During the past decade, improved technology has increased the sensitivity of many imaging devices, resulting in the discovery of subclinical nodules in the adrenal, pituitary, and thyroid glands [2-4]. The increasing use of sensitive, high-frequency ultrasonography has led to the identification of nonpalpable thyroid nodules during nonthyroidal ultrasonographic examination of the neck [2]. The discovery of one or more nodules within an otherwise clinically normal thyroid gland raises concern about malignancy and creates a difficult treatment decision for clinician and patient. These lesions, which are referred to as "incidentalomas," are small and nonpalpable and are incidentally discovered on ultrasonography. Although most authorities recommend fine-needle aspiration biopsy for palpable nodules, the optimal method for treating nonpalpable nodules is a matter of controversy. We review the frequency and clinical significance of incidentally discovered, nonpalpable thyroid nodules and offer a practical approach to their treatment.

Data Sources

We reviewed relevant articles published in major English-language medical journals during the past 15 years. We included prospective and retrospective studies that evaluated the prevalence of incidental thyroid nodules as determined by neck ultrasonography, by other imaging studies for thyroid and nonthyroid diseases, and by autopsy studies.

Search Strategy

The MEDLINE and Current Contents, Sciences Edition, computerized databases were used to search the medical literature published in the past 15 years. We used the keywords thyroid nodule and goiter, nodular, and we did keyword and textword searches using the terms occult, incidental, impalpable, and unexpected. We obtained the full text of the articles that met our criteria. We also reviewed articles cited in the articles identified in our database searches. To formulate our algorithm for approaching thyroid incidentalomas, we also searched using the textword thyroid neoplasms in combination with the previous textword searches.

Data Extraction

We examined each article identified in our search and determined which were eligible. A total of 135 articles and abstracts were reviewed. Three articles reported autopsy findings, and 11 reported ultrasonographic findings, either in comparison with findings on clinical palpation or as part of prospective studies in a given population. One study compared scintigraphic scanning with palpation. Several other reports on the risk for malignancy in irradiated thyroid glands, the natural history of benign thyroid nodules, and the prevalence of occult thyroid cancer were also reviewed. These articles formed the basis of our recommendations for treating thyroid incidentalomas.

Data Synthesis

Reliability of Clinical Examination

Because the thyroid gland is located superficially, it is easily palpated. There is disagreement about whether thyroid palpation is best done from the front or the back of the patient and about which system best describes the size of the thyroid gland-one based on estimated weight or one based on other variables, such as the presence of a visible prominence [5]. However, no study has compared the results of different methods of thyroid examination or size determination, which makes it difficult to recommend the use of one method over another [6, 7].

Most nodules that are 1 cm in diameter or larger can be palpated, especially when they are favorably situated. A careful examination should record the size, shape, and consistency of the gland and the number, dimensions, and consistencies of any nodules. A nodule that is located deep within or on the posterior surface of the gland is more difficult to palpate than is one located on the anterior surface [2]. In patients with short, fat necks, nodules may be extremely difficult to detect. Moreover, even with experience and careful technique, physicians may fail to detect many nodules smaller than 1 cm in diameter [8]. In a study by Brander and colleagues [2], one half of the nodules discovered on ultrasonography had escaped detection on clinical examination; approximately one third of the nodules that had not been detected by palpation were larger than 2 cm in diameter. However, a prominent but normal thyroid gland in a patient with a thin neck may be perceived by an examiner as an abnormality of the thyroid [5].

The accuracy of thyroid palpation depends greatly on the experience of the examiner. Interobserver variation in nodule examination has been assessed in two studies. Brander and colleagues [2] discovered a good correlation among examiners in the assessment of thyroid size and classification of nodularity. Most of the examinations were done by internists, but some were done by residents. In contrast, Veith and coworkers [9] found that in one third of cases, examiners disagreed about the number of nodules present. In another study [10], interobserver variation was shown to be less among examiners who had more experience than among those who had different levels of training. We found no study that compared the accuracy of palpation done by thyroidologists or endocrinologists with that of palpation done by general internists.

Clinical palpation is thus not a precise tool for assessing abnormality of the thyroid gland, and its reliability is influenced by the size and location of the nodule, the size and shape of the neck, and the experience of the examiner. In one study [11], the sensitivity of palpation of the thyroid gland in terms of size and nodularity was 38%.

Autopsy Data

In 1955, Mortensen and colleagues [12] examined thyroid glands removed during autopsy from 821 patients at the Mayo Clinic. These glands had all been found to be normal on clinical examination. The authors reported that 406 glands (49.5%) contained one or more nodules; 306 of these (37.3% of 821) were multinodular, and 100 (12.2% of 821) contained single nodules. Of the 406 nodular glands, 144 (35.5%) had nodules that were larger than 2.0 cm in diameter.

In an autopsy study of 200 patients with nodular goiter, Hermanson and associates [13] compared the clinical evaluation of thyroid nodularity with the results of pathologic examination in 190 patients. Of 137 patients who had solitary nodules found on clinical examination, 43 (31%) had several nodules found on pathologic examination. In an autopsy series of 215 patients who did not have thyroid disease, Furmanchuk and coworkers [14] documented nodules in the thyroids of 70 patients (32.5%).

Thyroid Imaging Studies

Current ultrasonographic technology permits high-resolution imaging of the thyroid gland that is more accurate than clinical palpation or other imaging techniques [8, 15, 16]. Ultrasonography is safe and sensitive and is capable of detecting lesions as small as 1 to 3 mm in the thyroid parenchyma [17]. Katz and colleagues [18] reviewed the accuracy of thyroid ultrasonography in 28 thyroid glands examined at autopsy. The correlation between the ultrasonographic finding of thyroid nodules and the pathologic finding of adenomatous goiter was good; ultrasonography thus had a sensitivity of 89% and a specificity of 84%.

Tan and associates [8] recently reported that in 151 patients with a clinical diagnosis of a solitary thyroid nodule, ultrasonography showed that 73 (48%) had other nodules (Table 1). Among the patients in whom subclinical nodules were discovered on ultrasonography, 49 (67%) had two nodules and the other 24 (33%) had three or more nodules that had escaped clinical detection. In that report [8], 89% of clinically palpable nodules were 1 cm in diameter or larger. In 72% of patients with more than one nodule, nodules that had not been identified by palpation were smaller than 1 cm in diameter.

In a retrospective analysis, Brander and colleagues [2] compared results of clinical examination with those of ultrasonography and found that only 12 of 32 (38%) clinically "solitary" nodules were truly solitary on ultrasonographic examination; 15 patients (47%) had several nodules, and 5 had normal glands. As did Tan and associates [8], Brander and colleagues also found that most nonpalpable nodules were smaller than 1 cm in diameter. Walker and coworkers [19] reported that of 200 patients with nodules that appeared to be solitary on clinical examination, 39 (20%) had more than one nodule found on ultrasonography.

The sensitivity of thyroid scintigraphy done using technetium Tc 99m pertechnetate was evaluated in the diagnosis of nodular glands. Arnold and colleagues [20] showed that among patients who had clinically normal thyroid glands or equivocal findings, 40% had evidence of one or more lesions on scintigraphy.

From these studies, it is clear that thyroid incidentalomas are common in apparently normal glands and in glands with solitary nodules and that they are detected on thyroid ultrasonography, scintigraphy, or both. A nodule smaller than 1 cm in diameter often escapes clinical palpation unless it is located superficially.

Prevalence Studies

In 1982, Carroll [21] found at least one incidental thyroid nodule in 13% of patients who had carotid ultrasonography (Table 2). In another study, Horlocker and associates [22] used a 10-MHz system to examine 1000 consecutive patients for suspected parathyroid disease and found an abnormal thyroid gland in 462 (46.2%). Of 709 patients who had clinically normal thyroids, 290 (40.9%) had abnormal findings on a thyroid radioisotope scan; nodules or masses were seen on the scans of 272 patients (38.4%). Stark and colleagues [23] reported a similar prevalence of thyroid incidentalomas (40%) in patients with hyperparathyroidism. In a recent prospective study that examined the prevalence of asymptomatic thyroid nodules in a North American population [24], only 33 of 100 patients (33%) had normal thyroid glands on ultrasonography. Among the 67 patients (67%) who had thyroid nodules, 30 (45%) had more than one nodule and 15 (22%) had solitary nodules.

Three studies done outside of the United States reported a lower prevalence of thyroid incidentalomas. Brander and coworkers [25] noted thyroid incidentalomas in 27% of patients studied in Finland, whereas Woestyn and colleagues [26] reported them in 19% of patients studied in Belgium, and Tomimori and associates [27] noted them in 17% of patients studied in Brazil. Genetic factors, iodine intake, and demographic features probably account for these differences.

Risks for Malignancy of Incidentalomas

We could not identify longitudinal studies that evaluated the outcome of patients with nonpalpable thyroid nodules. In the autopsy study by Mortensen and colleagues [12], most of the nodules examined histologically were classified as benign, and only 4.2% of nodular glands contained malignant nodules. Hermanson and associates [13] found 25 malignant nodules (12.5%) among 200 nodular growths examined at autopsy. In an autopsy study of 138 patients with no known thyroid disease, Komorowski and Hanson [28] reported malignant nodules in only 3% of thyroid glands. In a series by Horlocker and associates [22], only 14 (2%) incidental thyroid carcinomas were detected among 689 patients who had neck exploration for parathyroid disease. These included 6 clinically palpable nodules that had a mean diameter of 3.0 cm and 7 nodules that had a mean diameter of 1.0 cm and were identified only on ultrasonography. In addition, 1 patient had an occult carcinoma that was not detected ultrasonographically. The low risk for malignancy was also confirmed by Brander and coworkers [25], who reported that none of 30 patients who were randomly selected for fine-needle aspiration biopsy had thyroid cancer.

The prevalence of occult thyroid cancer in the United States ranges from 0.45% to 13%, with an estimated mean prevalence of 3.9% ± 4.1% [29, 30]. Similar prevalences have been seen in populations outside the United States [31-33]. Japanese persons living in Nagasaki, Hiroshima, or Sendai and those living in Hawaii had prevalences between 18.9% and 28.4% [31, 34, 35]. Most of the occult carcinomas were cases of papillary cancer. In an extensive study by Fukunaga and Yatani [31] of 67 occult carcinomas in different populations, all but 1 case were cases of papillary cancer. Most patients with occult cancer were in the fourth to seventh decades of life; no significant difference was seen in prevalence between the sexes [31]. Sampson and colleagues [29] compared the prevalence of thyroid cancer in Minnesota with that in Japan and found that differences in age, sex, and exposure to radiation did not influence prevalence.

These results suggest that geographic differences, including genetic, social, and cultural factors, should be considered when patients are examined for occult cancer, especially because no differences were noted in the prevalence of occult cancer between Japanese persons living in Hiroshima and those living in Hawaii [29].

Exposure of the upper body to radiation increases the risk for nodular growth in thyroid glands [36]. Numerous studies have confirmed the increased risk for malignancy in these nodules. In a prospective study of 2118 patients who had been exposed to radiation and had had surgery, Deaconson and coworkers [37] found a frequency of malignancy in nodules as high as 50%. Most of the malignant lesions were papillary carcinomas. Other studies have shown the overall incidence of malignancy in irradiated glands to be as high as 32% to 57% [36, 38].

Papillary cancer has an indolent course and an excellent prognosis [39]. Because most occult thyroid carcinomas are papillary and are not aggressive, conservative treatment is appropriate. Therefore, despite the high prevalence of occult thyroid cancer in Japanese persons, it is not surprising that the rate of death from thyroid cancer is low [29].

Possible Outcome of Thyroid Incidentalomas

It is known that most asymptomatic thyroid nodules are benign lesions. Kuma and associates [40] recently examined the long-term outcome of benign thyroid nodules that were untreated. Among 134 patients with cytologically benign thyroid nodules who were followed for 9 to 11 years, most nodules remained benign. Only 1 patient (0.7%) had a nodule that had been considered benign but that increased in size during follow-up; after surgical excision, this nodule was shown to be a papillary carcinoma. This clearly shows that benign thyroid nodules remain benign for a long time. Therefore, we speculate that the long-term course and final outcome of thyroid incidentalomas and palpable benign nodules would be similar if these conditions were left untreated.

Recommendations

Top Recommendations Author & Article Info References

Nonpalpable thyroid nodules found on imaging studies (thyroid incidentalomas) are usually smaller than 1.5 cm in diameter and present a management dilemma for clinicians. The incidence of incidentalomas appears to be high in the general population. Because of the continued widespread use and high sensitivity of ultrasonography, many small, nonpalpable thyroid nodules will be discovered incidentally during the course of carotid, parathyroid, or other ultrasonographic examinations of the neck. Furthermore, it appears that ultrasonography shows one or more additional nodules in about 50% of patients with clinically palpable solitary nodules.

An important question is, How should nonpalpable nodules be managed? We are not aware of any data on the potential for malignancy in nonpalpable incidental nodules. Most incidentalomas are histologically benign, and benign nodules remain so for a long time. The results of autopsy and ultrasonography suggest that fewer than 5% of asymptomatic nodules may be malignant. The great disparity between the relatively high prevalence of unsuspected nodules in the general population and the low frequency of malignancy in these nodules makes a conservative approach logical. Factors in the medical history that influence a more aggressive approach include a family history of thyroid cancer and a childhood history of neck or head irradiation. Ionizing radiation increases the incidence of benign and malignant thyroid nodules [41]. It has been estimated that 30% to 50% of palpable thyroid abnormalities in previously irradiated glands indicate thyroid carcinoma [39]. Most authorities have recommended surgical excision of nodules for patients who have been exposed to radiation; however, in a decision analysis by Stockwell and coworkers [42], results of a "no scan-reexamine" strategy were slightly better than those of a "scan-surgery" strategy with regard to surgical morbidity and mortality. The effect of fine-needle aspiration biopsy on management was not considered in that analysis. Ultrasonographic findings that suggest malignancy in a thyroid nodule (such as a hypoechoic pattern, an incomplete peripheral halo, an irregular margin, or internal microcalcification) should encourage a more aggressive approach [8, 43, 44].

In conclusion, in most patients with nonpalpable thyroid lesions, the risk for clinically malignant disease is low. Moreover, occult papillary tumors that are smaller than 1.5 cm in diameter grow slowly, and patients with such tumors have an excellent prognosis. It does not seem necessary, practical, or cost-effective to do a biopsy on or to surgically excise all nonpalpable nodules. Our treatment strategy is to observe incidentalomas that are smaller than 1.5 cm in diameter in patients who have a low risk for thyroid cancer (Figure 1). On the other hand, because nonpalpable papillary thyroid cancer is sometimes associated with local or even distant metastases, the incidentaloma cannot be dismissed as unimportant and careful follow-up is necessary. In patients who have nodules larger than 1.5 cm in diameter, a history of head or neck irradiation (particularly in childhood), a strong family history of thyroid cancer, or ultrasonographic findings that suggest malignancy, the next appropriate step is ultrasonographically guided biopsy. Treatment should then be determined on the basis of the cytologic diagnosis. We do not advocate thyroxine suppressive therapy for either a cytologically proven benign nodule that is larger than 1.5 cm in diameter or a benign-appearing nodule that is smaller than 1.5 cm in diameter detected ultrasonographically [45]. Therefore, most nonpalpable nodules do not require immediate treatment and should be followed up simply with palpation. Biopsy should be done if and when the nodules become palpable. During follow-up, costly imaging with ultrasonography or radioisotope scanning is rarely indicated.

View larger version (24K): [in this window] [in a new window]   Figure 1. Algorithm outlining an approach to thyroid incidentalomas. In patients who have a history of exposure to radiation during childhood or those who have ultrasonographic features that suggest malignancy, fine-needle aspiration biopsy should be done. Lesions that are proven to be malignant on biopsy should be treated surgically; a finding of benign nodules should be followed up by palpation at 6 months and annually thereafter. For patients who have nodules that are smaller than 1.5 cm in diameter or that have a benign appearance on ultrasonography, annual neck palpation is sufficient. Routine imaging for patients who have a history of radiation exposure is not recommended; a "no scan-reexamine" strategy [42] is preferable.

This paper was presented in part at the Fifth Annual Convention of the American Association of Clinical Endocrinologists and the American College of Endocrinology in Seattle, Washington, 1 to 5 May 1996.