells is a benign tumor, which is the nodule, and may grow large enough over time to lead to hyperthyroidism (“toxic” nodule, TAN). When TAN-produced thyroid hormone levels increase, TSH falls, leading to suppression of the normal paranodular tissue in both thyroid lobes.⁵ The chance of malignancy in a hot nodule is less than 1%.¹ Therefore, when TSH is undetectable in a patient with a thyroid nodule, a scintiscan may be the next appropriate test. In some cases, a hyperfunctioning nodule undergoes degeneration or hemorrhage, and therefore part of it, or uncommonly all of it, becomes hypofunctioning or cold. Some of the reported instances of cold areas in a hot nodule are in fact cases of small, coexisting carcinomas in close proximity to a larger, benign hot nodule. Although documented cases have been reported in which the entire hot nodule is malignant, these cases are quite rare.

Occasionally, follicular neoplasms (including follicular adenomas and even some carcinomas) may appear hot on a ^99mTc-pertechnetate image but cold on a radioiodine image, presumably because such tumors are able to trap but do not organify iodine. This type of discordance between ^99mTc and radioiodine images occurs infrequently and must be recognized when ^99mTc is used for thyroid scintiscanning.5–7

In the past, one spoke of a solitary thyroid nodule. We assumed incorrectly that if the rest of the thyroid gland was normal to palpation, it was free of nodules. Examination by ultrasound and correlation with pathology have disclosed that nonpalpable nodules occur commonly when a dominant nodule has been palpated.

Multinodular Goiter

Scintiscanning by itself does not reveal the etiology of a multinodular goiter. Scintiscanning has a role in the diagnostic assessment of multinodular goiter in some patients. When the nodules are discrete and larger than 1 cm in diameter, the scintiscan reveals the functional activity of a particular nodule relative to that of paranodular tissue (Fig. 7-3). When a nodule is clinically dominant—that is, a nodule that on palpation or by ultrasonography is different from the other nodules or is growing faster—the finding on scintiscan that this nodule contains all or most of the functional activity helps to guide management. Such a functioning nodule is unlikely to be malignant. When a multinodular goiter is large and causes symptoms and signs of compression of the trachea or esophagus, scintiscanning may complement other imaging modalities such as MRI or sonography. Either of the latter two methods depicts the extent of the goitrous mass, but only scintiscanning can reveal functioning tissue, information that is often helpful in making therapeutic decisions such as surgery or ¹³¹I therapy.

In a hyperthyroid patient with a nodular goiter, there may be several different patterns on scintigraphy. The image may show many functioning, widely dispersed nodules, generally smaller than 0.5 cm, that have a salt-and-pepper appearance; a coarsely patchy distribution may be evident, or a uniform pattern of uptake not unlike that of Graves’ disease can be seen. In contrast, there may be a cold zone correlating with a nodule that should be evaluated by other means, such as sonography and aspiration biopsy.

In a hypothyroid patient with nodular goiter, the uptake is usually low (unless there is iodine deficiency or dyshormonogenesis), and thus the scintigraphic image shows a “patchy” distribution of radioisotope because of low “counts” that result in poor scan statistics. However, scintigraphy may be of clinical value to identify a reactive, hyperplastic, functional area that is rarely malignant.

Diffuse Goiter Including Graves’ Disease

Assessing the thyroid uptake of radioiodine is diagnostically important to differentiate the low accumulation of the tracer in silent thyroiditis and its high accretion in hyperthyroid Graves’ disease and for precise dosimetry of ¹³¹I therapy. However, thyroid scintiscanning usually is not clinically useful in diffuse goiters. If assessing size is desired, an ultrasonogram offers superior results. Yet, some clinicians order scintiscans for patients with Graves’ disease.

The pattern of isotope accumulation in Graves’ disease usually is uniform, but irregular distribution may occur, especially in a geographic region where goiter is endemic. The heterogeneity should not be misconstrued as reflecting toxic nodular goiter when the clinical setting and serologic tests are more consistent with Graves’ disease. At times, a patient with Graves’ disease may have a palpable thyroid nodule or an unusually firm area, or a nonisotopic neck-imaging procedure might have disclosed an incidentaloma (discussed later) that leads to scintiscanning to assess whether it is cold, in which case sonography and aspiration biopsy might be appropriate. Similarly, a “routine” scintiscan might have revealed a cold area that should be investigated.

Substernal Goiter

A substernal goiter is suspected when chest radiographs or CT reveal an anterior mediastinal mass. A positive radioiodine scintigram definitively identifies such a mass as thyroid tissue. ¹²³I or ¹³¹I is preferred over ^99mTc in such cases because of interference by circulating ^99mTc in the mediastinal blood vessels. SPECT (using ¹²³I) is of particular value in visualizing substernal thyroid tissue and SPECT/CT or MR can precisely localize it to an anatomical structure.

Scintigraphy in Patients With Thyroid Carcinoma

In general, thyroid carcinoma takes up and retains radioiodine much less efficiently than normal thyroid tissue does. Therefore, most thyroid cancers are cold nodules. However, after the thyroid gland has been removed surgically and under the stimulation of elevated TSH levels and iodine deprivation, the uptake in differentiated thyroid carcinoma is often sufficiently high to permit detection by scintigraphy and therapy with ¹³¹I. Thus diagnostic imaging with radioiodine is indicated for patients who have recently undergone thyroid surgery for differentiated thyroid carcinoma. The rationale for imaging in such patients is to identify and quantify uptake in the thyroid remnant and to detect any functioning metastases in the neck or in distant sites. This information helps to determine the amount of ¹³¹I that should be administered to ablate the normal remnant and treat functioning metastases.

To stimulate uptake of radioiodine by malignant thyroid tissue, it is necessary to raise serum TSH levels to 30 mU/L or higher.⁷ In routine practice, levothyroxine therapy is discontinued for 5 to 6 weeks before administration of the diagnostic dose of radioiodine for imaging. To shorten the period of hypothyroidism, liothyronine (triiodothyronine [T₃]) is often given (25 to 50 µg/day in divided doses) for 3 weeks after discontinuation of levothyroxine therapy. T₃ therapy is discontinued 2 weeks before radioiodine administration. Alternatively, it is possible to prepare patients by giving their usual thyroxine dose every other day for 6 weeks, which induces mild hypothyroidism and raises the level of TSH to about 50 mIU/L.

Recently, recombinant human TSH (rhTSH, Thyrogen, Genzyme Corp., Cambridge, MA) has been developed and in clinical trials has been shown to stimulate uptake of ¹³¹I by thyroid remnants and metastases and to raise serum thyroglobulin (Tg) levels in patients who remain euthyroid while taking thyroid hormone therapy.8,9 The Food and Drug Administration has approved the use of rhTSH for radioiodine imaging and serum Tg testing in such patients and for the destruction (ablation) of thyroid tissue that remains after a total thyroidectomy has been performed because of thyroid cancer.

Patients who are scheduled to undergo diagnostic whole-body scintigraphy are advised to follow a low-iodine diet for at least 7 to 10 days before administration of the radioiodine. A simple low-iodine diet has been described.¹⁰ It is even more important to avoid iodine-containing medications and radiographic contrast agents.

Diagnostic administration of ¹³¹I may reduce the uptake of subsequent therapeutic ¹³¹I by normal thyroid remnants or functioning metastases. This phenomenon, which has been termed stunning, seems to involve a sublethal, presumably temporary suppression of iodine uptake (reviewed elsewhere11,12). To avoid stunning, many authors recommend limiting the quantity of ¹³¹I given for diagnostic imaging to 2 mCi (74 MBq)^11,13 or even less.¹⁴ An alternative is to use ¹²³I for whole-body imaging,¹⁵ but the present cost of this radionuclide in millicurie amounts is prohibitive for many centers. While ¹²³I offers superior imaging of deposits of thyroid tumor in the neck when compared to ¹³¹I, its efficiency for deposits deep in the rest of the body remains to be established convincingly.

Physicians who interpret whole-body radioiodine scintigrams must be familiar with the distribution of inorganic iodine in the blood pool and extracellular fluid and its dynamics over time, the normal, nonthyroid sites that accumulate the radioactive iodine tracer, and the behavior of isotope-labeled thyroxine that has been produced in thyroid gland or metastases. The salivary glands, gastric mucosa, kidneys, and lactating breasts concentrate iodide but do not convert it to thyroxine. Nasal secretions, saliva, sweat, urine, stool, and milk may contain high concentrations of inorganic radioiodine and can cause artifacts, depending on the time after isotope administration. Skin and hair are easily contaminated with saliva, urine, or vomitus. Radioiodine-labeled thyroxine is observed later than inorganic iodide and has a different pattern of distribution, including the liver as part of the enterohepatic circulation of thyroxine. Nonthyroid tumors, inflammatory lesions, and cysts may occasionally contain radioiodine as part of their vasculature and result in a false-positive scintigram.11,16

In interpreting ¹³¹I planar whole body scans, there may be atypical or cryptic findings. ¹³¹I SPECT/CT is useful to characterize challenging foci more accurately, differentiate thyroid cancer from physiologic activity or nonthyroid pathology, and reduce false positives. By identifying nonthyroid cancer activity, SPECT/CT can reduce inappropriate treatment with ¹³¹I.

¹³¹I scanning is not 100% sensitive for metastatic thyroid carcinoma. In some series, the rate of false-negative radioiodine scans approaches 35%.¹⁷ When diagnostic ¹³¹I scans are negative and metastases are suspected on the basis of elevated serum Tg levels, some advocate empirical treatment with ¹³¹I. Scans that are done 1 week after ¹³¹I therapy are frequently positive in these cases (reviewed by Clark and Hoelting¹⁸) (Fig. 7-4). Indeed, a routine pre-ablation scan after thyroidectomy for thyroid cancer may not be cost effective. Therefore, many centers employ a protocol of administering a standard therapeutic dose of ¹³¹I without prescanning and perform the whole body scan a week later. This protocol is limited by an inability to do dosimetry for the treatment, which some authorities deem essential, and fails to assess for undetected metastases in the head or spine, which may swell after ¹³¹I and cause neurologic complications.

It is useful in ¹³¹I scan-negative, Tg-positive patients to search for metastases by performing sonography of the neck and, if negative, performing MRI of the neck and chest. Especially if ¹³¹I therapy is anticipated, contrast CT studies are not employed, but noncontrast CT may suffice. The aim is to find metastases that either are surgically accessible or can be treated by external radiotherapy, if appropriate. When the above imaging techniques do not reveal a source of the elevated Tg, scintiscanning with other radiolabeled agents or fluoro-deoxyglucose (FDG)-PET scanning may succeed in localizing ¹³¹I-negative metastases¹⁹ (see Table 7-2).

Thallium (²⁰¹Tl) has been useful in localizing metastases in selected patients (reviewed by Cavalieri¹² and Sisson¹⁹). However, ²⁰¹Tl is concentrated by a variety of benign and malignant lesions other than thyroid carcinoma.

^99mTc-sestamibi (MIBI) is a cationic, lipophilic agent that concentrates in normal and neoplastic thyroid tissue and in a variety of other cancers. Experience indicates that like ²⁰¹Tl, ^99mTc-MIBI can be useful in ¹³¹I-negative patients in whom one has reason to suspect persistent or recurrent tumor.19,20 ^99mTc-MIBI is the agent of choice for imaging Hürthle cell carcinoma, which typically takes up radioiodine poorly.²¹

Other agents that have been shown to concentrate in some metastatic differentiated thyroid tumors are ^99mTc-tetrofosmin, which like MIBI is a myocardial perfusion imaging agent, and ^99mTc-labeled dimercaptosuccinic acid in the pentavalent form (DMSA [V]) (reviewed by Sisson¹⁹). Clinical experience with these agents is still limited. Indium-111-labeled octreotide is used to localize metastatic medullary thyroid carcinoma.²²

¹⁸Fluoro-2-Deoxyglucose Positron Emission Scanning

¹⁸F-fluoro-2-deoxyglucose (¹⁸FDG), a radiolabeled analogue of glucose, is actively concentrated in a variety of malignant tumors, including thyroid carcinoma.^{23 18}FDG uptake tends to be higher in thyroid tumors that are not well differentiated, in contrast to ¹³¹I, which is accumulated by differentiated cancers. For this reason, ¹³¹I-negative tumors are more often positive with ¹⁸FDG, and ¹⁸FDG-negative tumors tend to be positive with ¹³¹I.^{23 18}FDG-PET scans show low background in the chest and liver, which gives this agent a relative advantage over ²⁰¹Tl and ^99mTc-MIBI. Enhanced clinical value of PET images is obtained when they are fused with CT or MRI images. The last section of this chapter, Positron Emission Tomography and Bimodality CT or MRI Fusion Scanning, offers an in-depth discussion of the subject.

Technical Aspects

Sonography uses high-frequency sound waves (ultrasound) in the megahertz range to produce a photographic image of the internal structure of the thyroid gland and its region.24,25 No ionizing radiation is involved, nor is iodinated contrast material given. Sonography is safe; tissue damage has not been reported, and it is less costly than other imaging procedures. Preparation of the patient for the procedure is unnecessary, and it is performed without discontinuing TSH suppressive therapy. To image the thyroid gland and surrounding regions, the patient’s neck is examined in the sagittal, transverse, and oblique planes with a probe called a transducer that both generates the sound energy and receives the reflected signal. The sound enters the body and is transmitted or reflected by interfaces within the tissues. Air does not transmit ultrasound, and calcified areas block its passage. The images are produced quickly and are assembled electronically in “real time.” Each frame of the sonogram shows a static image, and sequential pictures depict motion. Swallowing is used to elevate the thyroid to examine the lower pole of an enlarged lobe, and this maneuver may facilitate identification of the esophagus. With the use of a signal having a frequency of 7.5 to 12 MHz, thyroid nodules and lymphadenopathy as small as 2 to 3 mm are identified in shades of gray. Dynamic information such as blood flow is added by using physics principles called the Doppler effect.²⁶ The signals are translated into colors to differentiate static fluid-filled cystic spaces and blood flowing through the vasculature. Thus the direction and velocity of flow and the degree of vascularity are revealed. Color is assigned to the signal by assuming that venous flow is parallel to, but in the opposite direction to, arterial flow. Arterial signals are made red, and the accompanying venous signals are made blue. The shade of a color is proportional to the direction of flow as it relates to the transducer and flow velocity.

Routine protocols for ultrasound scanning by a technologist are not satisfactory but unfortunately are used too often, degrading the value of the test. Rather, the ultrasound operator must be experienced and aware of the clinical question that has been posed to provide an appropriate answer. Close supervision by a sonographer-physician or endocrinologist who is expert in physical examination and ultrasonography is needed.

It is important to comprehend that the optimal clinical value of sonography depends on the quality of the ultrasound examination, including the maturity of the examiner and the characteristics of the equipment. Grossly misleading results are common with quick, incomplete studies and unsophisticated machines or readouts. Therefore, routine sonography in a medical or radiology office will require proper preparation of the professional and technical staff. Without study, training, and experience, there are likely to be unacceptable results and adverse outcomes.

Sonography of the Normal Thyroid Gland and Environs

With standard gray-scale technique, the normal thyroid gland has a homogeneous appearance like ground glass (Fig. 7-5). The surrounding muscles are of equal or lower echogenicity. Tissue planes are identified. The air-filled trachea, which does not transmit the ultrasound signal, is poorly imaged, and dense echoes represent its calcified tracheal ring anteriorly. The carotid artery and other blood vessels are echo-free unless calcified. Lateral and anterior to the carotid arteries is the jugular vein, which is frequently collapsed and can be identified when it is distended during a Valsalva maneuver. Small blood vessels on the surface of the thyroid and the inferior thyroid artery and vein can sometimes be seen. Color Doppler enhances the identification of blood vessels and flow. The esophagus is sometimes detected behind the thyroid and left of center, anteromedial to the longus colli muscle. It can be observed to distend after the patient swallows a sip of water. Lymph nodes can be seen normally as less than 1 × 3 mm, elliptical, uniform structures with an echo-dense central hilum. The parathyroid glands are not visualized unless they are enlarged. They are less dense to ultrasound than the thyroid gland because of the absence of iodine.

Generally, imaging procedures are not useful in patients whose thyroid gland is normal to palpation unless the patient has a history of exposure to therapeutic irradiation in youth or metastatic thyroid cancer has been discovered and a primary lesion in the gland is being sought. However, in selected circumstances, an ultrasound image may be used to supplement or confirm physical examination findings and identify the size and shape of regional structures accurately and relatively inexpensively. This test is useful when thyroid anatomy is in question but clinical perception is confused by obesity, great muscularity, distortion by abnormal adjacent structures, tortuous blood vessels, a prominent thyroid cartilage, metastatic tumor, lymphadenopathy, prior surgery, or examiner inexperience. On the other hand, the sonogram is so sensitive that many small, nonpalpable thyroid nodules may be detected. Management of these nodules, which are called incidentalomas, requires mature clinical judgment. Objective diagnostic evaluation of all of them is needless and impractical. Selective attention to those found in a patient with a high risk of cancer, those with special characteristics frequently seen in cancers, or those that grow is prudent. Neglecting all of them is inappropriate and occasionally dangerous.

Sonography When There Is Thyroid Enlargement (Goiter)

Enlargement of the thyroid gland is common. In general terms, enlargement may be diffuse and symmetrical, asymmetrical, smooth and uniform, or nodular. It is not generally necessary to obtain a sonogram to confirm thyroid enlargement unless a specific question has attracted clinical concern. Such questions may involve a dominant nodule, a tender spot, focal hardness, or substernal extension. At times, a physician may obtain a sonogram to explain a cryptic finding, such as differentiating goiter from fat or muscle, documenting a controversial observation, estimating the size of the thyroid gland for ¹³¹I dosimetry, or assessing volume changes in response to suppressive therapy with thyroid hormone. Sonography has been used in population studies to objectively identify thyroid enlargement as a screen for iodine deprivation.

Sonography can show alterations of the echo pattern of the thyroid gland and its size. Cystic and/or hemorrhagic degeneration, which is depicted by an echo-free zone, is common (Fig. 7-6). These findings are not specific for any particular type of pathology. However, sonography can identify one region in a uniform goiter whose echo pattern is different from the rest of the goiter, which is suggestive of a focal lesion, especially if that focus is surrounded by a distinctive rim or halo.

Sonography in Patients With Thyroiditis and Graves’ Disease

The greatest value of ultrasonography in patients with immune or inflammatory thyroid disease is to identify incidental focal lesions. However, because most focally distinct zones in these glands are not neoplastic, the need for their subsequent management requires judgment. In situations in which localized firm consistency, focal enlargement, or pain call attention to a part of Graves’ or Hashimoto’s goiter, or if a scintigram shows a cold area, a sonogram may demonstrate a region that has a distinctive appearance, which should be subjected to accurate aspiration biopsy (Fig. 7-7).

Sonography may demonstrate an image of the thyroid gland that correlates with subacute thyroiditis, Hashimoto’s thyroiditis, and Graves’ disease, but such correlation is of debatable practical diagnostic importance. Several types of thyroiditis show reduced echogenicity. During the active phase of subacute thyroiditis, the echogram is characterized by a severely reduced echo density of the thyroid gland that returns to a normal pattern with healing.²⁷ Some patients with Hashimoto’s thyroiditis have low echogenicity (see Fig. 7-7). Marcocci and co-workers reported that only 44 of 238 (18.5%) patients with autoimmune thyroiditis had diffuse hypoechogenicity, especially when they were hypothyroid.²⁸ The accuracy of a heterogeneous, hypoechoic sonographic pattern in diagnosing Hashimoto’s thyroiditis was compared to that of antithyroid peroxidase antibody thyroid peroxidase antibody (TPOAb) concentration in 451 ambulatory patients with unknown thyroid status, excluding those with suspected hyperthyroidism or who were on drugs known to cause hypothyroidism. There was high intraobserver and interobserver agreement on the abnormal thyroid ultrasound patterns, which were judged highly indicative of autoimmune thyroiditis and allowed the detection of thyroid dysfunction with 96% probability.²⁹ Furthermore, in another investigation, patients with postpartum thyroiditis who had both high levels of antithyroid peroxidase antibody and a hypoechoic thyroid gland also had a high risk of long-term thyroid dysfunction.³⁰ Graves’ disease goiters and a few goiters without thyroid autoantibodies have a similar appearance. In Graves’ disease, color Doppler imaging can detect diffuse hyperemia in the thyroid gland,³¹ a condition that has been called a “thyroid inferno.”³² Increased flow velocity in hyperthyroid patients has been demonstrated with duplex Doppler techniques.³³ However, neither the sensitivity nor the specificity of these observations is known.

Sonography of the Thyroid Nodule

Thyroid nodules are identified by sonography because they distort the uniform echo pattern or the shape of the gland. Most nodules have a less dense appearance than normal thyroid tissue does. Most of the remainder are more echogenic, and isoechogenicity is less common. Some nodules have a sonolucent rim called a halo. Nodules may contain regions of calcium (Fig. 7-8) that are extremely echo-dense. The echo texture within a small nodule tends to be uniform, but nodules larger than 2.5 cm usually have irregular zones that are free of echoes. These areas represent cystic and/or hemorrhagic degeneration that may occur in benign or malignant nodules (see Fig. 7-6). Such nodules are called complex. Careful examination of echo-free zones is necessary to discern internal echoes that represent septa or small solid regions that differentiate common, complex cystic nodules and a true thyroid cyst. A cyst is encountered once in approximately 500 to 1000 nodules and is globular shaped, smooth walled, and without internal echoes.

The prevalence of palpable thyroid nodules in members of the general population who are screened by palpation is 1.5% to 6.4%.³⁴ It is 10-fold greater when they have been screened by ultrasonography.³⁵ The prevalence of sonographically detectable nodules increases with age to around 50% in older adults. The risk of malignancy of palpable nodules is 5% to 15%; in ultrasonically detected nonpalpable nodules, the risk is smaller.³⁶ The major use of ultrasonography is to determine which of the many nodules are the relatively few malignancies. Sonography can help in this triage, but the ultrasonic appearance of a nodule cannot reliably differentiate benign lesions and cancer.³⁷ Table 7-3 lists the sonographic features of a nodule that are associated with high or low risk of thyroid cancer. Cancers may be minute or large, entirely solid or complex. Some features of a nodule can be helpful to select nodules that are likely malignant. The most reliable ultrasonic predictor of malignancy is vascular invasion, but this feature is observed uncommonly. Adenopathy associated with a nodule is also very good evidence, providing there is no other reason for enlarged nodes such as infection, as will be discussed later. It is important to realize that in children, adenopathy is a relatively nonspecific finding and must be interpreted with caution. The characteristics of nodules that studies have shown most useful for identifying carcinomas include the intensity of the echoes, the lack of sharpness and irregularity of the boundary of the nodule, the presence of an incomplete “halo” or calcifications, and internal structure, including vascularity.^38,39 However, there is considerable variation in the cancer-predictive value of these characteristics. The most optimistic data is a 97.2% positive predictive value for cytologically diagnosed cancer and 96.1% predictive value for benign disease among 1244 nodules in 900 patients who were stratified according to ultrasound characteristics on a scale of 1 to 5 assessing cancer-risk.⁴⁰ The author is considerably less confident. As a group, malignancies tend to be rather hypoechoic.^44,45 In one study, 62% of cancers were hypoechoic among 202 patients with nodules, and few were hyperechoic.⁴¹ In another series, none of the 14 cancers were hyperechoic among 132 consecutive ultrasound-guided fine-needle aspiration biopsies.⁴² However, most benign nodules are also of low echo density. Nevertheless, it can be said that hyperdense nodules are probably not cancerous and need not be biopsied unless there are other factors that indicate a high risk of cancer.⁴³ Internal cystic spaces in thyroid nodules are degenerative and have no cancer-predictive value. In distinction, cystic space in lymph adenopathy is important because it may be seen more often with thyroid cancer than in inflammatory nodes. Deposits of calcium may be seen in benign or malignant nodules. Frequently, large irregular plaques or eggshell calcifications are found, and because benign nodules are more common than malignant ones, these concretions do not correlate with cancer. In distinction, punctate calcifications or microcalcifications are not common in nodules but have high specificity for thyroid cancer (95.2%), low sensitivity (59.3%), and a diagnostic accuracy of 83.8%.³⁶ They may represent psammoma bodies in papillary cancer. A halo around a nodule is thought to represent a boundary, capsule, or vasculature that may be seen in benign or malignant conditions.^44,45 Demonstrating in multiple images and in several planes that a halo is incomplete correlates with cancer, but the finding is of low specificity and sensitivity. Lack of distinctness of a nodule has little diagnostic value, but an ill-defined edge has been reported with infiltrating, poor-prognosis lesions.⁴⁶ In general, the shape of a nodule has limited diagnostic significance except that cancers tend to exhibit a tall and thin shape. The patterns of blood flow as depicted by Doppler examination offer insight into the potential for malignancy. An internal or central flow pattern in a hypofunctioning (“cold”) nodule or lymph node should raise one’s suspicion of malignancy. In one study of 125 nodules, 55 of 92 (60 percent) cold nodules had a peripheral flow pattern, 34 no internal flow, and only 3 had increased internal flow. All 3 cold nodules with an enhanced internal color flow pattern were carcinomas. However, among the 27 patients who had pathology correlation, there were a total of 7 cancers, only 3 of which had increased internal flow; the other 4 had either diffuse or no internal flow.⁴⁷ Another study of 203 patients, the addition of color flow Doppler imaging to conventional sonography only slightly increased the screening sensitivity and accuracy from 71.9% to 83.3% in identifying the 36 malignant thyroid nodules.⁴⁸

Ultrasonic features of a nodule that specifically favor benign disease include hyperechogenicity and characteristics that suggest the presence of abundant colloid accumulation within the nodule. A “Napoleon” or puff-pastry, layered appearance of the nodule, or a bright spot with a trailing, fading comet tail have been reported with colloid nodules.

Thyroid palpation might not accurately predict the need for sonography in patients with a thyroid nodule. One suspects that routine sonography will be employed, especially when palpation is uncertain or skills are tentative.

Evidence is mounting in support of routine sonography for patients with palpable uninodular thyroid disease with or without a palpable goiter. Thyroid ultrasonography has been reported to show that among 114 patients who were referred because of a solitary thyroid nodule, ultrasonography detected additional nonpalpable thyroid nodules that were at least 1 cm in diameter in 27 patients and no nodules in 23. In this investigation, sonography provided information to the clinician that importantly altered management in 63% (109 of 173) of patients who were referred to a tertiary endocrine group. Sonography showed an indication for needle aspiration or demonstrated that the procedure was not necessary. Among 59 patients who were referred because of goiter, sonography revealed nonpalpable nodules that were at least 1 cm in diameter in 39 patients, requiring aspiration that was not anticipated.⁴⁹

Sonography in Patients With a Dominant Palpable Nodule in a Goiter

Improved technology has permitted the detection of thyroid nodules as small as 2 mm, which can be the source of problems.³⁶ Approximately 20% of all adults have nonpalpable micronodules that are of indeterminate significance, usually benign, and of no clinical consequence in most patients (Figs. 7-9, 7-10, and 7-11). Their discovery, usually by sonography of the neck but sometimes by CT or MRI during an investigation of cervical vascular or neurologic pathology or during thyroid sonography for a palpable thyroid nodule, may occasion needless expense, concern, and therapy. However, rarely one of these lesions represents occult thyroid cancer and could become a clinically significant malignancy.⁵⁰ Therefore, a micronodule, or incidentaloma, should not be dismissed simply for reasons of cost-effectiveness, as some authorities have suggested. Overreaction and surgery are not suitable either. Rather, yearly reassessment seems appropriate. However, it remains for future investigation to determine the value of periodic sonography in examining for changes in the size or characteristics of a nodule, as well as the benefit if any of suppressive therapy with thyroid hormone.

It is common, when a solitary nodule is palpable, for sonography to demonstrate micronodules in the rest of the thyroid (see Fig. 7-9). This occurrence has the same pathologic significance as a dominant nodule in a patient with clinical multinodular goiter. FNA biopsy and cytology of the dominant nodule appear to be the most cost-effective approach.

Sonography of Lymphadenopathy

Ultrasonography may be useful to diagnose and monitor lymphadenopathy in patients, especially if they have thyroid cancer or a history of therapeutic irradiation in youth. However, one must be mindful that even in cancer patients, enlarged benign nodes are more common than malignant ones.

Consensus is growing that the shape of benign lymph nodes tends to be a thin oval, whereas malignant ones are plump and rounded (Fig. 7-12), but differences in size or homogeneity are not reliable indicators of pathology. Solbiati and co-workers evaluated 291 lymph nodes in 143 patients before thyroid cancer surgery and reported that the ultrasonic characteristics of lymph nodes correlated with the histologic findings.⁵¹ A ratio of longitudinal diameter to transverse diameter of less than 1.5 was reported in 62% of metastatic nodes, and a ratio greater than 2 was reported in 79% of reactive nodes.⁵¹ The absence of a nodal hilus was observed in 44% of malignant lesions but in only 8% of benign nodes.⁵² Thus, ultrasound can detect head and neck cancer metastases to cervical nodes with a sensitivity of 92.6%.⁵³ It has been reported that cystic degeneration of a pathologic node occurs with some frequency in papillary thyroid cancer but is uncommon with other head and neck cancers.⁵⁴ It is not clear whether additional information about lymphadenopathy may be offered by color and spectral Doppler studies.⁵⁵ Vascularity along the convexity of a node, remote from the hilus, may represent neovascularization of malignancy.⁵⁶ Distortion of adjacent soft tissue by a node signifies hardness that may be seen with cancer. Even when thyroid cells are not observed in an aspirate from a node, the biochemical detection of thyroglobulin (even in the presence of antithyroid antibodies in the serum) indicates a thyroid metastasis.⁵⁷

Sonography in Patients With a History of Therapeutic Irradiation in Youth

Patients with a history of therapeutic irradiation in youth may have a risk of thyroid cancer that is as high as 30%. Therefore, some clinicians use sonography to screen irradiated people for tiny thyroid nodules before a mass becomes palpable. However, in the process, many more benign nodules are found than malignancies. The inefficiency of the selection process and the indolence of thyroid cancer have resulted in continuing controversy about the clinical relevance of sonographic identification of nodules and subsequent management. The approach to which the author subscribes is to obtain potentially useful baseline ultrasonic anatomic information but not act on it unless the sonogram strongly suggests malignancy (as discussed earlier), studies at intervals reveal changes, or other circumstances arise that heighten suspicion of malignancy.

Sonography to Monitor Changes in Thyroid Size

In a patient with thyroid disease, sonography can accurately and objectively assess the size of the thyroid gland or a nodule during the course of therapy or the emergence of a new nodule.⁵⁸ Because growth of a nodule can be difficult to perceive clinically, sonography may be useful in this context. Furthermore, inasmuch as most patients change doctors over the years, objective assessment of thyroid size greatly facilitates continuity of care. Comparison of serial records, even with different equipment, can demonstrate changes in the thyroid gland or in a nodule and lead to a change in treatment earlier than palpation alone would warrant.

Sonography in Patients With Known Thyroid Cancer

Sonography is useful in the management of a patient with thyroid carcinoma59,60 and has become the most frequently used imaging procedure in patients who have had either a partial or a complete thyroidectomy. After a hemithyroidectomy, the procedure will detect even nonpalpable nodules in the contralateral lobe or lymphadenopathy that could represent tumor (see Fig. 7-11). In these patients and in those who have undergone a total (or near-total) thyroidectomy, sonography that is done without interrupting thyroid hormone therapy will detect recurrent carcinoma either in the thyroid bed or in lymph nodes before the mass has grown sufficiently large to be palpable^59,60 (see Fig. 7-12). Sonography is particularly useful in searching for a nonpalpable malignant lesion in a patient who has had a thyroidectomy when periodic assessment has disclosed an elevated Tg concentration. In contrast, one group of investigators have reported that even when Tg levels remain low or undetectable after stimulation with rhTSH, ultrasonography may identify lymph node metastases from thyroid cancer.⁶¹

Another circumstance in which sonography is useful is when thyroid cancer was diagnosed by detecting metastases of thyroid origin and the thyroid gland is normal by palpation. A sonogram may detect an occult primary thyroid tumor.

Intraoperative ultrasonography may enhance the ability to locate and surgically remove recurrent thyroid cancer that does not accumulate radioactive iodine. Experience in seven patients suggests that sonography was particularly helpful after external beam radiotherapy to identify tumor nodules of 20 mm or less and were invasive or adherent to the airway.⁶²

Sonography in Conjunction With Needle Biopsy

In an attempt to enhance the utility of FNA biopsy, sonographic guidance has been used in selected circumstances to minimize sampling errors⁶³ (Fig. 7-13). Ultrasound guidance for needle biopsy is generally reserved for (1) unusually deep nodules, particularly in an obese, muscular, or large-framed patient; (2) very small nodules; (3) nonpalpable nodules; (4) ultrasonically detected incidentalomas that are associated with cancer risk factors; (5) complex degenerated nodules if a prior aspiration has not been diagnostic; (6) suspicious nodules in a nodular goiter; and (7) nonpalpable adenopathy.^64,65 A special transducer to guide the needle is available but is cumbersome and not required. It is easier to explore the thyroid area of the neck with a hand-held transducer to locate the nodule and then insert the needle into the lesion under ultrasonic direct vision from another angle. The success rate is low for nodules that are smaller than 8 mm. Generally, correlation of the anatomy with the sonographic film without guided puncture is less costly and is adequate for nodules that are palpable unless a prior aspiration has been unsuccessful.

It has been reported from a goiter zone in Italy that as many as 52% of histologically malignant nodules in goiters were found only with the aid of ultrasound-guided FNAB. Therefore, the authors concluded that ultrasound-guided aspiration should be used in areas where multinodular goiter is endemic to assess nodules that are deemed suspicious by virtue of a hypoechoic pattern, a “blurred halo,” microcalcifications, or intranodular color Doppler signal.⁶⁶

Ultrasound-guided aspiration can facilitate biochemical analysis—for instance, calcitonin assay—or can lead to a nonneoplastic diagnosis such as tuberculosis⁶⁷ or amyloidosis.⁶⁸ It seems probable that soon cytologic material will be studied for subcellular components that are useful as tumor markers.

Sonography in Conjunction With Percutaneous Therapeutic Intervention

After an aspiration and cytologic examination have demonstrated that a nodule is benign, ultrasound-guided puncture of a nodule can have a therapeutic role to deliver medication precisely into the lesion and to spare the surrounding tissue.

Percutaneous injection of ethanol has been used to reduce the function of autonomous thyroid nodules.⁶⁹ One investigation has observed 34 patients for up to 3 years, who had percutaneous ethanol injection of autonomous thyroid nodules with a volume larger than 40 mL. The patients required 1 to 11 sessions of 3 to 14 mL of ethanol injection (total amount of ethanol per patient: 20 to 125 mL). The authors report recovery of extranodular uptake on isotope scan and normalization of TSH levels within 3 months from the end of the treatment in 30 of 34 patients and an average reduction in nodule volume of 62.9%. Four of 34 patients were refractory to the treatment, three of whom had had nodule volumes greater than 60 mL. There were no recurrences during 6 to 36 months of observation.⁷⁰ Another study examined 20 patients with autonomous thyroid nodules for 763 ± 452 days after ethanol injection. A mean of 2.85 ± 1.1 injections per patient and a mean volume of 4.63 mL of ethanol were required (nodule volume-dependent). After a mean time of 50 ± 23 days, TSH normalized and was maintained in 16 patients (80%), whose nodular volume reduced 60.8%. Four patients (20%) did not completely respond to the treatment.⁷¹ Less impressive but clinically acceptable results have also been observed in a study that reported a “complete cure” in only 22 of 42 patients (52%), mainly in small nodules, and little or no hormonal response in four patients (9%). However, nodule volume decreased in all cases, and there were no recurrences or serious adverse effects.⁷² In the reported series, “mild to moderate” local pain often occurred after the injections and lasted a day or two, and local hematomas were seen. Major complications such as permanent dysphonia or vascular thrombosis seem to be very uncommon. However, transient paralysis of the laryngeal nerve may occur. Thus, this technique may be an option for some, but not very large, autonomous nodules that cannot or should not be treated surgically or with ¹³¹I.

Percutaneous injection of ethanol has also been used to treat toxic nodular goiter,72,73 and thyroid masses that recur after nontoxic nodular goiters have been treated surgically,⁷³ with results that are similar to those described for autonomous nodules.

Recurrent cysts and cystic spaces in a degenerated solid lesion have been obliterated in this fashion.⁷⁴ Perhaps the procedure will have use in cosmetically unacceptable or very large structures.

Sonographically guided percutaneous ethanol injection may become a treatment option for patients with cervical nodal metastases from papillary thyroid cancer that are not amenable to further surgical or radioiodine therapy. In a study of 21 metastatic nodes in 14 patients, all treated lymph nodes decreased in volume, some impressively. No major complications have been reported.⁷⁵ It seems to the author that this option may be palliative when there are large nodes that threaten to impact on surrounding structures. However, since ethanol-treated nodes may increase in size owing to inflammation, caution is warranted, especially when there are bulky nodes in the thoracic inlet or if they are adjacent to vital structures.

Prospective studies will be required to ascertain whether precise, ultrasound-guided placement of medication will reduce the intensity or duration of pain after the injection and improve success over palpation-directed injection.

Sonography of the Fetal Thyroid

Ultrasonography in pregnancy is an interesting tool to assess thyroid status in utero. Gestational age-dependent and age-independent nomograms for fetal thyroid size have been developed by performing ultrasonograms in 200 fetuses between 16 and 37 weeks of gestation.⁷⁶ Fetal goiters and hypothyroidism have been studied, and successful treatment has been reported. It is thought that intrauterine recognition and treatment of congenital goitrous hypothyroidism can reduce obstetric complications and improve the prognosis for normal growth and mental development of affected fetuses. One report cited a fetal goiter that was diagnosed at 29 weeks of gestation during routine ultrasound examination. Fetal blood sampling performed at this time documented fetal hypothyroidism, and treatment was given using a series of intraamniotic injections of triiodothyronine and subsequently thyroxine. Following birth, neonatal serum TSH levels were within the normal range.⁷⁷ A case of fetal goitrous hypothyroidism associated with high-output cardiac failure was diagnosed at 32 weeks of gestation on the basis of ultrasound examination. The fetus’s thyroid function was examined by amniocentesis and cordocentesis. The fetus was treated by injection of levothyroxine sodium into the amniotic fluid at 33 weeks of gestation. Thereafter, the goiter decreased in size, and the high-output cardiac failure improved.⁷⁸

Sonography of the Newborn Thyroid

Normative data for thyroid length (cm), 1.94 (0.24) 0.9 to 2.5; breadth (cm), 0.88 (0.16) 0.5 to 1.4; depth (cm), 0.96 (0.17) 0.6 to 2.0; and volume (mL), 0.81 (0.24) 0.3 to 1.7 was investigated in 100 (49 male) healthy term Scottish neonates. There was considerable variation (−0.8 to +0.7 mL) between the two lobes in individual babies.⁷⁹

Epidemiologic Use of Ultrasonography

Sonography has been used effectively even in the field in undeveloped areas to evaluate thyroid anatomy and size in iodine-deficient people or to search for cancer in radiation-exposed populations. Interobserver agreement on estimates of thyroid volume has been good in epidemiologic studies, but agreement on echogenicity has been poor.⁸⁰ One group correlated age, body size, and thyroid volume in an endemic goiter area.⁸¹ Another such study concluded that systematic ultrasound screening was useful in Belarus for the early detection of thyroid carcinoma in children 4 to 14 years of age who were exposed to radioactive fallout because of the Chernobyl accident.⁸²

The value of ultrasonographic mass screening to uncover thyroid carcinoma in a population with average cancer risk is controversial because of the presumed low benefit/cost of the screening. One group did thyroid sonograms on 1401 women who were scheduled to undergo a breast examination. Thyroid nodules were detected in 25.2% and thyroid cancer in 2.6% of all subjects. The size of the tumors was significantly smaller in the ultrasound-studied group than in a clinically detected cancer group (P < 0.05).⁸³

Computed Tomography

CT images are a reconstruction of a computer-assisted analysis of multiple x-rays of a region. Standard CT represents discontinuous thin slices through an anatomic region, and spiral CT provides a continuous assemblage of pictures.

In CT of the neck, the thyroid gland is distinctive in that it is relatively more radiopaque than the rest of the soft tissues of the neck because of its high iodine content. The thyroid is homogeneous except for regions of enhanced or reduced density that correspond to nodules, cysts, hemorrhage, and calcification. It is clear that sonography is far more sensitive than CT in detecting millimeter-sized nodules. To precisely define the gland for clinical purposes, the regional vasculature must be enhanced by the intravenous administration of iodinated contrast material, which is a major limitation to subsequent management when CT is used for thyroid diagnosis.86,87

Magnetic Resonance Imaging

MRI images are generated by computer-produced analysis of the interaction of electromagnetic waves of a specific frequency and the hydrogen atoms in a patient’s body. To perform the test, the person must be housed within a magnetic field. Varying the magnetic field can selectively emphasize special properties of the hydrogen atoms. The two properties that are conventionally used in MRI are termed T1 and T2. Because the hydrogen atoms of various tissues have specific T1 and T2 properties, differences between T1-weighted and T2-weighted images can be used to identify the thyroid gland, skeletal muscle, blood vessels, or lymph nodes.⁸⁸ The quality and diagnostic value of MRI are enhanced by the intravenous administration of noniodinated contrast agents such as gadolinium-labeled diethylenetriamine pentaacetic acid or by electronically repressing a relatively unique signal derived from fat (short τ inversion recovery). As with CT, MRI is not as sensitive as sonography in detecting small nodules.

In general, normal thyroid tissue tends to be slightly more intense than muscle on a T1-weighted image, and thyroid tumor usually appears even more intense, or brighter.⁸⁹ Although differences in the MRI characteristics of malignant and benign thyroid tissue have been suggested as a result of investigations in vitro,⁸⁹ the distinctions have rarely proved to be of clinical value.

Distinctions Between Computed Tomography and Magnetic Resonance Imaging

The relative clinical utility of CT and MRI in thyroid disease has not been examined critically and is controversial. However, the author is persuaded that when additional imaging is needed to supplement thyroid sonography, MRI is preferred, and CT should be used only when specifically needed. The major advantages of MRI are that ionizing radiation and iodinated contrast agents are not required. MRI seems to provide better spatial resolution, and reports have suggested superior differentiation of postoperative scar from recurrent tumor.90,91 The use of MRI has been limited by discomfort for a claustrophobic patient, considerable noise, long test time, great demand for the equipment for other types of examinations, relatively high cost, and incompatibility with pacemakers or ferrous prostheses. “Open” MRI systems address some of these problems but currently provide inferior images. CT is more sensitive in detecting small metastases to lymph nodes⁹² and the lungs.^93,94 The total examination time for CT is shorter than that for MRI, and access to CT scanners outside major centers is superior.

Biomarker Imaging of Thyroid Cancer

Surveillance with ¹³¹I WBS is the most specific imaging modality when thyroid cancer is iodine-avid. As discussed previously, this WBS offers a two-dimensional map of tissues that concentrate the radioactive iodine tracer and represents a biomarker of thyroid and other iodine-avid cells but only a general, imprecise indicator of anatomy. Its interpretation is highly objective and depends on an understanding of iodine metabolism, extrathyroid iodine-concentrating sites, thyroid physiology, and information about the physical structure of the body relative to the focus of radioactivity. Anatomic localization may be improved with SPECT, which may be employed after therapeutic doses of ¹³¹I have been administered. This method produces cross-sectional image slices through the patient that are reformatted by computer software to provide 3-D pictures.

PET scanning provides a non-iodine biomarker for rapidly metabolizing thyroid cancer. PET depicts the accumulation of radioactive-tagged glucose by rapidly growing or dividing tissue or organs, aggressive neoplastic tissue, lymphoma, and also regions that are infected or otherwise inflamed and thus is not as specific for thyroid cancer as ¹³¹I WBS. Indeed, benign thyroid adenomas, nodular hyperplasia, focal thyroiditis, Hashimoto’s thyroiditis, Graves’ disease, and thyroglossal duct cyst have been reported to accumulate excessive FDG.⁹⁷ Focal accumulation of FDG thought to be thyroid in origin on PET examination that is done to investigate nonthyroid malignancies has been reported in 1.2% to 4.3% of such studies.^98–102

As a general rule, thyroid cancer that is well differentiated and concentrates ¹³¹I utilizes relatively less glucose than poorly differentiated thyroid cancer that does not accumulate ¹³¹I. Consequently, PET may reveal cancer in WBS-negative patients, especially when thyroglobulin is elevated.

Anatomic Imaging of Thyroid Cancer

There are several anatomic imaging methods for monitoring patients with thyroid cancer. Since most recurrence of thyroid cancer is in the neck, ultrasonography of the neck is the most frequently used method. US may reveal nodules down to the millimeter range, but it lacks specificity for the type of the tissue or its benign or malignant nature. CT and MRI of the neck or whole body may also offer anatomical information in selected circumstances, such as mediastinal, pulmonary, or bone metastases, but also lack specificity for thyroid cancer. In order to derive optimal information from CT, iodinated contrast medium is usually required, which may complicate further diagnosis and treatment of patients with thyroid cancer.

Bimodality Imaging of Thyroid Cancer

The utility of PET is enhanced greatly by precise fusion of its images and simultaneous CT or MRI. Simple visual correlation of PET images with ultrasound, CT, and MRI is time consuming, imperfect, and sometimes misleading. Technologic advances have permitted multimodality scanning so that the same instrument can perform a PET scan and CT or MRI (which is in the late stages of development) with the patient in exactly the same position. Software can then digitally fuse the two images to produce an accurate superimposition of function and anatomy. Thus a composite image is produced that shows a supra-imposition or co-registration of both glucose concentration and a nodule or anatomic structure in the proper physical context. Extraordinary anatomic localization of FDG-avid lesions can be achieved, but mismatch of the two modalities may be a source of confusion. Fusion of WBS and anatomic images has not been achieved, but assembling thyroid gland scintiscan images and US on a single computer screen has been described.¹⁰⁵

A Biochemical Indicator of Thyroid Cancer

Assessment of the concentration of thyroglobulin in the serum when TSH is elevated is a very good marker for the presence of a postoperative thyroid remnant or recurrence of thyroid cancer only after a total thyroidectomy has been performed. Its presence in material aspirated from a metastatic lymph node is diagnostic of thyroid cancer.

Does It Matter If TSH Is Suppressed or Elevated During FDG PET Scanning?

The studies to assess if TSH should be elevated during PET scanning have suffered from low power and debatable design. Firm conclusions cannot be drawn. The author’s impression is that TSH elevation may improve sensitivity somewhat.

Some reports suggest that PET or PET/CT/MRI may be more sensitive to the presence of thyroid cancer when TSH is elevated rather than suppressed. In one study, 10 patients had FDG PET while TSH was suppressed and while it was elevated subsequent to the withdrawal of thyroid hormone. Although 17 lesions were found on both studies, the tumor-to-background ratio of activity increased in 15 of the patients when TSH was elevated compared to when it was suppressed. The change was attributable to increased uptake by the lesion and also to decrease in background activity.¹¹² A similarly designed examination of eight patients revealed more intense uptake of glucose by lesions in four patients and additional lesions in two of these four when TSH was elevated after the withdrawal of suppression.¹¹³ Similarly, after the administration of human recombinant TSH (rhTSH), the number of “tumor-like lesions” increased from 22 when TSH was low to 78 to after rhTSH. Furthermore, the tumor-to-lesion ratio and standardized uptake value (SUV) also increased.¹⁰³ Among seven patients, rhTSH stimulation identified four lesions that were not seen during TSH suppression, and one patient had lesions after rhTSH not seen during suppression.¹⁰⁴

Other studies have suggested that the two preparations are similarly effective. For instance, the relative efficacy of raising TSH by withdrawing L-thyroxine or administering rhTSH was studied in 15 patients who had elevated levels of thyroglobulin but negative WBS. The comparison consisted of seven patients whose TSH was elevated by withdrawing thyroxine and eight patients who had rhTSH. Abnormal FDG uptake was observed in four of the first and five of the second group.¹⁰⁵

The Thyroid Incidentaloma in PET Scanning

Thyroid incidentalomas in FDG PET occur in 1.2% to 4.3% of patients undergoing PET because of nonthyroid malignancy,98–104 as mentioned earlier. They are entirely similar to the incidentalomas that have been reported in thyroid ultrasonography. The majority of PET incidentalomas are benign but result in considerable anxiety, medical testing, and expense. There is considerable overlap in the SUV between benign and malignant incidentalomas.¹⁰³ One investigation disclosed that 70 of 1763 patients (4.0%) had excessive FDG accumulation in the thyroid region; 36.7% of the 70 lesions were malignant. The maximum SUV was significantly higher in the malignant lesions (10.7 verses 6.7). However, the overlap of maximum SUV was so great as to make the distinction meaningless (2.0 to 32.9 for malignant opposed to 2.3 to 33.1 for benign).⁹⁸ Therefore, lesions that are sufficiently large should be examined according to the criteria described for ultrasonography.

PET Scanning in the Initial Evaluation of Patients With Thyroid Nodules

Thyroid nodules are usually evaluated with fine-needle biopsy and cytology unless TSH is suppressed, in which case a ¹²³I scan is the appropriate next step. Ultrasonography is the cost-effective anatomic imaging study. PET scanning plays no role.

PET Scanning in the Initial Staging of Patients With Thyroid Cancer After Thyroidectomy

Ultrasonography (for the neck), TSH stimulated ¹³¹I WBS, and TSH-stimulated, in vitro measurement of thyroglobulin play complementary and essential roles in the initial postoperative evaluation of patients with thyroid cancer. Uncontrolled clinical testing of patients with thyroid cancer and a few actual investigations have explored the feasibility, utility, and cost effectiveness of using PET scanning for initial screening. PET scanning can reveal some metastatic lesions but is not adequately useful clinically to routinely warrant this expensive, time-consuming, and isotope-requiring technique. However, PET scanning may be useful in the initial evaluation and staging of local cancer or metastases in patients with poorly differentiated or anaplastic thyroid cancer. The sensitivity of this application is unknown, but since anaplastic cancers usually do not adequately concentrate ¹³¹I, and there may be extensive distant metastases, PET with CT or MRI are the appropriate imaging techniques in this setting rather than WBS. Similarly, with Hürthle cell thyroid cancer, where ¹³¹I accumulation tends to be low, and medullary thyroid cancer, where there tends to be no uptake, PET scans are especially useful. Several studies have shown the value of PET scanning for Hürthle cell lesions117–120 and for medullary tumors.^121–129

PET Scanning in the Detection of Recurrence of Thyroid Cancer

As in the initial staging after surgery, ultrasonography of the neck, ¹³¹I WBS, and measurement of thyroglobulin are the main tools to assess patients for recurrence of typical thyroid cancer. In selected cases, percutaneous fine-needle biopsy of suspicious nodules together with cytology and sometimes measurement of thyroglobulin in the aspirate complement these examinations. PET scanning seems to be cost effective when ¹³¹I WBS is negative, but thyroglobulin is identified in either the blood or an aspirated specimen of tissue or fluid.

The role of FDG PET scanning (PET/CT or PET/MRI) in patients with thyroid cancer is the detection and localization of recurrence of the cancer when ¹³¹I WBS is negative, but there is reason to believe that thyroid cancer deposits are present because thyroglobulin is elevated and/or ultrasonography shows a suspicious lesion.130–136 This conclusion is based on observations that thyroid cancers that do not take up ¹³¹I frequently concentrate FDG and are detectable on PET scanning. As an example, in one investigation among patients who were strongly suspected of persistent or recurrent cancer by traditional clinical criteria, FDG PET identified 17 of 18 sites that concentrated ¹³¹I on WBS. Eleven additional sites that concentrated FDG but not ¹³¹I were also disclosed. In contrast, there was normal FDG distribution in 19 of 27 patients who did not concentrate ¹³¹I.¹³⁷

Studies that evaluate the utility of PET scanning agree about a general value but vary as to specificity and sensitivity. False positives are uncommon. The result of PET alters clinical management in many of the patients.

PET Scanning in Patients With Hyperthyroidism and Lymphocytic Thyroiditis

PET scanning may be used in a variety of disorders to study the metabolism of glucose. For example, it has been found that there is enhanced uptake of glucose in hyperthyroid patients that correlates with increased levels of antithyroid antibodies, but it is unclear if the activity reflects thyroid cell or lymphocyte function.¹³⁸

Novel Observations from PET Scanning

It is believed that thyroid cancers that are PET positive and ¹³¹I WBS negative behave clinically in a more aggressive fashion than ¹³¹I-positive lesions and are less differentiated histopathologically. A retrospective analysis demonstrated relatively reduced survival of patients older than 45 years, abnormal focal FDG uptake, and large volume of the FDG lesions. The variable that most strongly predicted survival was large FDG volume of lesions.¹³⁹ In contrast, a negative PET scan in a patient with known metastatic disease seems to confer a favorable survival advantage when compared with patients who are PET-positive.¹⁴⁰

The expression of glucose transporters has been linked to abnormal PET FDG accumulation and to unfavorable prognosis in patients with thyroid cancer. This observation may offer insight into the mechanism of enhanced labeled-glucose accumulation by certain thyroid cancers. Studying the expression of types 1 to 5 glucose transporters in formalin-fixed, paraffin-embedded tissue samples from 45 patients with carcinoma, increased levels of glucose transporter type 1 (GLUT1) was found in the cancers of patients who had an unfavorable prognosis. Among these were anaplastic thyroid cancer and aggressively behaving follicular cancers. Low or no GLUT1 expression was observed in both normal thyroid tissue and well-differentiated tumors that had a good prognosis.¹⁴¹