This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Hegedüs, L. Articles by Bennedbæk, F. N. Search for Related Content PubMed PubMed Citation Articles by Hegedüs, L. Articles by Bennedbæk, F. N.

Management of Simple Nodular Goiter: Current Status and Future Perspectives

Department of Endocrinology and Metabolism, Odense University Hospital, DK-5000 Odense C, Denmark

Correspondence: Address all correspondence and requests for reprints to: Laszlo Hegedüs, M.D., D.M.Sc., Department of Endocrinology and Metabolism, Odense University Hospital, DK-5000 Odense C, Denmark. E-mail: laszlo.hegedus{at}ouh.fyns-amt.dk

	Abstract

Top Abstract I. Introduction II. Etiology III. Epidemiology IV. Natural History V. Diagnosis VI. Treatment VII. Unresolved Issues and... References

 
The simple nodular goiter, the etiology of which is multifactorial, encompasses the spectrum from the incidental asymptomatic small solitary nodule to the large intrathoracic goiter, causing pressure symptoms as well as cosmetic complaints. Its management is still the cause of considerable controversy.

The mainstay in the diagnostic evaluation is related to functional and morphological characterization with serum TSH and (some kind of) imaging. Because malignancy is just as common in patients with a multinodular goiter as patients with a solitary nodule, we support the increasing use of fine-needle aspiration biopsy (cytology).

Most patients need no treatment after malignancy is ruled out. In case of cosmetic or pressure symptoms, the choice in multinodular goiter stands between surgery, which is still the first choice, and radioiodine if uptake is adequate. In addition to surgery, the solitary nodule, whether hot or cold, can be treated with percutaneous ethanol injection therapy. If hot, radioiodine is the therapy of choice. Randomized studies are scarce, and the side effects of nonsurgical therapy are coming into focus. Therefore, the use of the optimum option in the individual patient cannot at present be based on evidence. However, we are of the view that levothyroxine, although widely used, should no longer be recommended routinely for this condition.

Within a few years, the introduction of recombinant human TSH and laser therapy may profoundly alter the nonsurgical treatment of simple nodular goiter.

I. Introduction

II. Etiology

III. Epidemiology

IV. Natural History

V. Diagnosis

A. Clinical evaluation

B. Laboratory investigations

C. Diagnostic imaging

D. Fine-needle aspiration biopsy

VI. Treatment

A. Multinodular nontoxic goiter

B. The solitary nontoxic nodule

VII. Unresolved Issues and Future Perspectives

	I. Introduction

Top Abstract I. Introduction II. Etiology III. Epidemiology IV. Natural History V. Diagnosis VI. Treatment VII. Unresolved Issues and... References

 
NODULAR GOITERS ARE clinically recognizable enlargements of the thyroid gland characterized by excessive growth and structural and/or functional transformation of one or several areas within the normal thyroid tissue. In the absence of thyroid dysfunction, autoimmune thyroid disease, thyroiditis, and thyroid malignancy, they constitute an entity described as simple nodular goiter (SNG; Ref. 1).

Simple goiter occurs both endemically, mainly related to iodine deficiency when goiter prevalence in children 6–12 yr of age within a population is more than 5%, and sporadically, when this number is 5% or less. The clinical evaluation of size, morphology, and function is highly inaccurate (2) as illustrated, for example, by up to 50% of subjects with a solitary palpable nodule or a diffusely enlarged gland actually having multiple nodules when investigated by sonography (3, 4, 5). Additionally, up to 50% of the general population has thyroid nodules by sonography, even when the gland is normal to palpation (3, 4, 5). This may be one of the explanations for the increasing tendency, at least for thyroidologists, to use diagnostic imaging in the evaluation of such patients, in addition to functional characterization with serum TSH (6, 7, 8, 9).

SNG covers a spectrum from the clinically normal gland to diffuse, uninodular, or multinodular and cystic enlargement of the thyroid, using the most sensitive of imaging techniques (4, 10). The clinical manifestations are related to those of growth and functional autonomy, unpredictable in the individual patient, leading to cosmetic and pressure symptoms on the one hand and that of hypersecretion of thyroid hormone on the other (1). The latter is increasingly recognized as a health hazard (11, 12). Another question is the risk of thyroid cancer in SNG. Although concern of an increased prevalence in SNG seems unfounded (3, 5), accumulating evidence suggests that the risk of cancer in SNG, approximately 3–5%, is no different whether the gland contains a single nodule or multiple nodules (3, 5).

Pathogenesis of SNG, excellently reviewed in recent publications (1, 13, 14, 15, 16, 17), is considered beyond the scope of this article. We will, however, in brief, comment on etiological aspects of this condition (1, 13, 14, 15, 16, 17, 18), because consideration of possible preventive measures is a prerequisite for costeffective management. Consequently, the discussion will emphasize management aspects, i.e., diagnostic and therapeutic aspects. For this, we rely heavily on our recent questionnaire studies among clinician members of the European Thyroid Association (ETA) and the American Thyroid Association (ATA). A number of issues are unresolved, and it should be evident from the following that major controversies exist (6, 7, 8, 9, 19, 20). Despite this, we aim at providing a balanced recommendation and point at future therapeutic perspectives whenever this is at all possible. Although a distinction between the solitary nontoxic nodule and the multinodular nontoxic goiter is not always straightforward, we have in the following, whenever appropriate, tried to separate the two.

	II. Etiology

Top Abstract I. Introduction II. Etiology III. Epidemiology IV. Natural History V. Diagnosis VI. Treatment VII. Unresolved Issues and... References

 
Familial clustering of simple goiter has long been recognized, but classical genetic analysis based on Mendelian principles has shown no simple mode of transmission. In recent years, it has become clear that both endemic and sporadic simple goiter belong to the group of diseases referred to as complex diseases (18, 21) along with, for example, autoimmune thyroid disease (Graves’ disease and Hashimoto’s thyroiditis; Refs. 22 and 23). These conditions are common, vary in their severity, and are multifactorial, with the clinical phenotype representing the net effect of all the contributing genetic and environmental factors. In these conditions, it has been difficult to separate environmental influences from genetic susceptibility (Fig. 1).

View larger version (19K): [in this window] [in a new window]   FIG. 1. The development of simple nodular goiter seems to involve complex interactions between environmental, genetic, and endogenous factors. At present, it is not clear how and to what degree the susceptibility genes interact with the environmental risk factors.

Although familial aggregation of simple goiter has repeatedly been demonstrated (35, 36, 37), family studies cannot determine whether this results from shared genes or shared environment. However, because concordance rates for simple goiter in female monozygotic twins have been reported higher in endemic (80%; Refs. 38 and 39) as well as nonendemic areas (42%; Ref. 21) as compared with female dizygotic twins [40–50% (Refs. 38 and 39) and 13% (Ref. 21), respectively], evidence of a genetic component in the etiology of simple goiter is provided. In an endemic goiter area, the heritability of the liability to goiter development in females has been calculated to 39% [95% confidence interval (CI), 0–79%; Ref. 39 ]. On the basis of path analysis (structural equation modeling), Brix et al. (21) estimated this heritability to be 82% (95% CI, 67–92%) in a nonendemic goiter area. The remaining 18% (95% CI, 8–33%) of the phenotypic variance resulted from individual-specific environmental factors not shared by the twins (21). Although the twin data cannot uncritically be interpreted as the relative importance of heredity and environment, in any population, both endemic and sporadic simple goiter seems to develop on the basis of genetic susceptibility interacting with environmental factors, the two most important being level of iodine intake and cigarette smoking. Consequently, possible susceptibility genes or genetic markers will be recognized more easily in nonendemic than endemic goiter areas.

Studies assessing the role of specific candidate genes in the etiology of simple goiter have given conflicting results. Corral et al. (40) demonstrated a thyroglobulin (Tg) gene point mutation in chromosome 8 associated with nonendemic goiter, which could, however, not be confirmed by others (41). Using classic linkage analysis in combination with a genomewide screening in a large French-Canadian pedigree, Bignell et al. (42) identified a region of interest on chromosome 14, multinodular goiter 1 (MNG-1). Two other studies (41, 43) have examined the role of MNG-1 and other possible candidate genes/markers, including Tg, TSH receptor (TSHR), and sodium iodide symporter (NIS), in the etiology of simple goiter, with conflicting results. Recently, Capon et al. (44) mapped a dominant form of multinodular goiter to chromosome Xp22. This finding awaits confirmation. It follows that genetic heterogeneity, i.e., no single gene being either necessary or sufficient for disease development, is highly likely. The results of these studies, all in single or small groups of families, cannot at present be extrapolated to the general population. It is likely that single genes play a role in certain families. However, studies in larger numbers of pedigrees, taking gene-environment interaction into consideration, are necessary.

In light of the observed linkage between MNG-1 and goiter (41, 42), it is intriguing that Tomer et al. (45) have found linkage of Graves’ disease to another locus on chromosome 14, Graves’ disease 1. If confirmed, this could indicate the presence of a gene complex on chromosome 14, etiologically related to thyroid disease in general. However, as pointed out in Section I, simple goiter certainly comprises a number of phenotypes, complicating any comparison between available studies. This, in addition to the small number of families studied, hinders any firm conclusion on possible candidate genes in SNG.

	III. Epidemiology

Top Abstract I. Introduction II. Etiology III. Epidemiology IV. Natural History V. Diagnosis VI. Treatment VII. Unresolved Issues and... References

 
Epidemiological studies of nodular goiter are hampered by problems such as selection criteria (e.g., age, sex), influence of environmental factors (e.g., iodine and drug intake, smoking and drinking habits), evaluation of size and morphology (palpation, sonography, scintigraphy), and determination of thyroid function and whether subjects with subclinical thyroid dysfunction are included (46). Most studies have focused on middle-aged women and the elderly, whereas only a few have documented the prevalence of nodular thyroid disease in a population-based manner. Large-scale population-based longitudinal studies using diagnostic imaging allowing distinction between uninodular and multinodular disease and morphological as well as functional characterization do not exist. Therefore, adequate data on prevalence, incidence, etiological risk factors, and the natural history are lacking. These limitations should be born in mind when considering the available data. Despite these shortcomings, a pattern of increased thyroid volume and nodularity in areas with iodine deficiency is the rule (46, 47, 48, 49, 50). Furthermore, it has been suggested that relatively small differences in the iodine intake of a population may lead to notable differences in median thyroid volume, goiter prevalence, and nodule size (51). There are insufficient data allowing estimation of the optimum level of iodine intake to minimize goiter occurrence. Additionally, it is emphasized that goiter and thyroid nodules also exist in the face of iodine sufficiency, and even in the case of iodine excess (52, 53).

In the Whickham survey of a representative sample of the adult population from a geographic area of the United Kingdom, 15.5% of the participants had a palpable goiter (8.6% had a small goiter), with a female to male ratio of 4.5:1 (54). There was a weak association between goiter and thyroid autoantibodies and no relation to urinary iodine excretion. In Framingham, Massachusetts, where iodine intake was also sufficient, 1% of persons between 30 and 59 yr of age had multinodular goiter by palpation (55). In Connecticut, 2% of adults were reported to have nodular glands (56, 57). In Denmark, palpable goiter was demonstrated in 9.8% of a mildly iodine-deficient population and 14.6% of a moderately iodine-deficient population (49). This frequency increased to 15.0% and 22.6%, respectively, when goiter was defined by sonographic determination of thyroid volume (49). There are a vast number of such studies underscoring the inaccuracy of and also the large observer variation in the determination of goiter and thyroid size by clinical examination (2, 58, 59).

In the Whickham survey (54), solitary thyroid nodules were estimated to be present in 5.3% of women and 0.8% of men (ratio, 6.6:1). No details were given about nodule size, function, or their association with goiter. In Framingham, this frequency was 4.6% in all (6.4% in women and 1.6% in men; Ref. 55). However, these numbers are markedly changed if sonography is used. Then, the prevalence of thyroid nodules, even if defined as more than 10 mm in diameter, is usually 20–30% in unselected populations (46, 49, 60, 61) and even higher in older age groups and in areas with insufficient iodine intake (46, 62). These findings are substantiated by autopsy studies demonstrating that half the population had either single or multiple thyroid nodules (46, 63).

Little is known of the functionality of these solitary nodules. Although not systematically investigated, available data suggest that approximately 5% are toxic, 10% are warm, and 85% are cold (49, 64).

	IV. Natural History

Top Abstract I. Introduction II. Etiology III. Epidemiology IV. Natural History V. Diagnosis VI. Treatment VII. Unresolved Issues and... References

 
It is often stated that the natural history of nodular goiter is that of gradually increasing size with development of multiple nodules, local compression symptoms, and/or cosmetic complaints. However, the natural history with respect to growth and function varies and is difficult to predict in a given patient because no specific growth parameters exist. Therefore, it is difficult to decide whether an individual patient can be monitored without treatment or should have treatment before the goiter grows any further and possibly affects treatment outcome adversely (65). This dilemma is clearly illustrated by recent European and North American questionnaire surveys dealing with the solitary (6, 7) as well as the multinodular goiter (8, 9), disclosing profound differences within and between countries in the proneness to offer treatment, once malignancy has been ruled out.

In the Whickham survey (66), 20% of the women and 5% of the men who had goiters in the initial survey had no goiter in the follow-up survey, whereas only 4% of the women and none of the men acquired a goiter between the two surveys. In a 20-yr follow-up study of 11- to 18-yr-old subjects in the southwestern United States, 60% of the 92 subjects who had diffuse goiters initially had spontaneous regression by the age of 30 yr (67). In a large epidemiological study, Knudsen et al. (51) only found an increase of thyroid volume up to the age of 40 yr. On the basis of cross-sectional data and ultrasonic scanning, an average annual growth rate of 4.5% in multinodular goiter has been reported (68). In patients referred because of SNG and who qualified for treatment, it has been estimated to be up to 20% yearly in a noniodine-deficient region (69) but is usually much lower. The suggested decreasing frequency of goiter with age could possibly be explained by a fall in lean body mass with age, known to be related to thyroid volume (70). A fall with age in IGF-I and GH, which is associated in the elderly (71), in patients with anorexia nervosa (72), and in pygmies (73) with a decrease in goiter frequency and thyroid volume, is another possibility. At least in women, the loss of estradiol-mediated increase in proliferation and down-regulation of the NIS gene may be of importance (74).

The natural history of thyroid nodules is poorly understood. In the Framingham survey, new nodules appeared with an incidence of 1 per 1000 per year (57), leading to an estimated lifetime risk for developing a nodule to between 5 and 10%. In a study of 140 selected Japanese patients who had not been treated and were reexamined 15 yr later (75), 13% had an increase in nodule size, 34% were unchanged, 23% had decreased, and in 30% the nodule was no longer palpable. The nodules that increased were originally predominantly solid, whereas those that disappeared were predominantly cystic. With the same reservations, that is the patients with rapid growth, for example, symptoms and clinical suspicion of malignancy were offered treatment, others have found that nodules on average do not change significantly over time (49).

Patients with nontoxic nodular goiter can become hyperthyroid or, less commonly, hypothyroid. However, thyroid dysfunction usually develops only after the nontoxic goiter has existed for many years. Hyperthyroidism often develops insidiously, in contrast to that of Graves’ disease. It often begins with a prolonged period of subclinical hyperthyroidism characterized by low serum TSH and normal serum free T₄ and free T₃ concentrations (76). Increasing nodularity and size are related to a decrease in serum TSH (68). The true rate of progression from normal thyroid function to subclinical and finally overt hyperthyroidism is poorly described. However, two studies suggest an incidence of 9–10% of overt hyperthyroidism during a 7- to 12-yr follow-up period (77, 78). In part, this progression depends on genetic predisposition (21), somatic mutations in individual nodules (17), and extrinsic factors such as iodine intake (52). The latter may be alimental or related to iodine-containing drugs such as disinfectants and amiodarone or from radiographic contrast agents, which, in a goiter with increasing autonomous iodine metabolism, leads to the production of excessive amounts of thyroid hormone. In autonomously functioning solitary nodules (hot nodules), the evolution into a toxic nodule has been estimated at annual rates of 0–6% (79, 80). In one study, 67 of 375 patients developed hyperthyroidism over a mean follow-up of 53 months, corresponding to 4.1% per year (80). Size of the nodule seems crucial because Hamburger (81), during a 6-yr follow-up, demonstrated that nodules greater than 3 cm in size at diagnosis carried a 20% risk of developing hyperthyroidism, whereas nodules less than 2.5 cm in size had only a 2–5% risk of developing hyperthyroidism within the same time frame.

	V. Diagnosis

Top Abstract I. Introduction II. Etiology III. Epidemiology IV. Natural History V. Diagnosis VI. Treatment VII. Unresolved Issues and... References

 
A. Clinical evaluation
1. Manifestations.
There is no clear-cut relation between thyroid size, morphology, and function of the thyroid gland on one hand and the complaints of the individual patient on the other hand. The majority of patients with SNG have few or no clinical symptoms (Table 1). Therefore, given euthyroidism and exclusion of malignancy, many need no treatment. A simple management algorithm for the majority of patients with a solitary or a dominant nodule (index nodule, i.e., the palpable nodule in a thyroid gland with other nodules found incidentally by imaging or a growing nodule in a preexisting multinodular gland) is given in Fig. 2.

View larger version (17K): [in this window] [in a new window]   FIG. 2. Algorithm outlining a cost-effective evaluation and treatment of the solitary or dominant thyroid nodule. In case of strong suspicion of malignancy, surgery is advised irrespective of FNAB results. In case of a nondiagnostic result, repeat FNAB yields a satisfactory aspirate in 50%. FNAB guided by US (US-guidance) allows sampling from the periphery of a solid nodule or solid part of a mixed solid-cystic nodule, increasing the sufficiency rate. The options in case of a diagnostic FNAB cover both solid and cystic nodules. In case of recurrent cysts, the possibilities are reaspiration, surgery, or ethanol injection. The shaded boxes indicate treatment options. *, PEIT or ILP.

The symptoms of tracheal compression are dyspnea, stridor, cough, and choking sensation. With an intrathoracic localization, respiratory distress is present in 30–85% of patients referred to surgery (84, 85, 86, 87). However, such goiters are often very large, and it is unsettled whether the tracheal compression is related to the goiter size or the substernal localization per se. The detrimental influence by the goiter on the respiration is amplified in the recumbent position (88). Concomitant with tracheal narrowing, dyspnea and stridor develop initially only on exertion, but later also at rest. An acute exacerbation in upper airway obstruction may have a benign origin like hemorrhage into a nodule or cyst or upper respiratory infections causing endotracheal swelling. In patients with severe or acute airway distress, often requiring acute intubation, 20–50% may be caused by thyroid malignancy (89, 90). Even in asymptomatic individuals, attention should be paid to the trachea. Thus, Gittoes et al. (91) found that in patients referred with a moderately enlarged goiter, flow-volume loops revealed upper airway obstruction among a third, and correlated poorly with symptoms. Complaints due to esophageal compression are less common and most prevalent in patients with large and partly intrathoracic goiters (86, 87). Even rarer is vocal cord paralysis, which can be transient or permanent, caused by stretching or compression of one or both recurrent laryngeal nerves (86, 87, 92). Phrenic nerve paralysis (93) and Horner’s syndrome (94), due to compression of the cervical sympathetic chain, have been described but are extremely rare. If present, malignancy should always be suspected.

Toxic nodular goiter is beyond the scope of SNG; therefore, symptoms and signs of thyrotoxicosis have not been included. However, a number of patients with SNG certainly have subclinical hyperthyroidism biochemically (17% in one study; Ref. 95), and some of these patients have symptoms compatible with thyroid hyperfunction. Treatment aspects of these conditions will be dealt with subsequently.

2. Clinical examination.
The evaluation of a patient with SNG comprises initially a careful history and physical examination focusing on inspection of the neck (including regional lymph nodes) and the upper thorax and palpation of the goiter to determine its size and nodularity. This clinical evaluation, preferably done with the patient swallowing gulps of water and the head tilted slightly backward, essential for deciding on subsequent investigations and treatment, seems to be a lost or unlearned art for many physicians, as demonstrated by the considerable inter- and intraobserver variation regarding size and morphology of the thyroid (2, 59, 96). Recognition of this may well be one of the reasons for the increasing use of diagnostic imaging by European as well as North American thyroidologists in the evaluation of SNG (6, 7, 8, 9).

With practice, the thyroid gland can also be palpated when of normal size, but to most the thyroid gland does not become palpable until the volume has doubled. A visibly diffusely enlarged goiter has often reached a volume of 30–40 ml. In general, the size of smaller goiters is overestimated, and the size of larger goiters is underestimated (2). Detection of nodules depends on their size, morphology, location within the thyroid parenchyma, anatomy of the patient’s neck, and most of all the training of the physician. The patient is usually unaware of the presence of a nodule smaller than 1.5–2 cm in diameter. Awareness may, however, depend on localization, speed of growth, and the possible pain or discomfort related to, for example, hemorrhage into a nodule. Most goiters remain stable or increase in size slowly. Disregarding the possibility of malignancy, rapid increases may also be due to menstrual cycle-related alterations (97) or release of norepinephrine from a pheochromocytoma (98).

3. Assessment of risk of malignancy.
SNG does not include thyroid cancer, but one of the main aims of the clinical evaluation is to exclude or at least minimize the risk of overlooking thyroid cancer. Table 2 lists the most important factors suggesting thyroid malignancy.

The risk of harboring thyroid cancer is highest in the young and the old, and therefore the diagnostic approach should be more aggressive in these age groups. Nodular thyroid disease is 5–10 times more common in females, whereas the rates of thyroid carcinoma are nearly equal in men and women. Therefore, nodular goiter in a man is more likely to be a carcinoma. Head and neck irradiation in infancy or childhood, for a number of benign conditions, is strongly associated with a subsequent occurrence of carcinoma (100). The possibility that many naturally occurring thyroid carcinomas may be due to fallout radiation from various radiation sources or natural background is strengthened by the observed epidemic of childhood papillary thyroid cancer seen in Belarus and Ukraine after the Chernobyl nuclear accident (101).

Rate of growth is of the essence. Thyroid carcinomas usually grow slowly over a period of weeks or, more often, months. Growth during thyroid hormone therapy is particularly worrisome. Sudden growth is most likely a thyroid cyst or a previously undetected nodule into which a hemorrhage has occurred. Rapid enlargement is also encountered in cases of anaplastic carcinoma or the development of lymphoma in a patient with previous chronic autoimmune thyroiditis.

The presence of discomfort in the neck, jaw, or ear and dysphagia, hoarseness, or dyspnea can occur in patients with benign thyroid nodules, particularly in those with large multinodular goiters, but may also indicate thyroid carcinoma. A hard and fixed nodule is suggestive of thyroid carcinoma, as is paralysis of the vocal cords and ipsilateral lymphadenopathy. Although virtually all patients with thyroid carcinoma are euthyroid (as are most patients with benign thyroid nodules), a suppressed serum TSH suggests subclinical hyperthyroidism. Usually, this rules out clinically significant malignancy, but it should be noted that coexisting hyperthyroidism does not exclude malignancy in the multinodular goiter. Thus, in one large series of operated hyperthyroid patients, thyroid cancer was found in 1.6% of patients with a toxic multinodular goiter, contrasting with none among those with Graves’ disease (102).

It should be remembered that glands that harbor malignancy are in many cases indistinguishable from those that do not. In case of a high clinical suspicion of malignancy (Table 2), thyroidectomy should be advocated irrespective of a benign cytology because the likelihood of malignancy is very high (Ref. 103 ; Fig. 2). When two factors suggesting high suspicion are present, the likelihood of thyroid malignancy approaches 100% (103).

B. Laboratory investigations
Serum TSH is by far the most used test in the initial evaluation of SNG (6, 7, 8, 9). This approach is indeed justifiable, because present TSH assays are very sensitive in detecting thyroid dysfunction. Although preferred by more than half of clinicians (Table 3), it can be strongly questioned whether an initial measurement of thyroid hormones adds further information in this setting, provided the serum TSH is within the normal range. When it comes to the use of other biochemical tests than TSH, North American clinicians are clearly more restrictive than are Europeans (Table 3; Refs. 6, 7, 8, 9).

Although serum Tg correlates with the iodine status and/or the size of the thyroid gland, this marker is too inaccurate at the individual level to have any independent value in the diagnosis of goiter (107, 108). This is reflected by less than 10% of clinicians measuring serum Tg (6, 7, 8, 9).

A test giving rise to much discussion during recent years is serum calcitonin. This hormone is a marker of medullary thyroid carcinoma (MTC), and the serum levels correlate with the tumor burden (109). MTC accounts for less than 10% of all thyroid cancers and is supposed to be present in less than 0.5% of all thyroid nodules (16). Of particular interest is its use in the early diagnosis of MTC. A two-site immunometric calcitonin assay should be used (110), but presence of heterophilic antibodies may still cause falsely elevated values (111). With modern assays, serum calcitonin is below 10 pg/ml in 99% of healthy subjects (112) and is slightly higher in men than in women (113). Raised levels of calcitonin not related to thyroid diseases are seen in conditions such as impaired renal function (114), pseudohypoparathyroidism (115), or treatment with proton pump inhibitors (116). Pentagastrin-stimulated calcitonin was much used in the past to test family members for MTC associated to the multiple endocrine neoplasia 2 syndrome, but this can now easily be done by genetic test for mutations in the RET proto-oncogene (117). Pentagastrin stimulation may still have a role in sporadic cases in which basal calcitonin exceeds 10 pg/ml (118). In the United States, where pentagastrin is no longer available, the calcium-calcitonin test may be used instead (119). The normal upper limit of serum calcitonin after stimulation with pentagastrin is approximately 40 pg/ml (113). If the level exceeds 100 pg/ml, this strongly indicates the presence of MTC or C cell hyperplasia (120). In case of familial disposition, C cell hyperplasia is considered as carcinoma in situ. In other cases, the significance of C cell hyperplasia is unclear, because it is a common finding in the elderly, present in 20% at autopsy (121). The results of six large-scale studies (112, 118, 122, 123, 124, 125) of the routine use of serum calcitonin in patients with nodular thyroid disease are summarized in Table 4. The studies vary with regard to the diagnostic set-up and the fraction of patients with histological verification. The prevalence of MTC was up to 1.4% in a large French study (123) in which 41% of 34 patients with an elevated basal calcitonin had MTC, but also included two false-negative cases. In the study of Hahm et al. (112), 56 of 1448 patients (3.9%) with nodular thyroid diseases had a serum calcitonin level above 10 pg/ml. Ten patients (0.7%) proved to have MTC. Half of these subjects had a basal serum calcitonin level above 100 pg/ml. Vierhapper et al. (118) found 3 of 1062 patients having a basal calcitonin level exceeding 100 pg/ml. More important, in 10 of 31 patients with basal calcitonin ranging from 10–100 pg/ml, a pentagastrin-stimulation increased calcitonin above 100 pg/ml (118). MTC was found in three patients, and another six had C cell hyperplasia. Basal or stimulated calcitonin levels were generally more sensitive than FNAB in detecting MTC, and the routine use of serum calcitonin was recommended by all authors of these studies (112, 118, 122, 123, 124, 125). However, a clear conclusion is not easy to draw from the existing data. Cost-benefit must be taken into consideration, and a high false-positive rate of 60–80% (112, 123) will result in many unnecessary thyroidectomies. It is also evident from our surveys (6, 7, 8, 9) that there is no consensus on this issue (Table 3). Thus, more than 30% of Europeans measure basal serum calcitonin routinely (6, 8), perhaps reflecting that five of the six large studies mentioned (118, 122, 123, 124, 125) were performed in Europe. In contrast, only very few clinicians in North America routinely use basal calcitonin measurement, except in the case of a family history of thyroid cancer (7, 9). This approach is in accordance with existing ATA guidelines (126) as well as guidelines published by the American Association of Clinical Endocrinologists (127).

C. Diagnostic imaging
Although simple and cheap, neck palpation is notoriously imprecise with regard to both thyroid gland morphology (2, 58, 128) and size determination (2, 96, 129). For this purpose, several imaging methods are available: sonography, scintigraphy, computed tomography scan, magnetic resonance (MR) imaging, and perhaps positron emission tomography (PET). Table 5 lists characteristics, advantages, and disadvantages of these imaging modalities. Of these, sonography clearly has first priority among clinicians (6, 7, 8, 9).

2. Scintigraphy.
Thyroid isotope imaging has been used for many years. Although the scintigraphic resolution can be enhanced to 6–7 mm by tomography (SPECT; Ref. 143), this is still far below that of sonography. Therefore, scintigraphy at present has little place in the anatomotopographic evaluation of the nodular goiter (Table 5). However, scintigraphy is very helpful in the determination of the functionality of the thyroid nodules. TSH suppression by levothyroxine (L-T₄) administration emphasizes autonomous nodular function (144), but is not routinely used (6, 7, 8, 9). Nodules with a high uptake by scintigraphy almost never harbor clinically significant malignancy, although exceptions have been reported (145). ^99mTc used as a tracer may result in falsely positive uptake in 3–8% of thyroid nodules (146), whereas iodine isotopes are devoid of this problem. This may explain why the use of ¹²³I is favored by 63% (solitary nodule) and 49% (multinodular goiter) of ATA members (7, 9). The fact that ^99mTc is inexpensive and easier to use, in addition to several studies being unable to demonstrate any significant dissimilarities between the two tracers (147, 148, 149), may explain why this isotope is recommended by 85% of ETA members (6, 8).

The fraction of inactive or "cold" lesions constitutes 77–94% of nodules in consecutive series of thyroid scintigrams (150, 151, 152). The a priori risk of malignancy among cold nodules is reported to be 8–25% or even higher (151, 153, 154, 155). However, such risk estimates are without a doubt heavily influenced by selection bias, because cold scintigraphic nodules greater than 10 mm are found in 2.4% of the subjects living in a borderline iodine deficient area (49). Tracers like ²⁰¹Thallium and ^99mTc-MIBI have an increased uptake in differentiated malignant thyroid nodules, but the sensitivity and specificity do not support their general use (156, 157, 158, 159). Thyroid scintigrams have been used through the decades for measurement of the thyroid volume. However, in comparison with modern imaging techniques, it is very inaccurate (160, 161, 162) and cannot be recommended for this purpose.

Many disregard thyroid scintigraphy in the initial evaluation of patients with nontoxic nodular goiter (16, 126). Nevertheless, more than two thirds of ETA members (6, 8) routinely use scintigraphy, whereas less than 25% of ATA members prefer such a strategy in this condition (7, 9). Indisputable indications for scintigraphy in the setting of a nodular goiter are hyperthyroidism (to visualize hot nodules suitable for ¹³¹I therapy) and a follicular neoplasm shown by FNAB, because warm nodules with great certainty are benign (163).

3. Computed tomography and MR imaging.
Computed tomography and MR provide high-resolution three-dimensional visualization of the thyroid gland. Some aspects of MR and computed tomography for thyroid imaging have recently been reviewed (164). None of these methods have any advantages over sonography when it comes to detailed visualization of the intrathyroidal structure (Table 5). The major strength of computed tomography and MR is their ability to diagnose and assess the extent of substernal goiters much more precisely than any other method (Fig. 3). However, in this situation, only 15–20% of clinicians use these techniques (8, 9). The only comparative study done so far showed that MR is more precise than computed tomography in the anatomotopographic evaluation of the substernal goiter (165). However, whether computed tomography or MR is preferred probably depends on cost and availability. Another advantage of computed tomography and MR is the possibility for planimetric volume estimations, especially useful in irregularly enlarged goiters. For this purpose, the observer variation of MR imaging is 2–4% (141, 161). Interestingly, MR measurements based on the ellipsoid principle underestimate the thyroid volume by approximately 12% in comparison with MR planimetry (138). With regard to computed tomography planimetric estimates, the median intra- and interobserver variations are 5% and 11%, respectively (140), and compared with surgical specimens, computed tomography overestimates the volume by up to 57% (mean, 12%; Ref. 166). The volume of surgical and postmortem thyroid specimens may, however, be dissimilar to in vivo determinations due to the influence of thyroid perfusion. The agreement between computed tomography and MR estimates is unknown, and little is known about the accuracies of the methods, because no gold standard is attainable.

View larger version (142K): [in this window] [in a new window]   FIG. 3. Coronal MR image of a large goiter, demonstrating narrowing of the trachea (white arrow) and left-sided substernal extension (black arrows).

4. Tracheal imaging and relation to goiter size and lung function.
The air flow rate, particularly in the inspiratory phase, will obviously be critically compromised if the lumen of the trachea is reduced beyond a certain point. Imaging of the trachea can be done by plain x-ray, computed tomography, or MR imaging. Plain x-ray has a poor sensitivity with regard to detection of airway obstruction and shows poor correlation with air flow rate (91, 170, 171, 172, 173). However, the tracheal area, which can be measured by computed tomography or MR imaging, is a more relevant variable. Melissant et al. (173) found that computed tomography estimates of the smallest cross-sectional tracheal area, as well as plain tracheal radiograms, were unrelated to the lung function. In contrast, a relationship between the corresponding MR estimates and the inspiratory capacity has been demonstrated in large goiters (65). Changes in goiter volume and the tracheal area resulting from ¹³¹I therapy also correlated well in the study by Huysmans et al. (174). However, results have been inconsistent, maybe due to the low precision of the imaging of the small tracheal dimensions (141). Other sources of error may be changes of the tracheal lumen caused by the respiration itself, the applied intrathoracic pressure, and the position of the patient (175, 176, 177). Furthermore, even with a small increase of the tracheal area, the flow rate becomes much less dependent of the luminal space. Thus, tracheal imaging cannot stand alone in the evaluation of respiration in the patient with a goiter. Because asymptomatic tracheal compression is prevalent (91), a flow volume loop should be considered, in particular if the goiter is very large, but in this case less than 15% of ATA members and even fewer of ETA members include a lung function test in the diagnostic setup (8, 9).

D. Fine-needle aspiration biopsy
FNAB provides the most direct and specific information about a thyroid nodule, and it is used by the majority of thyroidologists given a solitary nodule or a dominant nodule in a multinodular goiter (Table 7; Refs. 6, 7, 8, 9). As the cornerstone in the evaluation, it is virtually without complications, inexpensive, and easy to learn. Figure 2 depicts the central role of FNAB in the investigation of nodular thyroid disease, with special emphasis on the solitary or the dominant nodule. The use of FNAB reduces the number of thyroidectomies by approximately 50% (178), roughly doubles the surgical yield of carcinoma, and reduces the overall cost of medical care in these patients by 25% (163).

Diagnostically useful FNAB specimens are obtained in approximately 80% of the cases (Table 8; Refs. 130 and 179). The number of insufficient samples depends on operator experience, number of aspirations, the character of the nodule (cystic or solid), the experience of the cytopathologist, and especially the criteria used for adequacy of a sample. The number of sufficient samples increases if FNAB is guided by ultrasonography (US; Refs. 180 and 181). Rebiopsy will typically halve the number of insufficient FNABs (130, 182). Needle-steering devices and pistol-grip equipment are used by some. The specimens should be smeared immediately (pull-apart technique); most use air drying and staining with May-Giemsa-Grünwald stain, good for cytoplasmic details, or alternatively Papanicolaou’s stain, good for nuclear details (183).

Most cytopathologists agree that it is nearly impossible to distinguish benign from malignant follicular neoplasms on the basis of fine-needle aspiration cytology. In a large 3-yr prospective study from The Mayo Clinic comprising nearly 2000 patients, 17% had suspicious cytological findings, and in 24% of these patients malignancy was found (186). If scintigraphy of the cytologically suspicious nodule suggests that it is hyperfunctioning, also indicated by a suppressed TSH in case of the solitary nodule, such a lesion almost never harbors malignancy. In this situation, as also supported by others (16, 163), surgery can be avoided. In fact, clinicians should refrain from FNAB in scintigraphically hot nodules. Figure 2 shows our suggestion for the role of FNAB in the diagnostic setup. The application of a more complex cytological classification system to follicular lesions has been suggested by one group, but they were unable to demonstrate improved cancer yield with this cumbersome system (187). Large-needle biopsy does not have a higher diagnostic accuracy, i.e., correctly identifying follicular carcinomas, than FNAB; furthermore, large-needle biopsy is more likely to yield an insufficient specimen, and more often hematomas are seen due to this procedure (188). In a smaller prospective study, a smaller needle size seemed to provide more cellular specimens, but diagnostic accuracy was unrelated to needle size (189).

Sonography has the advantage of providing multiple axial, longitudinal, and oblique planes of section and, most importantly continuous real-time needle localization. The maneuver can be performed by freehand, with the needle being inserted through the skin directly into the plane of view of the transducer without a guide or by the use of built-in or attachable needle slots. US guidance is reliable because it allows the maneuver to be performed quickly, with no exposure to x-rays and under permanent visual surveillance. US-guided FNAB dramatically reduces the risk of sampling error, and the rate of adequate material aspirated during the procedure greatly increases (180, 190). In a retrospective Italian study by Danese et al. (181) comparing conventional FNAB with US-guided FNAB in 4697 and 9683 patients, respectively, sensitivity, specificity, and overall diagnostic accuracy were significantly improved when using US guidance. In another retrospective study from California, sensitivity and specificity improved to 100%, and adequacy rate improved from 84% to 93% when compared with conventional FNAB (191).

The true risk of malignancy in the cystic or mixed cystic/solid thyroid lesion is enigmatic to many clinicians. The usefulness of FNAB in thyroid cysts is hampered by the fact that specimens suitable for cytological examination cannot be collected from all cystic lesions. Attempts to add biochemical analysis of thyroid cyst fluid obtained by FNAB to analyze various enzyme concentrations, Tg, and other components has not given clear evidence of any simple test that distinguishes benign from malignant lesions (192, 193). Given clinical factors (Table 2) or cytological findings increasing the likelihood of thyroid malignancy in a euthyroid patient with a cystic hypofunctioning nodule, the risk of neoplasia and cancer is the same as in solid hypofunctioning nodules, and surgery should be recommended (Fig. 2; Refs. 194 and 195). Certain sonographic features (Table 6) are at best suggestive but are no proof of malignancy (4, 10, 196), and the macroscopic appearance of the cyst fluid, cyst resolution after one aspiration, or other characteristics of the cyst provide no additional information in the context of ruling out malignancy (195, 197, 198). Some report of a low rate of accurate diagnosis for the cystic malignant lesion (199) with a high false-negative rate (200), whereas others have been able to obtain diagnostic FNAB in 71% of patients with cystic nodules with no false-negative results (198). If the solid portion of a cystic mass is aspirated under US guidance and cytology is performed after centrifugation of the aspirated fluid, diagnostic accuracy may be improved (199).

The risk of malignancy in thyroid nodules occurring within a multinodular goiter has not been completely clarified, but some authors find a similar frequency in uni- and multinodular goiters (62, 201). This is consistent with the low accuracy of physical examination in detecting thyroid nodules. When diagnostic errors occur in identifying the number of nodules, the results are biased toward classifying multiple nodules as single because the sensitivity of the physical examination is rather low, in the range of 30% (202). The single nodule, if identified by palpation, will often be reclassified as a dominant (or prominent) nodule within a multinodular thyroid gland if sonography is used. Additional nodule(s) identified by sonography will not decrease the a priori risk of malignancy in the index nodule. For example, in nodules larger than 10 mm, the prevalence of cancer is unrelated to whether they can be detected by palpation (203). The possibility of thyroid malignancy should be considered in all patients with multinodular goiters, and the use of US guidance has been shown to enhance the diagnostic efficacy of FNAB (204). It has been recommended that nodules less than 10 mm, detected incidentally, do not require a FNAB (61, 126). However, in the recent study by Papini et al. (205), thyroid malignancy was found in 6% of nonpalpable lesions of 8–15 mm in size in multinodular goiters (9% in solitary thyroid nodules). The risk was similar in nodules smaller or greater than 10 mm. A similar independency of size with regard to malignancy in nonpalpable nodules was found in an earlier study (190). However, whether carcinomas found in nodules other than the index nodule constitute clinically significant cancers or just incidental microcarcinomas remains an unsolved issue, leaving the clinician with no clear-cut guidelines for management. Sonographic features (205) may guide the clinician to include FNAB in nodules other than the index nodule. If scintigraphy is performed, we recommend FNAB in up to two nodules, provided these are scintigraphically cold. This strategy is based purely on pragmatism and not evidence.

So far, results clearly indicate that FNAB remains superior to other diagnostic tests. Further improvement of diagnostic sensitivity could arise from TPO immunodetection as a tool to assist diagnosis of thyroid nodules by FNAB (206, 207). Immunodetection of other candidate molecules, such as the lectin-related molecules, as markers of malignantly transformed thyrocytes may add valuable information in the selection of those nodules that need to be surgically resected, but confirmatory studies are needed (208). At present, telomerase assays do not add significant information to that of FNAB alone (209, 210).

	VI. Treatment

Top Abstract I. Introduction II. Etiology III. Epidemiology IV. Natural History V. Diagnosis VI. Treatment VII. Unresolved Issues and... References

 
A. Multinodular nontoxic goiter
There is no ideal treatment for the simple goiter. This is reflected by the fact that one third of clinicians would refrain from treatment facing a patient with moderate discomfort due to a multinodular nontoxic goiter of 50–80 g in which malignancy has been ruled out (Refs. 8 and 9 ; Fig. 4).

View larger version (21K): [in this window] [in a new window]   FIG. 4. Treatment options in nontoxic multinodular goiter and preferences among ATA and ETA members according to questionnaire surveys (8 9 ). Index case: "A 42-yr-old premenopausal Caucasian woman admitted to your hospital/clinic due to an irregular nontender bilaterally enlarged thyroid. Clinically judged to be 50–80 g, absence of lymphadenopathy. No family history of thyroid disease. No previous external irradiation. No symptoms of thyroid dysfunction or anterior neck pain. The goiter has been present for 3–5 yr, and the patient reports of moderate local neck discomfort. Your investigations, whatever these have been, have demonstrated a benign condition." In the six listed variations, only one variable has been changed.

When it comes to the nontoxic multinodular goiter, there are only three controlled studies (69, 223, 224) in which sonography was used for objective size monitoring (Table 10). In the nonrandomized study of Lima et al. (223), the control group only comprised 15 patients. Thirty per cent of patients were regarded as responders (>50% reduction of the total nodular volume), whereas partial responders (20–50% reduction) constituted 23%. In the control group, 87% showed no change or an increase in goiter size. Berghout et al. (69), in a randomized double-blind trial, showed that the goiter volume was only reduced by 25% within a period of 9 months in the subgroup comprising 58% of the patients, who responded significantly to L-T₄ therapy (>13% decrease of volume). In the placebo group, the goiter continued to increase in size by more than 20% on average after 9 months. If this rather high growth rate in the placebo group was taken into account, L-T₄ therapy caused a 45% size reduction. In the L-T₄ group, the goiter volume returned to baseline values 9 months after discontinuation of the therapy. Finally, Wesche et al. (224) compared L-T₄ with ¹³¹I therapy in a randomized trial (Fig. 5). Among ¹³¹I-treated patients, L-T₄ was commenced in 45% in the follow-up period due to increasing serum TSH levels. The median reduction of goiter volume in the ¹³¹I-treated group was 38% and 44% after 1 and 2 yr, respectively, whereas the corresponding values in the L-T₄-treated group were 7% and 1%, respectively. The difference in number of responders, 97% (¹³¹I) vs. 43% (L-T₄), was highly significant (224). Disfavoring L-T₄ therapy, more than one third developed symptoms of thyrotoxicosis, and a significant bone mineral loss as evaluated by dual-energy x-ray absorptiometry scanning was also noticed in that group (224). The above-mentioned trials (69, 223, 224) show that the effect of L-T₄ is at best very modest and that a sufficient goiter reduction can only be expected in a minority of the patients.