Otolaryngol Clin N Am 36 (2003) 1733

Evaluation of a thyroid nodule

Namou Kim, MD, Pierre Lavertu, MD, FRCS, FACS*
Department of OtolaryngologyHead & Neck Surgery, Case Western Reserve University School of Medicine, 11100 Euclid Avenue, Cleveland, OH 44106, USA

Thyroid nodules are very common clinical entities. In the United States alone, approximately 275,000 new thyroid nodules are detected annually [1]. The prevalence varies according to the method of screening used. Clinical palpation suggests a prevalence of thyroid nodules of 1% to 7% [1]. In a prospective study conducted in Framingham, Massachusetts, in the 1950s and 1960s, the estimated lifetime risk of developing a nodule was projected to be between 5% and 10%. Five thousand, one hundred and twenty-seven persons were examined in this cohort; 4.2% had a thyroid nodule (6.4% of females and 1.6% of males) [2]. An additional 1.3% developed nodules during a 15-year follow-up. Another study from Wickham, England reported a prevalence of 3.2% in a randomly selected sample population [3]. Ultrasound studies revealed that the prevalence of thyroid nodules in the general population was between 19% and 46% [4]. This increase in prevalence is due to the discovery of smaller nodules that are undetected by palpation (mostly smaller than 1.5 cm). In addition, ultrasound studies revealed that a signicant number of clinically solitary nodules occur in the context of multiple nodules. Brander et al [5] reported that only 38% of clinically solitary nodules were demonstrated to be solitary on highresolution ultrasound. The true prevalence of thyroid nodules can be extrapolated from studies involving pathologic examinations of thyroid glands. Mortensen et al [6] reported a nodule prevalence of 49.5% in 821 thyroid glands that were removed during autopsies of patients with clinically normal thyroid glands. This may be an underestimation of the true prevalence of nodules because only nodules that were larger than 1cm in diameter were counted. Regardless of the incidence of thyroid nodules, the critical question remains whether a detected nodule is malignant. In the past few decades,
* Corresponding author. E-mail address: pierre.lavertu@uhhs.com (P. Lavertu). 0030-6665/03/$ - see front matter 2003, Elsevier Science (USA). All rights reserved. PII: S 0 0 3 0 - 6 6 6 5 ( 0 2 ) 0 0 1 3 0 - 5

18

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

many advances in diagnostic capabilities have helped clinicians to identify malignant nodules. These include radioisotope scans, ultrasonography, CT and MRI scans, and ne needle aspiration (FNA). Despite all of these tools, no single test is entirely accurate and the clinician must be judicious in their use.

Clinical evaluation History and physical examination The usual presentation of a thyroid nodule is an asymptomatic mass that is discovered by either the patient or the clinician. Upon initial contact, a systematic inquiry of the patients medical history is necessary. Factors that increase the risk of malignancy include: previous head and neck irradiation, rapid growth, symptoms of compression or invasion such as dysphagia, dysphonia, and hemoptysis, male sex, pain, age younger than 20 or older than 60, family history of thyroid cancer or multiple endocrine neoplasia. The appearance or enlargement of a nodule while receiving thyroid suppression therapy is especially troublesome. Physical examination by palpation is subject to the clinicians experience and the patients body habitus. Nodules of at least 0.5 cm to 1.0 cm can usually be detected by palpation, although estimates of nodule size vary from physician to physician. It can be dicult to palpate any nodule in a patient with a thick, short neck. Worrisome ndings upon physical examination include nodules that are larger than 4 cm, xation to adjacent skin and soft tissue which may indicate extraglandular invasion, rm nodules, as they may more frequently harbor cancer than softer nodules, and palpable cervical lymphadenopathy in the presence of a thyroid nodule, which raises the possibility of regional metastasis. Although a solitary nodule palpated within a diuse, irregular, rm thyroid gland may indicate chronic thyroiditis, the possibility of malignancy cannot be excluded because 14% to 20% of cancerous glands demonstrate diuse or focal thyroiditis [7]. A laryngeal examination should be an integral part of the initial physical examination. The presence of a vocal cord paralysis suggests a malignant process with involvement of the recurrent laryngeal nerve. The proximal trachea may also be inspected for possible compression or intraluminal extension of a neoplasm. A barium swallow study or an esophagoscopy may be indicated in the presence of dysphagia. In summary, a careful and systematic history and examination may reveal features that are suggestive of malignancy. This is of paramount importance when other diagnostic tests are equivocal or inconclusive, because the decision to proceed with surgery may be based solely on risk factors.

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

19

Laboratory evaluation Laboratory analysis is an essential part of the evaluation. These various measurements are complementary because none of the available ones, except for possibly the calcitonin assay, is an accurate predictor for malignancy. The laboratory values most commonly obtained include thyrotropin (TSH), T4 and T3, thyroglobulin, thyroglobulin antibody, erythrocyte sedimentation rate, thyroid peroxidase antibody (TPO antibody, microsomal antibody), CBC, and calcitonin. The functional status of the thyroid gland has to be evaluated when working up a solitary thyroid nodule. A total T4 and total T3 are often obtained during the initial clinical visit. Their levels can uctuate based on the circulating level of thyroid-binding globulin levels. A more accurate assay is the free T4 level which correlates better with the activity of thyroid hormones in the body. A more useful assay is the high-sensitivity thyrotropin assay (TSH). This assay is the only test that is necessary to detect abnormalities in thyroid function. This high-sensitivity thyrotropin assay will pick up thyroid function abnormality before uctuations in T4 can be detected. Most patients with thyroid nodules are euthyroid, and thus thyrotropin will most often be normal. When an abnormal thyrotropin result is identied, levels of T4 and T3 should be obtained. A reduced thyrotropin level may indicate hyperthyroidism, such as an autonomic functioning gland or thyrotoxicosis, whereas an elevated thyrotropin level indicates hypothyroidism or thyroiditis. In both of these situations, the gland can be enlarged or nodular. A complete blood count and ESR can be obtained if the clinician suspects an inammatory or infectious thyroiditis. A thyroid peroxidase antibody assay is useful for the diagnosis of Hashimotos thyroiditis. The diagnosis of Hashimotos thyroiditis does not decrease the risk of malignancy and in fact, there is a higher incidence of carcinoma and lymphomatous lesions in these glands. A thyroglobulin level can also be obtained. Many factors exist that may produce falsely elevated or decreased levels. Some of these factors include the degree of thyrotropin receptor stimulation, the volume of the gland itself, the presence of thyroid inammation, injury to the gland (radiation, hypoxia, hemorrhage, biopsy, or surgery), multinodular goiter, decreased renal clearance, tobacco smoking, and estrogen level [8]. One of the major limiting factors of the serum thyroglobulin assay is the presence of thyroglobulin antibodies. They may be present in up to 10% of normal subjects and in approximately 15% to 30% of patients with dierentiated thyroid cancers [8,9]. Thyroglobulin antibodies are present in 100% of patients with Hashimotos thyroiditis and in between 89% to 98% of patients with Graves disease [911]. Benign and malignant conditions can produce elevated levels of thyroglobulin. A preoperative assay cannot be used to diagnose or exclude cancerous lesions and is probably not relevant

20

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

when evaluating a thyroid nodule. The best use of this test is in the followup of patients with thyroid cancer after thyroidectomies. Although not used routinely, calcitonin level should be considered in high-risk patients, such as patients with familial medullary thyroid carcinoma or multiple endocrine neoplasia. Recent studies advocated the routine use of calcitonin assay, followed by a pentagastrin-stimulating test if an abnormal calcitonin level is detected in the evaluation of patients with nodular diseases. This is suggested for the following reasons. First, the incidence of sporadic medullary thyroid cancer seems to be higher than previously estimated, with a prevalence of 0.6% to 1.4% in patients with nodular thyroid glands [12,13]. Second, sporadic medullary thyroid carcinoma is very dicult to diagnose on FNA because of its various morphology [14]. Unlike familial medullary thyroid cancer which is diagnosed earlier as the result of family history and gene testing, sporadic medullary thyroid cancer presents at a later stage, often with regional metastasis. A 10-year survival of 60% to 70% is reported in this group [13,15]. A routine calcitonin level may allow identication of this aggressive tumor at an earlier stage (microscopic tumors without nodal metastasis). Further investigation is needed before oering this assay on a routine basis. A survey that was conducted by the American Thyroid Association in 1998 showed the diagnostic trend of endocrinologists in North America. In this survey, a ctional patient with a 2 cm by 3 cm solitary nodule was presented. Ninety-nine percent of physicians obtained a thyrotropin level. Sixty-one percent also included a serum T4 or free T4 level. Thyroid peroxidase antibodies were obtained by 36% of respondents. Thyroglobulin antibodies, thyroglobulin level, and serum calcitonin were included in the initial work-up by 18%, 4.9%, and 5% of respondents, respectively [16]. Radioisotope imaging Nuclear scans of the thyroid, once the cornerstone of thyroid nodule evaluation, have fallen out of favor in the past few decades. Recent surveys by the American Thyroid Association and the European Thyroid Association showed that 23% and 66% of endocrinologists, respectively, would still routinely obtain a nuclear scan in the evaluation of a solitary nodule [16,17]. The most commonly used radioisotopes are technetium (99mTc) and 123I. The choice of radioisotope is dependent on the preference of the clinician and radiologist, because they provide similar information. 123I is more physiologic than 99mTc. 99mTc quickly washes out of the thyroid gland before being organied inside the gland. This property of 99mTc allows for a shorter scanning time (2030 minutes) and the scanning can be performed immediately after the administration of 99mTc. 123I imaging needs to be performed 24 hours after administration of 123I and the scanning time can run 4 to 6 hours in length. Radiation exposure is comparable for both

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

21

agents, and is not signicant. The whole body exposure from 123I and 99mTc scanning is 0.04 cGy and 0.07 cGy, respectively [18,19]. Imaging resolution is better with 99mTc than radioiodine. Nodules that are smaller than 1 cm cannot be detected reliably by either scan, as they are below the discriminating power of scintigraphic devices. Approximately 80% to 85% of nodules are cold on scintigraphy, with 14% to 22% of them ultimately proven to be malignant. Five percent of nodules are hot with less than a 1% risk of malignancy. The remaining 10% to 15% are warm or indeterminate nodules. It was suggested that these nodules harbor a higher risk of malignancy than hot nodules, with a reported range of less than 10% up to 36% [20,21]. Thus, when grouping cold and warm nodules together, the sensitivity of scintigraphic scans for cancer diagnosis is 89% to 93%, but specicity is only 5% with a positive predictive value of only 10% [20]. Although scintigraphy does not oer any additional diagnostic value, it is a powerful adjunct test to more accepted modalities, such as FNA and ultrasound. The following circumstances would be indications for nuclear scans [18,20].  Identication of a functional solitary thyroid nodule when initial serum thyrotropin is decreased.  If an FNA is reported as follicular neoplasm or suspicious, the nding of a hot nodule may decrease the suspicion of a malignancy.  Detecting neck metastasis. Recently, Thallium-201 scan was reported to be a useful diagnostic tool to dierentiate between benign and malignant thyroid nodules [2224]. Sinha et al [24] presented a series of 78 patients with nonfunctioning (cold) solitary nodules. Three mCi of Thallium-201 was used to image the thyroid gland and the nodule uptake was categorized into low-, intermediate-, or high-uptake. All patients then underwent thyroidectomies and histopathology was correlated with the Thallium-201 scan. Risks of malignancy for low-uptake, intermediate-uptake, and high-uptake lesions were 0%, 6%, and 55%, respectively. They concluded that Thallium-201 imaging was a useful adjunct to FNA in the evaluation of solitary nodules, especially when the FNA is inconclusive. Further investigation is needed before incorporating this new modality in the evaluation of the solitary thyroid nodule. Computed tomography and magnetic resonance imaging Although CT scans are a highly sensitive technique in detecting thyroid nodules, they have a very limited role in the initial management of a solitary thyroid nodule. A CT scan can be helpful if the nodule occurs in a diusely enlarged gland that makes palpation dicult. CT scans are more useful in detecting thyroid tissue in retrotracheal and retoclavicular regions and allow the assessment of mediastinal involvement and cervical lymphadenopathy.

22

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

The inherent high iodine content of the thyroid gland will increase the brightness of the gland on CT scan even without contrast material. It is preferable to obtain noncontrasted CT scan because the use of iodine-based contrast will delay any nuclear scintigraphy for 4 to 8 weeks because of the saturation of iodine agents in the cellular components of the thyroid gland. MRI also plays a minor role in the evaluation of the thyroid nodule. MRI is more expensive than CT and US, but it demonstrates exquisite soft tissue details and vascular anatomy. This allows identication of extraglandular invasion and involvement of the great vessels. An advantage of the MRI scan over the CT scan is the possible use of contrast (gadolinium) without interfering with nuclear scintigraphy. A few studies investigated the use of MRI to study nodules in dierent functional states. Tancredi et al [25] reported the use of gadolinium-enhanced T1 ash-2D sequence MRI in the evaluation of thyroid nodules. In a series of 31 patients, in whom an FNA cytology was suspicious or malignant, each thyroid nodule was evaluated for maximum enhancement pre- and postgadolinium. Thyroidectomies were performed and histopathology was compared with the imaging studies. Based on the percent dierence between post- and pregadolinium enhancement, the sensitivity and specicity of detecting a cancerous nodule were 42% and 76%, respectively, for nodules that were less than or equal to 5 mL, and 84% and 100%, respectively, for nodules that were larger than 5mL. They concluded that this technique could be useful for larger nodules (>5 mL), when an FNA was suspicious, and could possibly decrease unnecessary surgeries. Further investigation is probably needed before drawing nal conclusions and including it as a useful adjunct to ultrasound and FNA. Ultrasound Ultrasonography is the most widely used imaging technique for the evaluation of thyroid nodules. Modern ultrasound is performed with high frequency transducers (713 MHz) and can detect solid nodules of 3 mm to 4 mm and cystic nodules of 2 mm in diameter. When routinely used for solitary nodules, it can discover coexisting nodules in approximately 50% of patients. The routine use of ultrasound for solitary nodules was investigated by Marqusee et al [26] at the Thyroid Nodule Clinic of Brigham and Womens Hospital. Ultrasonography changed the clinical management of 44% of patients who were referred for solitary nodules. The ndings that altered management included the discovery of multiple nodules, no actual nodule identied, and very small solitary nodules (<1 cm). Based on their series, the investigators concluded that routine US should be seriously considered in the evaluation of solitary nodules. Nodules can be solid or cystic on ultrasound. Purely cystic nodules are uncommon (1%), with partially cystic lesions accounting for up to 20% of nodules. Cystic lesions were reported to carry a lower risk of malignancy (0.5% to 3%). The discovery of a purely cystic nodule should not discourage

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

23

a needle aspiration for cytologic analysis, however. Predominantly solid nodules carry a higher risk of malignancy (10%). Many studies looked into the echogenic pattern of nodules to predict malignancy, but currently, none of them has discovered a denitive pattern [27,28]. The characteristics that suggest a malignant nodule are described in Box 1. A hypoechoic signal is more typical of a malignant nodule, although many benign nodules also demonstrate hypoechogenicity. Benign nodules, such as adenomas, are generally surrounded by a well-dened capsule, and thus demonstrate a halo sign on ultrasound. Calcium deposits can be ne, punctate ndings in papillary cancer that correspond with Psammoma bodies histologically, but can also be coarse and amorphous. Kakkos et al [27] reported a series of 82 solitary thyroid nodules that were imaged with ultrasound and managed by surgical excision. Ultrasound showed that 22 patients had calcication in their thyroid glands. Histopathological slides were compared with preoperative US. They noted a malignancy incidence of 55% (12 out of 22) in patients with solitary nodules with calcications versus 23% (14 out of 60) for patients with noncalcied nodules. In another study, Takashima et al [29] reported a series of microcalcications with a specicity of 93% and positive predictive value of 70% for cancer, albeit with a sensitivity of only 36%. Koike et al [25] applied multiple logistic regression analysis on ve dierent ndings (margin, shape, echo structure, echogenicity, and calcication) in a retrospective series of 329 nodules (all 5 mm) that were imaged with ultrasound. Patients then underwent thyroidectomies and US ndings were correlated with histopathology. Two hundred and sixty-ve patients had nonfollicular neoplasms and 64 patients had follicular neoplasm. Their sensitivity of preoperative diagnosis was 86.5% for patients with nonfollicular neoplasms, and 18.2% for patients with follicular neoplasms. The specicity was 92% and 89%, respectively. An ultrasound is a safe, noninvasive, nonradioactive test that should be ordered judiciously. Recommendations for ultrasound are:  Nonpalpable or difcult to palpate nodules for US-guided FNA.  Follow-up imaging for solitary nodules that are managed medically or by observation.  Nondiagnostic ne needle aspirate (as an adjunct to repeat FNA).

Box 1. Ultrasound features suggesting malignancy Absent halo sign Solid or hypoechogenicity Heterogeneous echo structure Irregular margin Fine calcications Extraglandular extension

24

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

Routine use of US is probably not indicated at this time because of its cost, its subjective interpretation, and the existence of alternative diagnostic tools, such as FNA. The cytology analysis of the nodule still remains the gold standard, but features on ultrasound may provide the clinician with crucial information, especially when a needle aspirate is inconclusive. The best use of US as a diagnostic modality is in combination with FNA. Fine needle aspiration Needle aspiration of the thyroid was pioneered in the 1930s by Martin [30]. Scandinavian workers introduced ne needle aspiration. It eventually gained wide acceptance in the US in the 1980s, and was endorsed by the American Thyroid Association [31]. This technique is now the gold standard in the evaluation of thyroid nodule. There are many reasons for this. (1) Most nodules are benign, and surgical excision is not required in most cases. Even when scintigraphy was part of the routine evaluation, most cold nodules were histologically benign. With the wide acceptance of FNA, there has been a decrease in the number of thyroidectomies performed and an increase in the yield of malignancies in excised glands [32]. a-Mayor et al [33] reviewed their practice pattern for thyroid nodules. Garc They noticed that the rate of surgery in patients with nodules dropped from 89.9% in 1980 to 46.6% in 1993. During the same time period, the frequency of cancer in surgical specimens increased from 14.7% to 32.9%. (2) FNA is a safe and quick procedure with few complications and it does not involve radiation exposure. (3) FNA decreases the overall cost of care by approximately 25%, mainly by eliminating unnecessary surgeries and limiting the use of frozen section intraoperatively. This results in a savings of $500 to $1300 per patient [34]. FNA is performed with a 10 cc or 20 cc pistol syringe coupled with a ne 22, 23, or 25-gauge needle. The nodule is secured rmly between ngers, and while suction is applied, two to six passes are made into the nodule. Suction is released before exiting the nodule or as soon as material appears at the hub of the needle. If the sample enters the syringe, part or the entire specimen may be lost in the syringe [30]. This technique, although simple to perform, is subject to the operators experience. It is dependent on the placement of the needle, the sensitivity of ngertips, and the amount of suction applied. To maximize the yield of FNA, this technique is probably best performed by experienced physicians coupled with experienced cytopathologists. Recently, a technique of needle biopsy without suction was advocated to reduce traumatic blood aspiration and has been particularly useful in smaller nodules. This technique uses the needle for cutting and the capillary pressure of tissue for sampling [34]. When a cytopathologist is not immediately available to handle the biopsy specimens, the operator will have to prepare the material. The choice of smearing technique or biological preparation is based on preference and

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

25

convenience. Air-dried smears and wet smears can be obtained concurrently, as they are complementary. Air-dried smears include the Di-Quick and the May-Grunwald-Giemsa methods. After carefully expelling the sample (to prevent aerosolization) on a glass slide, the aspirate is allowed to air dry. This method is best for immediate reading by a cytopathologist. The airdried smear highlights the background colloid, the cell architecture, and the cytoplasmic details. This technique is helpful for the diagnosis of medullary and lymphoid tumors. The Papanicolaou stain is a wet smear that requires immediate xation with 95% alcohol. This method enhances nuclear morphology, such as grooving and inclusions, and is better suited for detecting papillary cancer [35]. Zubair et al [36] reported a denitive diagnosis in 65% of Di-Quick smears and in 88% of Papanicolaou smears. The aspirate can also be diluted in xative or preservative for later preparation. The aspirate content can be spinned down in the laboratory and cells that have precipitated can be processed as a larger tissue sample (cell block). Cell block processing has the potential of increasing the accuracy of FNA, particularly for FNA that are labeled as suspicious [35]. The aspirate can also be processed by ThinPrep (Cytye Corporation, Boxborough, MA), which is an automated preparation method that results in a concentrated, thin layer of cells with improved nuclear details [34]. This technique requires added experience and is more expensive. Ancillary tests to improve the accuracy of FNA include immunohistochemistry, ploidy studies, molecular markers, and more recently, Reverse Transcription-Polymerase Chain Reaction (RT-PCR), to detect thyroglobulin mRNA and thyrotropin-receptor mRNA [9,34,35]. This RT-PCR technique is most useful when an FNA is obtained from an adjacent lymph node. Arturi et al [37] reported that if the lymph node biopsied was positive for thyroglobulin mRNA and thyrotropin-receptor mRNA, then there was a 100% concordance with metastatic cancer originating from the thyroid gland. Complications from FNA are neither signicant nor frequent. Box 2 includes some of the potential complications. Pain is localized but can also radiate to the ipsilateral ear and may persist for weeks. When a thyroglobulin level is part of the evaluation, it should be obtained prior to the FNA since a transient but substantial thyroglobulin elevation can be noted after an FNA. Seeding along the tract is extremely rare and only one case has been reported [38]. Reports of cytology need to be standardized, and the current guidelines were discussed and reported by the Papanicolaou Society in 1997 [39]. The four recognized categories of FNA are malignant (3.4% to 5%), benign (60% to 75%), suspicious (7.2% to 30%), and insucient (7% to 29.5%) [34,40,41]. The malignant category includes nodules that have unequivocal typical cytologic characteristics of a malignant neoplasm. This category includes papillary, medullary, poorly dierentiated or undierentiated thyroid cancers, lymphomas, and metastatic nonthyroid cancers [34]. Benign reports include hyperplastic colloid nodule in 90% of the

26

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

Box 2. Complications of FNA Pain Hematoma Entry into trachea Transient thyroid swelling Cystic degeneration Transient bradycardia Transient vocal cord paralysis Formation of calcication Necrosis of nodule Capsular pseudoinvasion Fibrosis Transient thyrotoxicosis Elevation of thyroglobulin level

cases and chronic inammatory lesions (Hashimotos thyroiditis, subacute lymphocytic thyroiditis, De Quervains thyroiditis) in the other 10% [34,40]. The category of suspicious lesions on FNA is due to the inability to unequivocally detect cytologic features of either benign or malignant neoplasm. The most common cause of categorizing an FNA aspirate as suspicious is the inability to dierentiate a follicular adenoma from a well-dierentiated follicular carcinoma. Some of the other possible causes of a suspicious report are as follows [34]:        Hu rthle cell neoplasm. Follicular variant of papillary carcinoma. Low-grade papillary carcinoma. Hyalinizing trabecular adenoma. Hashimotos thyroiditis with metaplasia. Any cancer with suboptimal sampling. Adenomatous goiter with microfollicular structure predominance.

Nondiagnostic reports caused by the lack of cellular components in the aspirate or because of improper handling of the specimen. An adequate smear consists of at least ve to six groups of follicular cells, with each group containing at least 10 cells [42]. Operator- and patient-dependent factors exist. The highest rates of nondiagnostic reports are found in centers where multiple clinicians perform FNA [34]. Patient factors include brosis, cystic lesions, calcications, small nodules (<1 cm), and necrosis, among others. The main goal of FNA is to accurately predict which nodule is cancerous. Numerous studies cited the following data. Sensitivity is 65% to 100% and specicity is 70% to 100% [34,43,44]. Overall accuracy is estimated at 92%

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

27

to 95% [41,45]. One of the most recent and largest series was conducted by Amrikachi et al [39] who retrospectively reviewed 6226 consecutive FNAs from 1982 to 1998. They reported a sensitivity and specicity of 93% and 96%, respectively, when the FNA material was adequate for cytological analysis. In most series, the false positive rate is approximately 0.8% to 9%, and the false negative rate is 0% to 16% [40,41,44,45]. Most experts agree that the actual false negative ratio is less than 5% [46]. One of the most common causes of a false negative reading are cystic nodules, especially when larger than 3 cm, because they carry a rate of inadequate sampling of 20% and a false negative reading of up to 30% [34,47]. It was previously believed that cystic nodules were benign, but purely cystic nodules were demonstrated to be uncommon. The presence of colloid is not always a reassuring nding, and it is imperative that a few clusters of normal cells be found to minimize nondiagnostic and false negative FNAs [35]. Pseudocysts are believed to occur in cancerous and benign nodules at a rate of 23% to 33% and 27% to 35%, respectively; cystic nodules should be considered to have the same cancerous potential as more solid nodules [34]. Papillary cancer is the most likely culprit in cystic degeneration of cancerous lesions. Repeat FNA, ultrasound-guided FNA, or excision is recommended in patients with cystic nodules.

Management of the solitary thyroid nodule Surgical excision based on clinical indicators The greatest challenge of the management of a solitary thyroid nodule is to dierentiate between benign and malignant neoplasms. The wide acceptance of the FNA technique has enabled us to be more selective about which patients should undergo surgical excision of the nodule. The following clinical factors may warrant surgery despite an FNA that is reported as benign or non-diagnostic: (1) Age extreme (younger than 20 or older than 45); (2) male sex; (3) nodules that occur in patients with Graves or Hashimotos thyroiditis; (4) previous radiation to the cervical region; (5) strong family history of thyroid cancer; (6) pain, (7) compressive, inltrative/invasive features; (8) cervical metastasis, (9) large nodule (>4 cm), (10) rapid growth; and (11) growth despite thyroid suppression therapy. Some investigators rely mainly on clinical prognostic factors to determine the need for surgery and the extent of surgery. Brooks et al [48] prospectively studied 564 patients between 1996 and 1999, and concluded that preoperative FNA had no direct impact on the selection of the surgical procedure; the majority of thyroid operations for thyroid nodules were planned and performed based on known prognostic and intraoperative factors.

28

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

Management algorithm according to FNA results The natural history of a FNA benign nodule is unpredictable, and, therefore, close observation is recommended. Although yearly follow-up with or without US is probably sucient, some nodules may warrant further work-up when poor clinical indicators are present (see earlier discussion). A repeat FNA in 6 to 24 months is advisable in these patients. This is done to reduce the false negative results of initial FNAs. Hamburger [49] repeated an FNA at a mean of 2.39 years in 205 patients with benign initial FNAs and reported a reversal to suspicious in 12 patients and to malignant in 6 of them [26]. Although it is unclear how cost-eective a routine repeat FNA is, some experts recommend it. Chehade et al [41] reported a series of 235 routine repeat FNAs at a mean of 2.95 years and noticed that the false negative rate decreased from 2.55% to 1.11%. Based on their experience and a literature review, they concluded that a repeat FNA would reduce the false-negativity rate by 50%. Thyroid suppressive therapy was once used as a short-term diagnostic tool, but its diagnostic value has become obsolete because of the advent of FNA and other radiographic studies. Although long-term suppressive therapy is still advocated by some experts, it is now falling out of favor. Several series were conducted to determine the benets of T4 suppressive therapy for patients with benign solitary nodules (Table 1). Gharib and Mazzaferri [50] reviewed the literature and concluded that because only 10% to 20% of patients respond to suppressive therapy, benign solitary nodules are best followed without suppression, thus avoiding potential complications of long-term therapy [21]. The management of FNA malignant nodules is more straightforward because the predictive value of malignancy in FNA is close to 100% with a specicity also close to 100% [35]. Surgical excision is warranted for malignant nodules. The FNA suspicious lesion, which includes follicular neoplasm and Hu rthle cell neoplasm, is probably the limiting factor of FNA. Suspicious FNA results can account for up to 30% of FNAs in some series; the overall malignancy rate of these lesions is 10% to 20% [34]. Some series quoted malignancies in this category as high as 41% to 60% [40,43]. These numbers strongly suggest the need for surgical management.
Table 1 T4 suppressive therapy for solitary thyroid nodules Series Gharib et al [57] Cheung et al [56] Reverter et al [58] Papini et al [51] Mainini et al [52] # Patients 53 74 40 101 55 Duration (months) 6 18 11 12 21 Measurement US P US P, US US Daily T4 3.0 lg/kg >150 lg 2.8 lg/kg 2.0 lg/kg 1.7 lg/kg Benet NS NS NS NS NS

NS, not signicant; P, palpation; US, ultrasound.

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

29

The last category of FNA result is the nondiagnostic or insucient for diagnosis reading. When this result is obtained, a repeat FNA is performed, possibly with ultrasound guidance to increase the yield. Carmeci et al [53] reported that the rate of insucient sampling decreased from 16% to 7% when ultrasound guidance was used. Despite repeated aspirations, a small group of patients will still have nondiagnostic FNAs. McHenry et al [54] reported a 9% incidence of malignancy in 92 patients who underwent thyroidectomies after repeat nondiagnostic FNAs. Schmidt et al [44] reported an incidence of 5% in 59 patients. Their data suggest that nodules with repeated nondiagnostic FNA are probably benign. They concluded that clinical follow-up, with surgical intervention only when poor prognostic indicators were present, was acceptable. In this situation, the management should be tailored to the individual patient. Extent of surgery With the exception of a lesion that is limited to the isthmus, the least extensive procedure recommended is a lobectomy and isthmusectomy [50]. When limited to the isthmus, simple excision of the nodule can be done leaving the two lobes intact. In cases where a denite diagnosis was not made before surgery or in cases of suspected benign lesions where a malignancy still needs to be ruled out, the specimen is often examined on frozen section to determine the need for further surgery at the same setting. An exception to this practice may be in the presence of follicular neoplasm that falls in the suspicious category of FNA results where a frozen section analysis is unlikely to provide further information. The need for a total thyroidectomy in the presence of malignancy is controversial [55]. Although some investigators advocate total thyroidectomy as the procedure of choice for patients with well-dierentiated thyroid cancer, others reserve total or near-total thyroidectomy to the high-risk group of patients. The patients in the high-risk group are older than 45 years of age (also younger than 20), have lesions that are larger than 4 cm, have tumors with extraglandular extension, have regional or distant metastases, or have histologically more aggressive tumors. There is consensus that a total thyroidectomy should be performed for patients in the high-risk group. Although many investigators suggest a similar approach in patients in the low-risk group, the advantages of total thyroidectomy must be weighed against the risks of recurrent laryngeal nerve paralysis and hypoparathyroidism which are denitely increased with the more aggressive operation, especially in the hands of the occasional operator.

Summary Solitary thyroid nodules are common entities and the majority of these are benign. The main goal in evaluating these nodules is to identify the

30 N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

Fig. 1. Management algorithm of a thyroid nodule based on FNA results.

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

31

nodules with malignant potentials. A multitude of diagnostic tests is available to the clinician. FNA biopsy is now considered the gold standard diagnostic test in the evaluation of a thyroid nodule. Ultrasound and nuclear scans are also useful tests, but are best used in conjunction with FNA. Clinical decisions are often based on the results of the FNA. A diagram of the clinical decision process based on FNA ndings is presented in Fig. 1. Despite the clinical advantages that are provided by the FNA results, the clinician should not dismiss the importance of other clinical prognostic indicators to optimize timely treatment.

References
[1] Castro MR, Gharib H. Thyroid nodules and cancer: when to wait and watch, when to refer. Postgrad Med 2000;107(1):11324. [2] Vander JB, Gaston EA, Dawber TR. The signicance of nontoxic thyroid nodules: nal report of a 15 year study of the incidence of thyroid malignancy. Ann Intern Med 1968;68:537. [3] Tunbridge WMG, Evered DC, Hall R, et al. The spectrum of thyroid disease in a community: the Wickham survey. Clin Endocrinol 1977;7:48193. [4] Burguera B, Gharib H. Thyroid incidentalomas: prevalence, diagnosis, signicance, and management. Endocrinol Metab Clin North Am 2000;29(1):187203. [5] Brander A, Viikinkoski P, Nickels J, et al. Thyroid gland: U/S screening in a random adult population. Radiology 1991;181:6837. [6] Mortensen JD, Woolner LB, Bennett WA. Gross and microscopic ndings in clinically normal thyroid glands. J Clin Endocrinol Metab 1955;15:127080. [7] McCaffrey TV. Evaluation of the thyroid nodule. Cancer Control 2000;7(3):2238. [8] Spencer C. Thyroglobulin. In: Braverman LE, Utiger RD, editors. The thyroid. 8th edition. Philadelphia: Lippincott Williams & Wilkins; 2000. p. 40213. [9] Torrens JI, Burch HB. Serum thyroglobulin measurement: utility in clinical practice. Endocrinol Metab Clin North Am 2001;30(2):42967. [10] Amino N. Antithroglobulin antibodies. In: Ingber SH, Braverman LE, editors. The thyroid. 5th edition. Philadelphia: JB Lippincott; 1986. p. 54659. [11] Tamaki H, Karsumaru H, Amino N. Usefulness of thyroglobulin antibody detected by ultrasensitive enzyme immunoassay. Clin Endocrinol 1992;37:26673. [12] Niccoli P, Wion-Barnot N, Caron P, et al. Interest of routine measurement of serum calcitonin: study in a large series of thyroidectomized patients. J Clin Endocrinol Metab 1997;82:33841. [13] Raue F. Routine calcitonin determination in thyroid nodules: an effective approach? Exp Clin Endocrinol Diabetes 1998;106(4):28991. [14] Hahm JR, Lee MS, Min YK, et al. Routine measurement of serum calcitonin is useful for early detection of medullary thyroid carcinoma in patients with nodular thyroid diseases. Thyroid 2001;11(1):7380. [15] Raue F, Kotzerke J, Reinwein D, et al. Prognostic factors in medullary thyroid carcinoma: evaluation of 741 patients from the German medullary thyroid carcinoma register. Clin Invest 1993;71:712. [16] Bennedbaek FN, Hegedus L. Management of the solitary thyroid nodule: results of a North American survey. J Clin Endocrinol Metab 2000;85(7):24938. [17] Bennedbaek FN, Perrild H, Hegedus L. Diagnosis and treatment of the solitary thyroid nodule: results of a European survey. Clin Endocrinol 1999;50:35763.

32

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

[18] McDougall IR, Cavalieri RR. In vivo radionucleide test and imaging. In: Braverman LE, Utiger RD, editors. The thyroid. 8th edition. Philadelphia: Lippincott Williams & Wilkins; 2000. p. 35575. [19] Meier DA, Kaplan MM. Radioiodine uptake and thyroid scintiscanning. Endocrinol Metab Clin North Am 2001;30(2):291313. [20] Cases JA, Surks MI. The changing role of scintigraphy in the evaluation of thyroid nodules. Semin Nucl Med 2000;30(2):817. [21] Sabel MS, Staren ED, Gianakakis LM, et al. Effectiveness of the thyroid scan in evaluation of the solitary thyroid nodule. Am Surg 1997;63(7):6603. [22] El-Desouki M. Thallium 201 thyroid imaging in differentiating benign from malignant thyroid nodules. Clin Nucl Med 1991;16:42530. [23] Hardoff R, Baron E, Sheinfeld M. Early and late lesion-to-non-lesion ratio of Thallium 201 chloride uptake in the evaluation of cold thyroid nodules. J Nucl Med 1991;32:18736. [24] Sinha PS, Beeby DI, Ryan P. An evaluation of Thallium imaging for detection of carcinoma in clinically palpable solitary, nonfunctioning thyroid nodules. Thyroid 2001; 11(1):859. [25] Tancredi M, Foppiani L, Giordano GD, et al. Magnetic resonance imaging: a complimentary tool in the evaluation of thyroid nodules. J Endocrinol Invest 2001;24(5): 384385. [26] Marqusee E, Benson CB, Frates MC, et al. Usefulness of ultrasonography in the management of nodular thyroid disease. Ann Intern Med 2000;133:696700. [27] Kakkos SK, Scopa CD, Chalmoukis AK, et al. Relative risk of cancer in sonographically detected thyroid nodules with calcications. J Clin Ultrasound 2000;28:34752. [28] Koike E, Noguchi S, Yamashita H, et al. Ultrasonographic characteristics of thyroid nodules. Arch Surg 2001;136:3347. [29] Takashima S, Fukuda H, Nomura N, et al. Thyroid nodules: re-evaluation with ultrasound. J Clin Ultrasound 1995;23:17984. [30] Martin HE, Ellis EB. Biopsy made by needle puncture and aspiration. Ann Surg 1930; 92:169. [31] Kaplan MM. Evaluation of thyroid nodule by needle aspiration. In: Braverman LE, Utiger RD, editors. The thyroid. 8th edition. Philadelphia: Lippincott Williams & Wilkins; 2000. p. 44151. [32] Sidawy MK, Del Vecchio DM, Knoll SM. Fine needle aspiration of thyroid nodules: correlation between cytology and histology and evaluation of discrepant cases. Cancer 1997;81:2539. rez Mendez LF, Pa ramo C, et al. Fine needle aspiration biopsy of [33] Garc a-Mayor RV, Pe thyroid nodules: impact on clinical practice. J Endocrinol Invest 1997;20:4827. [34] Belore A, La Rosa GL. Fine needle aspiration biopsy of the thyroid. Endocrinol Metab Clin North Am 2001;30:361400. [35] Orell SR, Philips J. Problems in ne needle biopsy of the thyroid. Pathology 2000;32:1918. [36] Baloch ZW, Tam D, Langer J, et al. Ultrasound-guided ne-needle aspiration biopsy of the thyroid: role of on-site assessment and multiple cytologic preparations. Diagn Cytopathol 2000;23:4259. [37] Arturi F, Russo D, Giuffrida D, et al. Early diagnosis by genetic analysis of differentiated thyroid cancer metastases in small lymph nodes. J Clin Endocrinol Metab 1997;82:163841. [38] Hales MS, Hsu FS. Needle tract implantation of papillary carcinoma of the thyroid following aspiration biopsy. Acta Cytol 1990;34:801. [39] Guidelines of the Papanicolaou Society of Cytopathology for ne needle aspiration procedure and reporting by the Papanicolaou Society of Cytopathology Task Force on standards of practice. Diagn Cytopathol 1997;17:23947. [40] Amrikachi M, Ramzy I, Rubenfeld S, et al. Accuracy of ne needle aspiration of thyroid: a review of 6,226 cases and correlation with surgical or clinical outcome. Arch Pathol Lab Med 2001;125:4848.

N. Kim, P. Lavertu / Otolaryngol Clin N Am 36 (2003) 1733

33

[41] Chehade JM, Silverberg AB, Kim J, et al. Role of repeated ne needle aspiration of thyroid nodules with benign cytologic features. Endocr Pract 2001;7:23743. [42] Goellner JR, Gharib H, Grant CS, et al. Fine needle aspiration cytology of the thyroid. Acta Cytol 1987;31:58790. [43] Boyd LA, Earnhardt RC, Dunn JT, et al. Preoperative evaluation and predictive value of ne needle aspiration and frozen section of thyroid nodules. J Am Coll Surg 1998;187: 494502. [44] Schmidt T, Riggs MW, Speights VO. Signicance of non-diagnostic ne needle aspiration of the thyroid. South Med J 1997;90:11836. [45] Sabel MS, Staren ED, Gianakakis LM, et al. Use of ne-needle aspiration biopsy and frozen section in the management of the solitary thyroid nodule. Surgery 1997;122:10217. [46] Gharib H, Goellner JR. Fine needle aspiration biopsy of the thyroid: an appraisal. Ann Intern Med 1993;118:2829. [47] Meko JB, Norton JA. Large cystic/solid thyroid nodules: a potential false-negative ne needle aspiration. Surgery 1995;118:996. [48] Brooks AD, Shaha AR, DuMornay W, et al. Role of ne-needle aspiration biopsy and frozen section analysis in the surgical management of thyroid tumors. Ann Surg Oncol 2001;8:92100. [49] Hamburger JI. Consistency of sequential needle biopsy ndings for thyroid nodules: management implications. Arch Intern Med 1987;147:979. [50] Gharib H, Mazzaferri EL. Thyroxine suppressive therapy in patients with nodular thyroid disease. Ann Intern Med 1998;128:38694. [51] Papini E, Bacci V, Panunzi C, et al. A prospective randomized trial of levothyroxine suppressive therapy for solitary thyroid nodules. Clin Endocrinol 1993;38:507. [52] Mainini E, Martinelli I, Morandi G, et al. Levothyroxine suppressive therapy for solitary thyroid nodule. J Endocrinol Invest 1995;18:796. [53] Carmeci C, Jeffrey RB, McDougall IR, et al. Ultrasound-guided ne needle aspiration biopsy of thyroid masses. Thyroid 1998;8:283. [54] McHenry CR, Walsh PG, Rosen IB. Non-diagnostic ne needle aspiration biopsy: a dilemma in management of nodular thyroid disease. Am Surg 1993;59:4159. [55] Shaha AR. Controversies in the management of thyroid nodule. Laryngoscope 2000;110: 183193. [56] Cheung PSY, Lee JMH, Boey JH. Thyroxine suppressive therapy of benign solitary thyroid nodules: a prospective randomized study. World J Surg 1989;13:818. [57] Gharib H, James EM, Charboneau JW, et al. Suppressive therapy with levothyroxine for solitary thyroid nodules. N Engl J Med 1987;317:70. [58] Reverter JL, Lucas A, Salinas I, et al. Suppressive therapy with levothyroxine for solitary thyroid nodules. Clin Endocrinol 1992;36:25.