- Email: [email protected]
Diagnostic imaging techniques in thyroid cancer
DIAGNOSTIC IMAGING TECHNIQUES
Diagnostic Imaging Techniques in Thyroid Cancer Michael Friedman, MD, Dean M. Toriumi, MD, and Mahmood F. Mafee, MD, Chicago, Illinois
The main objective in imaging the thyroid is to identify lesions likely to be malignant. Once the diagnosis of malignant disease is confirmed by histologic examination, imaging is useful to delineate the extent of tumor or metastatic disease. In the past, a routine workup typically began with thyroid function studies and a radionuclide scan. Recently, technologic advances and the need for strict cost containment have resulted in an alteration in the common algorithm used for the diagnostic evaluation of a thyroid nodule. In addition to the particular tests and sequence of the workup, the rationales behind them have changed as well. Aside from the need of the physician to obtain information for the treatment decision, the need to provide adequate information to the patient so he or she can participate in that decision is also important. Although recommendations can be made for specific testing and for consideration of operation, there are many situations where the indications for operation or the extent of the resection can be controversial. The final decision on treatment is based on input from three areas: clinical, laboratory findings, and personal factors (Figure 1). The history and physical examination, which includes such important aspects as prior irradiation, family history, and growth rate, as well as consistency, shape, and size of the mass all play a role. As this report discusses, laboratory testing includes a variety of imagFrom the Departments of Otolaryngology-Head and Neck Surgery and Radiology, University of Illinois College of Medicine; the Department of Otolaryngology-Head and Neck Surgery, Illinois Masonic Medical Center; and the Department of Otolaryngology Head and Neck Surgery, Northwestern University Medical School, Chicago, IlUnois. Requests for reprints should be addressed to Michael Friedman, MD, Department of Otolaryngology-Head and Neck Surgery, University of Illinois Eye and Ear Infirmary, 1855 West Taylor Street, Chicago, Illinois 60612. Editor's Note: This specifically solicited material contains an element of case report format that was found essential to convey the education messages of Figures 6, 7, and 8. It does not represent a willingness to review or accept case reports for the Journal.
Volume 155, February 1988
ing techniques that will add information. The physician's attitude and philosophy will play a major role. Finally, the patient's input in terms of his or her anxiety and philosophy must be added to the overall picture when reaching a decision in treatment. Although fine-needle aspiration is not a method of thyroid imaging, it cannot be excluded from this report. Besides being the first step in evaluating nodular thyroid diseases, fine-needle aspiration can be the determining factor in selecting appropriate tests, and it can be used in combination with thyroid imaging. Herein, we present an updated protocol for the use of diagnostic imaging to evaluate a discrete lesion, a diffusely enlarged gland, and regional, recurrent, and distant metastatic disease. We also discuss in depth the current diagnostic applications of plain films, contrast studies, radionuclide scanning, and ultrasonography, as well as computerized tomography and magnetic resonance imaging. The initial evaluation of a thyroid mass obviously begins with a thorough history and physical examination. The head and neck examination should include indirect laryngoscopy to assess vocal cord mobility, which, if impaired, may indicate recurrent laryngeal nerve invasion. Blood tests to evaluate thyroid function can be limited to a thyroxine level if the patient is euthyroid by history and physical findings. On the other hand, if an abnormality in thyroid function is suggested by the history, triiodothyronine resin uptake, free thyroxine, and thyroidstimulating hormone measurements should be obtained. If operation is planned, calcium levels should be obtained preoperatively. Approach to Thyroid Imaging
Solitary lesion: Thyroid neoplasia can present as either a discrete nodule or nodules or a diffusely enlarged gland, although the former is more likely to 215
Friedman et al
SURGERY OF THYROID MASSES .~LABORATORY FNDNG5
CLINICAL
J. PERSONAL TLFACTORS
~FNA
*History
I
=
reported false-negative rate of 0.3 to 10 percent and a false-positive rate of 0 to 2.5 percent [2]. Our experience, as well as that of others, is closer to the 10 percent false-negative rate and less than the 2.5 percent false-positive rate. For practical application, a malignant finding on fine-needle aspiration is a strong indication for operation. In contrast, a negative finding on fine-needle aspiration cannot definitely rule out malignancy. In the case of a benign or inadequate fine-needle aspiration or when fine-needle aspiration is unavailable, other tests and clinical evaluation become more important in planning treatment. When fine-needle aspiration reveals follicular or H~irthle cells, there is some controversy as to whether a distinction between a benign and malignant neoplasm can be made. In most cases, a tissue specimen is needed to document capsular or vascular invasion to confirm malignancy. If fine-needle aspiration indicates that a mass is solid but benign, the next step is to obtain a technetium-99m pertechnetate thyroid scan. This scan can provide an important source of information for the patient regarding the likelihood of malignancy and the frequency of follow-up. Most solid benign lesions are cold on radionuclide scan, with less than 10 percent being
~
H~tDAttitude
,Phys~alExarr~rmtion ,ThyroidScan
~Patient's Anxiety
*GrowthPatternof Mass ,Ultrasound ,Patient's Inputin Decision ,Age & GeneralHealth )OtherLab Studies
Figure 1. Multiple factors Influencing the decision to treat a
thyroid mass surgically.
be malignant. Thyroid nodules are relatively common and can be found in 4 to 7 percent of the population [1-4]. The incidence of solitary thyroid nodules that prove to be malignant on surgical excision ranges from 8 to 33 percent (median 20 percent). Fine-needle aspiration is the initial test used to evaluate a discrete thyroid nodule (Figure 2). Fineneedle aspiration can help classify a thyroid nodule as benign or malignant, as well as differentiate a solid mass from a cystic one. The incidence of inadequate aspiration depends largely on the experience of the person performing fine-needle aspiration, as well as the pathologist. Fine-needle aspiration has a
E V A L U A T I O N OF DISCRETE THYROID NODULE(S)
Fine N e e d l e
/
so,,f \
Malignant
Sur
ry %%
Hot
Follow
A s p i r a t i o n (FNA)
Benign
Cystic
Tc-99m Scan
A s p i r a t e Fluid
Cold
Follow with S u p p r e s s i o n
Ultrasound"
No S m a l l e r
Consider Surgery
216
Follow
Smaller
Follow with S u p p r e s s i o n
Reforms
Ultrasound
J
Purely C y s t i c
Re-aspirate
\
C y s t i c with Solid Component U l t r a s o u n d - g u i d e d FNA of Solid Component
Reforms
Benign
Malignant
Consider Surgery
T r e a t as Benign Solid Lesion
Surgery
F i g u r e 2. Protocol for workup and management of a solitary thyroid nodule. Single asterisk Indicates one prominent nodule in a mul#nodular gland; double asterisk Indicates ultrasonography only If the mass Is difficult to palpate; dagger Indicates computerized tomography or magnetic resonance imaging to assess substernal extension, Invasion, or both. FNA = fine-needle aspiration; T c - g g m = technetlum-ggm.
The American Journal of Surgery
Diagnostic Imaging Techniques in Thyroid Cancer
EVALUATION OF DIFFUSE THYROID ENLARGEMENT
Tc-99m Scan
\
Hot Nodule
Cold Nodule :~
\
FNA (Ultrasound-guided) ~'~
Benign
Follow
FNA (Ultrasound-guided) ~>~
Malignant
Surgery ~
Figure 3. Protocol for the evaluation and
management of a diffusely enlarged thyroid gland. Single asterisk Indicates ultrasonography only ff the mass is difficult to palpate; double asterisk Indicates computerized tomography or magnetic resonance Imaging for large masses to assess substernal extension, Invasion, or both.
hot. The patient should be informed that malignancy occurs in an estimated 10 percent of similar situations, as in benign solid lesion on fine-needle aspiration, but cold on scan. We prefer to treat a patient with a cold lesion that is benign by fine-needle aspiration with a 3 month course of thyroxine (levothyroxine sodium) suppression, with monthly followup to monitor the size of the nodule. Most lesions will not regress, but those that do are almost always benign. Ultrasonography is performed before and after the course of suppression to document size if the mass is not easily palpable. If the mass remains the same size or grows larger, operation is indicated. If a mass is easily palpable and fine needle aspiration has ruled out a cystic lesion, ultrasonography is superfluous. A mass that is benign on fine-needle aspiration and hot on technetium-99m scan may be an autonomous hot nodule and should be followed every 3 to 6 months. The patient can be informed that the likelihood of a malignancy is minimal. We believe that surgical management is usually unnecessary in such cases. On occasion, lesions are seen that are hot on technetium-99m scan but cold on radioiodine scan. These are termed discordant lesions and should be treated as cold nodules. The incidence of malignancy in discordant nodules is 15 percent, and therefore, surgical management should be considered [5]. A nodule that is cystic on fine-needle aspiration
Volume 155, February 1988
Benign
Suppress for 3 Months
( r e p e a t ultrasound) " ~
Smaller
Follow
No Change or Larger
Consider Surgery'f
should be aspirated and then reevaluated monthly for 3 months. If a cyst reforms, ultrasonography can be performed followed by ultrasonographically guided fine-needle aspiration if a solid component is detected. Cystic thyroid nodules are rarely purely cystic and ultrasonography can locate any solid component. Ultrasonographically-guided fine-needle aspiration can improve the yield and accuracy of the technique. If the lesion is indeed purely cystic, repeat aspiration may be attempted. D i f f u s e l y enlarged thyroid gland: Diffuse thyroid enlargement usually represents a benign goiter but may actually hide malignant disease. The initial diagnostic test in the workup of a diffusely enlarged gland is a technetium-99m thyroid scan (Figure 3). If a cold area is isolated but is not palpable on examination, an ultrasonographically-guided fineneedle aspiration can be performed on any irregular region that corresponds with the cold region on the scan. If the fine-needle aspiration findings are benign, the patient is placed on thyroxine suppression for 3 months to shrink the gland and make the cold, nonfunctioning area easier to evaluate. The patient should be examined every month. If after 3 months, the mass has become larger, operation is considered. If the lesion becomes smaller, the patient is followed every 3 months thereafter. Diffusely enlarged glands with a prominent nodule found to be hot on technetium-99m scan are examined every 3 months.
217
Friedman et al
Current Applications of Diagnostic Imaging R a d i o n u c l i d e s c a n s : Thyroid scans remain a first-line diagnostic tool if accurate fine-needle aspiration is unavailable. Scans are also performed to evaluate the functional status of nodules that are benign on fine-needle aspiration. One of the benefits of scanning after fine-needle aspiration is ensuring against false-negative results by further assessing the characteristics of the mass, and the clinician is also protecting himself legally. Also, the results of the scan will help determine the frequency of followup visits for patients on a 3 month course of suppression. Patients with cold nodules are seen approximately every month and patients with hot nodules, every 3 months. If a scan reveals that the nodule is cold and the physician informs the patient that he recommends suppression in spite of a 10 percent chance of malignancy, he has fully informed his patient. Three major types of radionuclide scans are currently available to image the thyroid. Technetium99m is the preferred agent for screening because it has the greatest sensitivity, low radiation exposure, and a short test time of 1 hour compared with 20 to 24 hours for radioiodine. In addition, it is readily available and the cost is relatively low. Technetium99m, which is a d m i n i s t e r e d intravenously, is trapped by the thyroid gland but not organified. Maximal thyroidal uptake is seen in 20 to 40 minutes. The technetium-99m scan can readily identify a nonfunctioning nodule, which does not trap any pertechnetate. A hot or warm nodule seen on early scan (3 minute s ) that becomes cool on late imaging may represent a malignant vascular lesion, which must be ruled out [6]. A hot nodule on late imaging (after 20 minutes) may be a functional nodule but may also represent trapping. A true functional nodule can only be confirmed by radioiodine scan. Radioactive iodine (iodine-131, iodine-125, and iodine-123) has a distinct advantage over all other tracers. Radioiodine is the only agent that is both trapped and organified by the thyroid gland, thus permitting evaluation of the true physiologic features of the thyroid gland and any masses within it. The quality of the image is better than that of technetium-99m. Radioiodine is taken by mouth and requires nearly 24 hours for organification. Iodine123 is the most favorable of the more than 20 radioactive isotopes of iodine. It has a short half-life (13.2 hours), a photon energy that permits good imaging (159 keV), and no particulate emissions [7]. A large r tracer dose can be administered at an acceptable radiation level. Iodine-125 and -131 require much higher radiation levels to yield a comparable image. However, iodine-123 is very expensive and has a short shelf life, rendering it practical only in large institutions where several scans are performed on a daily basis. A nodule that is hot on technetium-99m scan may 218
appear cold on radioiodine scan. Such discordant lesions theoretically trap iodine but do not organify it [7]. The majority of discordant lesions represent benign colloid nodules, follicular adenomas, or thyroiditis, but are nonetheless considered suspicious for Carcinoma, just as other cold nodules. Therefore, nodules that are hot on technetium-99m scan and have enlarged on follow-up examination, should be evaluated with radioiodine scan to verify t h e functional status of the nodule. Some institutions routinely perform iodine-123 scans if a hot nodule is noted on a technetium-99m scan. Indeterminate nodules can be palpated on physical examination, but cannot be visualized On thyroid scan. These account for 6 percent of all solitary palpable nodules. Having similar clinical significance to cold nodules, i n d e t e r m i n a t e nodules should be treated as such [7,8]. Usually such nodules lie anteriorly to a functioning portion Of the thyroid gland and, therefore, are not seen on the scan due to superimposed normal activity [7]. Thallium-201 chloride is a relatively new agent for thyroid scanning. It is readily incorporated into well-perfused cellular lesions and is very sensitive to cancer in nodules that are cold on technetium-99m scan [7]. The disadvantage of thallium-201 is that tumors are indistinguishable from adenomas and thyroiditis [9,10]. Ochi et al [11] found that most malignant tumors retained thallium, whereas benign lesions cleared t h e t r a c e r faster; therefore, delayed thallium-201 scanning improved the specificity. Ikekubo et al [12] found that if technetium-99m and thallium-201 scans are combined, 90.8 percent of differentiated carcinomas will be detected. They concluded that the major advantages of the thallium-201 scan are detection of lymph node involvement or substernal extension; detection of residual, recurrent, or metastatic disease in postthyroidectomy patients whose initial tumor incorporated thallium-201 chloride; and detection of functioning nodules and suppressed normal thyroid with a functioning tumor. Although these studies support the continued use of thallium-201 scanning in large centers for academic purposes, they do not support its use for routine evaluation. Thyroid scintigraphy using technetium-99m (V) dimercaptosuccinic acid has been shown to be specific for medullary carcinoma with no sign of uptake in other forms of thyroid carcinoma [i3]. This new scintographic agent can help determine the extent of medullary thyroid carcinoma and locate its metastasis [13]. U l t r a s o n o g r a p h y : The primary purpose of thyroid echography is to predict the potential for malignancy of a thyroid mass by determining if it is solid, cystic, or multinodular. Ultrasonography can also be used after thyroxine suppressive therapy to evaluate a change in the size of a nodule that is difficult to palpate. There is no need for cessation of thyroid suppression if ultrasonography is used rather than The American Journal of Surgery
Diagnostic Imaging Techniques in Thyroid Cancer
radionuclide scanning. Other uses include guiding fine:needle aspiration, screening patients with prior head and neck irradiation, and detecting recurrent disease. Ultrasonography can be performed with grayscale B-mode scanning or high-resolution real-time ultrasound. Although gray-scale B-mode echography is available in most hospitals and can detect lesions as small as 5 mm, it is not reliable for detecting smaller nonpalpab!e lesions [6]. Abdel-Nabi et al [14] found a 32 percent false-negative rate, with several lesions being missed on gray-scale B-mode u!trasonography but being detected on radionuclide scan-directed operation. High-resolution real-time ultrasonography can detect occult nodules as small as 2 to 6 mm [15]. The incidence of solid cold nodules that are malignant ranges from 15 to 25 percent, which is significantly higher than purely cysti c lesions. Approximately 25 percent of Cystic lesions that have a solid component (mixed or complex) are malignant. Mixed lesions should undergo ultrasonographically-guided fine-needle aspiration from the solid component of the lesion. Cystic thyroid lesions identified by gray-scale Bmode ultrasonography have a relatively high rate of malignancy (7 percent), indicating that many nodules that appear to be purely cystic by this imaging technique actually are not. Indeed, a great majority of lesions that appear to be cystic with this instrumentation are found to contain small amounts of solid tissue on high-resolution real-time ultrasonography. Using high-resolution real-time ultrasonography, Simeone et al [i5] found only one purely cystic nodule in 550 patients evaluated for thyroid masses. On pathologic examination, the majority of these cystic tumors were found to be follicular adenomas with cystic degeneration. High-resolution real-time ultrasonography cap differentiate a purely cystic lesion with a reported incidence of 95 to 100 percent [4]. A ring of light seen surrounding a mass on thyroid ultrasonography is known as the halo sign. Previously associated with benign follicular adenomas, this sign has been found with well-differentiated carcinomas and, in reality, demonstrates a lack of specificity [15,16]. Despite improved sensitivity and higher resolu tion, there is no reliable ultrasonographic criteria on which to base the diagnosis of malignancy. Furthermore, fine-needle aspiration and physical examination can achieve most of the goals of ultrasonog r a p h y at a m u c h lower cost. T h e r e f o r e , ultrasonography must only be used in specific clinical situations to prevent unnecessary testing and misuse of the time and funds of the patient. When Ultrasonography is performed, we suggest high-resolution real-time ultrasonography as opposed to gray-scale B-m0de scanning. Conventional r a d i o g r a p h y : A recent chest roentgenogram should be available to rule out lung Volume 155, February 1988
metastasis and tracheal deviation. A soft-tissue lateral neck roentgen0gram can identify foci of calcification or tracheal compromise but should not be routinely obtained because such nonspecific findings will not alter treatment. If a patient complains of dysphagia or a foreign body sensation, a barium swallow may demonstrate esophageal compression or invasion. Angiography, formerly used in an attempt to differentiate benign from malignant lesions, is impractical due to its invasive nature, equivocal results, and expense [17]. Of conventional roentgenograms, only a recent chest roentgenogram need be routine. Computerized t o m o g r a p h y and magnetic resonance imaging: Computerized tomography and magnetic resonance imaging have similar roles in the evaluation of thyroid disease. Although extremely helpful when indicated, neither should be obtained routinely. Both computerized tomography and magnetic resonance imaging can demonstrate the normal anatomic characteristics of the thyroid and surrounding structures. On computerized tomography scan, thyroid tissue has increased density against surrounding structures because of its high iodine content [18]. Magnetic resonance imaging will image the thyroid giand with a signal intensity distinct from surrounding structures on both T1 and T2 weighted images. Distortion o f normal tissue planes or fatty compartments by tumor extension can be detected by both computerized tomography and magnetic resonance imaging. Computerized t omography and magnetic resonance imaging serve five primary purposes: to determine exact location and degree of invasiveness of large thyroid cancers; to evaluate substernal or retrotracheal extension of large thyroid tumors or goiters (Figures 4 and 5) [47]; to detect regional metastatic disease; to detect preoperatively cervical lymphadenopathy; and to detect local recurrences. Both computerized tomography and magnetic resonance imaging have been shown to correlate well with radionuclide images of the thyroid gland. Silverman et al [18] found regions of decreased density on computerized tomography scan to correspond with areas of decreased tracer activity on radionuclide scanning, especially in patients with multinodular goiter. Kulkarni et al [19] demonstrated areas of decreased uptake on technetium99m thyroid scans that corresponded to regions of increased signal intensity with magnetic resonance imaging on a T2 weighted image. Magnetic resonance imaging has revealed prolonged T1 and T2 relaxation times with solid thyroid nodules, frequently Correlating with findings on radionuclide scan [19]. Both computerized tomography and magnetic resonance have specific advantages for thyroid imaging. Computerized tomography scan is !ess expensive, more accessible, and more familiar to most radiologists; however, magnetic resonance imaging 219
Friedman et al
Figure 4. Thyroid carcinoma. Axial computerized tomography scan shows a round tumor ( t) Involving the right lobe of the thyroid. Note the deformity of the trachea (arrows) adjacent to the mass because of tumor Invasion.
can provide superior soft-tissue contrast and vascular imaging without contrast medium or radiation exposure. Furthermore, the contrast between muscle and tumor or lymphadenopathy provided by magnetic resonance imaging is superior to that of computerized tomography. Magnetic resonance imaging can identify lymphadenopathy, but does not differentiate between hyperplasia and metastatic disease. When the muscle is invaded by tumor, there is a change in signal intensity that further enhances the contrast between muscle and tumor [20]. Streak shoulder artifact is not a problem with magnetic resonance imaging as it is with computerized tomography thyroid imaging. Of particular importance in the examination of the neck and mediastinum is the ability of magnetic resonance imaging to provide coronal and sagittal sections to evaluate mediastinal extension [20]. For these reasons, magnetic resonance imaging is preferred in those cases where computerized tomography or magnetic resonance imaging is indicated. P o s t o p e r a t i v e evaluation: After operation, certain patients will require additional diagnostic tests to rule out residual disease or metastasis. In recent years, the indications for postoperative iodine-131 scans have been liberalized. In general, scans are not needed after the excision o f well-differentiated follicular carcinomas without capsular invasion or solitary papillary thyroid carcinomas less than 1.5 cm. A postoperative iodine-131 scan should be obtained in cases of invasive follicular carcinoma and papillary carcinomas greater than 1.5 cm in size.
220
Figure 5. Thyroid carcinoma. Top, axial view showing substernal extension, 2,000/20 msec, relatively T1 weighted magnetic resonance scan shows a large mass (arrow) arising from the left lobe of the thyroid, t = trachea. Bottom, axial view showing subslernal extension, 2,000 msec, relatively T1 weighted magnetic resonance scan obtained 2.3 cm below that of the top figure, shows the superior medlastinal mass with marked posferior extension (arrow) behind the trachea (T).
Postoperative preparation for iodine-131 total body scan is crucial to obtaining accurate data because high levels of thyroid-stimulating hormone are needed for good visualization of residual primary thyroid cancer or its metastasis [7]. Subtotal, or preferably, total thyroidectomy must be performed to ensure an accurate metastatic survey as appropriate elevations of thyroid-stimulating hor-
The American Journa| of Surgery
Diagnostic Imaging Techniques in Thyroid Cancer
mone will not be obtained if considerable thyroid tissue remains (enough to take up 5 percent of the tracer dose). During the 6 weeks before the scan, all thyroid medication is withheld, and I vJeek prior to the scan, the patient can be placed on a low-iodine diet in an effort to increase visualization [21]. The usual dose of radioactive iodine for a metastatic survey is 2 to 3 mcI of sodium iodine-131 taken orally. Total body imaging is performed at 24 to 72 hours [6,7]. The survey can be repeated whenever there is a suspicion of metastasis or recurrence. Radionuclide scanning can be performed with technetium-99m medronate to detect skeletal metastasis at an earlier stage than with conventional roentgenograms [22]. Systemic staging of thyroid lymphoma can be performed with gallium-67 [6]. These radionuclides are used much less frequently than iodine-131. Iodine-131 scans are not very helpful in determining the presence or extent of local recurrence. In any case where extensive local disease has been resected, postoperative computerized tomography or magnetic resonance imaging is performed as a baseline 4 weeks postoperatively after surgical edema has resolved. Repeat scans should then be performed whenever the question of local recurrence arises. If a tumor has been highly invasive, routine follow-up with computerized tomography or magnetic resonance imaging scanning can be carried out 1 and 5 years postoperatively, even if there is no clinical suspicion of recurrence. Two examples are presented to demonstrate the applications of thyroid imaging. Hyperthyroidism developed in a 56 year old white woman and was treated with radioactive iodine. Hypothyroidism then 'occurred. The patient was controlled with throxine for 2 years with no problems. Right vocal cord paralysis subsequently developed. Examination of the neck by three surgeons revealed no palpable thyroid tissue and no distinct mass. An iodine-131 scan showed essentially no uptake in either lobe, which was consistent with the history of radioactive ablation. A computerized tomography scan clearly showed a 2 by 2 cm mass behind the sternocleidomastoid muscle that was lateral but adjacent to the thyroid gland (Figure 6). At operation, a benign thyroid nodule was found under the sternocleidomastoid muscle, stretching and displacing a nonrecurrent laryngeal nerve [23]. The mass, demonstrated by computerized tomography, was undetectable by physical examination or iodine-131 scan. Another patient presented with hyperthyroidism 14 years after treatment of a nontoxic goiter by subtotal thyroidectomy at the age of 16. Physical examination revealed fullness in the lower neck but no discrete mass. The findings on technetium-99m scanning were consistent with partial thyroidecto-
Volume 155, February 1958
Figure 6. Thyroid adenoma. Postcontrast axial computerized tomography scan shows a mass (arrow) posterior and lateral to the right thyroid lobe. Notice displacement of the vascular structures adjacent to It. There is no evidence of Invasion of the tracheal or paratracheal regions. C -~ left common caroUd artery; E = esophagus;
Figure 7. Technetlum-99m scans showing absence of a portion of the left lobe consistent with previous operation.
my but were otherwise normal (Figure 7). Magnetic resonance imaging clearly showed extensive substernal tumor invading the trachea (Figure 8). At operation, it was confirmed to be a papillary carcinoma.
Summary With the refinement of fine-needle aspiration, the specific applications of thyroid imaging techniques need to be reevaluated for efficiency and cost containment. No thyroid imaging test should be
221
Friedman et al
Figure 8. Thyroid carcinoma. The left lobe is partially gone in all four views. Top left view, 800/25 msec, T~ weighted magnetic resonance scan shows a large tumor mass (NI) with marked compromise of the tracheal lumen (open arrow). Note the common carotid artery (curved arrow). Top right, axial view, 800/25 msec, 7"1 weighted magnetic resonance scan at a different level shows marked enlargement of the thyroid lobes ( straight arrows), deformed trachea (T) and right common carotid artery (curved arrow). Bottom left, coronal view, 2,900/20 msec, 7"2weighted magnetic resonance scan shows thyroid masses (solid arrows) with Invasion of the trachea ( open arrow). Bottom right, coronal view, 2,000/80 msec, T: weighted magnetic resonance scan shows the tumor (solid arrows) and tracheal involvement ( open arrow). Notice that the tumor is slightly hyperintense relative to muscle In T1and becomes very hyperintense in 7"2 weighted magnetic resonance images.
routinely obtained. Radionuclide scanning is most beneficial in evaluating the functional status of thyroid nodules when fine-needle aspiration is inadequate, the findings are benign, or when there is no discrete nodule that is palpated in an enlarged gland. When fine-needle aspiration is unavailable or unreliable, radionuclide scanning becomes a firstline diagnostic tool. Ultrasonography should be used primarily for identifying a solid component of a cystic nodule, determining the size of nodules on thyroxine suppression that are not easily palpable, or for performing guided fine-needle aspiration. Computerized tomography and magnetic resonance imaging both have a definite role in the evaluation
222
of thyroid tumors. Magnetic resonance imaging is superior to computerized tomography for the evaluation of metastatic, retrotracheal, or mediastinal involvement of large thyroid tumors or goiters. Careful selection of the diagnostic techniques will ensure more accurate diagnosis and reduce unnecessary patient costs in the treatment of thyroid cancer. Acknowledgment: We thank D.G. Pavel, MD, from the University of Illinois Hospital, Department of Nuclear Medicine, who reviewed the manuscript and Terri Strorigl, RN, who edited it.
The American Journal of Surgery
Diagnostic Imaging Techniques in Thyroid Cancer
References 1. Ashcraft MW, VanHerle A. Management of thyroid nodules. I. History and physical examination, blood tests, x-ray tests, and ultrasonography. Head Neck Surg 1961; 3: 216-30. 2. Frable WJ. The treatment of thyroid cancer. The role of fine needle aspiration. Arch Otolaryngol Head Neck Surg, 1986; 112: 1200-3. 3. Ashcraft MV, VanHerle AJ. Management of thyroid nodules. I1. Scanning techniques, thyroid suppressive therapy and fine needle aspiration. Head Neck Surg 1981; 3: 297-332. 4. VanHerle AJ, Rich P, Lijung BE, et al. The thyroid nodule. Ann Intern Med 1982; 96: 221-32. 5. Massin JP, Planchon Z, Perez R. The comparison of 99-m Tc pertechnetate and 131-1scanning of thyroid nodules. Clin Nucl Med 1977; 2: 324-33. 6. Noyek AM, Greyson ND, Steinhardt MI, et al. Thyroid tumor imaging. Arch Otolaryngol 1983; 109: 205-24. 7. Freitas JE, Gross MD, Ripley S, Shapiro B. Redionuclide diagnosis and therapy of thyroid cancer. Semin Nucl Med 1985; 15: 106-31. 8. Okerlund MD, Greyson ND, Steinhardt MI, et al. Thyroid tumor imaging. Arch Otolaryngol 1983; 109: 205-24. 9. Tanami N, Hisada K. Clinical experienceof tumor imaging with 201-TI chloride. Clin Nucl Med 1977; 2: 75-81. 10. Makimoto K, Ohmura M, Tamada A, et al. Combined scintiscans in the diagnosis of thyroid carcinomas. Arch Otolaryngol (Stockh) 1985; 419: 189-94. 11. Ochi H, Sawa H, Fukuda T, et al. Thallium-201 chloride thyroid scintigraphy to evaluate benign and/or malignant nodules. Cancer 1982; 50: 236-40. 12. Ikekubo K, Higa T, Hirasa M, et al. Evaluation of radionuclide thyroid echography in the diagnosis of thyroid nodules. Clin
Volume 155, February 1988
Nucl Med 1986; 11: 145-9. 13. Miyauchi A, Endo K, Ohta H, et al. 99m Tc(V)-Dimercaptosuccinic acid scintigraphy for medullary thyroid carcinoma. World J Surg 1986; 10: 640-5. 14. AbdeI-Nabi H, Falko JM, Olsen JO, Freimanis AK. Solitary cold thyroid nodule: cost-effectiveness of ultrasonography. South Med J 1984; 77:1146-8. 15. Simeone JF, Danials GH, Mueller PR, et al. High-resolution real-time sonography of the thyroid. Radiology 1982; 145: 431-5. 16. Propper RA, Skolnick ML, Weinstein BJ, Dekker A. The nonspecificity of the thyroid halo sign. JCU 1980; 8: 129-32. 17. Damascelli B, Casinelli N, Terno G, Dragoni G, Saccozzi R. Second thoughts on the value of selective thyroid angiography. Am J Roentgenol Radium Ther Nucl Med 1972; 114: 822-9. 18. Silverman PM, Newman GE, Korobkin M, Workman JB, Moore AV, Coleman RE. Computed tomography in evaluation of thyroid disease. AJR 1984; 141: 897-902. 19. Kulkarni MV, Sandier MP, Shaft MI, et al. Clinical magnetic resonance imaging with nuclear medicine correlation. J Nucl Med 1985; 26: 944-57. 20. Mafee MF, Rasouli F, Spigos DG, et al. Magnetic resonance imaging in the diagnosis of nonsquamous tumors of the head and neck. Otolaryngol Clin North Am 1986; 19: 523. 21. Maxon HR, Thomas SR, Boehringer A, et al. Low iodine diet in 1-131 ablation of thyroid remnants. Clin Nucl Med 1983; 8: 123-6. 22. Noyek AM. Bone scanning in otolaryngology. Laryngoscope 1979; 89(suppl): 1-87. 23. Friedman M, Toriumi DT, Grybauskas VT. Nonrecurrent laryngeal nerves and their clinical significance. Laryngoscope 1986; 96: 87-9.
223
Diagnostic Imaging Techniques in Thyroid Cancer Michael Friedman, MD, Dean M. Toriumi, MD, and Mahmood F. Mafee, MD, Chicago, Illinois
The main objective in imaging the thyroid is to identify lesions likely to be malignant. Once the diagnosis of malignant disease is confirmed by histologic examination, imaging is useful to delineate the extent of tumor or metastatic disease. In the past, a routine workup typically began with thyroid function studies and a radionuclide scan. Recently, technologic advances and the need for strict cost containment have resulted in an alteration in the common algorithm used for the diagnostic evaluation of a thyroid nodule. In addition to the particular tests and sequence of the workup, the rationales behind them have changed as well. Aside from the need of the physician to obtain information for the treatment decision, the need to provide adequate information to the patient so he or she can participate in that decision is also important. Although recommendations can be made for specific testing and for consideration of operation, there are many situations where the indications for operation or the extent of the resection can be controversial. The final decision on treatment is based on input from three areas: clinical, laboratory findings, and personal factors (Figure 1). The history and physical examination, which includes such important aspects as prior irradiation, family history, and growth rate, as well as consistency, shape, and size of the mass all play a role. As this report discusses, laboratory testing includes a variety of imagFrom the Departments of Otolaryngology-Head and Neck Surgery and Radiology, University of Illinois College of Medicine; the Department of Otolaryngology-Head and Neck Surgery, Illinois Masonic Medical Center; and the Department of Otolaryngology Head and Neck Surgery, Northwestern University Medical School, Chicago, IlUnois. Requests for reprints should be addressed to Michael Friedman, MD, Department of Otolaryngology-Head and Neck Surgery, University of Illinois Eye and Ear Infirmary, 1855 West Taylor Street, Chicago, Illinois 60612. Editor's Note: This specifically solicited material contains an element of case report format that was found essential to convey the education messages of Figures 6, 7, and 8. It does not represent a willingness to review or accept case reports for the Journal.
Volume 155, February 1988
ing techniques that will add information. The physician's attitude and philosophy will play a major role. Finally, the patient's input in terms of his or her anxiety and philosophy must be added to the overall picture when reaching a decision in treatment. Although fine-needle aspiration is not a method of thyroid imaging, it cannot be excluded from this report. Besides being the first step in evaluating nodular thyroid diseases, fine-needle aspiration can be the determining factor in selecting appropriate tests, and it can be used in combination with thyroid imaging. Herein, we present an updated protocol for the use of diagnostic imaging to evaluate a discrete lesion, a diffusely enlarged gland, and regional, recurrent, and distant metastatic disease. We also discuss in depth the current diagnostic applications of plain films, contrast studies, radionuclide scanning, and ultrasonography, as well as computerized tomography and magnetic resonance imaging. The initial evaluation of a thyroid mass obviously begins with a thorough history and physical examination. The head and neck examination should include indirect laryngoscopy to assess vocal cord mobility, which, if impaired, may indicate recurrent laryngeal nerve invasion. Blood tests to evaluate thyroid function can be limited to a thyroxine level if the patient is euthyroid by history and physical findings. On the other hand, if an abnormality in thyroid function is suggested by the history, triiodothyronine resin uptake, free thyroxine, and thyroidstimulating hormone measurements should be obtained. If operation is planned, calcium levels should be obtained preoperatively. Approach to Thyroid Imaging
Solitary lesion: Thyroid neoplasia can present as either a discrete nodule or nodules or a diffusely enlarged gland, although the former is more likely to 215
Friedman et al
SURGERY OF THYROID MASSES .~LABORATORY FNDNG5
CLINICAL
J. PERSONAL TLFACTORS
~FNA
*History
I
=
reported false-negative rate of 0.3 to 10 percent and a false-positive rate of 0 to 2.5 percent [2]. Our experience, as well as that of others, is closer to the 10 percent false-negative rate and less than the 2.5 percent false-positive rate. For practical application, a malignant finding on fine-needle aspiration is a strong indication for operation. In contrast, a negative finding on fine-needle aspiration cannot definitely rule out malignancy. In the case of a benign or inadequate fine-needle aspiration or when fine-needle aspiration is unavailable, other tests and clinical evaluation become more important in planning treatment. When fine-needle aspiration reveals follicular or H~irthle cells, there is some controversy as to whether a distinction between a benign and malignant neoplasm can be made. In most cases, a tissue specimen is needed to document capsular or vascular invasion to confirm malignancy. If fine-needle aspiration indicates that a mass is solid but benign, the next step is to obtain a technetium-99m pertechnetate thyroid scan. This scan can provide an important source of information for the patient regarding the likelihood of malignancy and the frequency of follow-up. Most solid benign lesions are cold on radionuclide scan, with less than 10 percent being
~
H~tDAttitude
,Phys~alExarr~rmtion ,ThyroidScan
~Patient's Anxiety
*GrowthPatternof Mass ,Ultrasound ,Patient's Inputin Decision ,Age & GeneralHealth )OtherLab Studies
Figure 1. Multiple factors Influencing the decision to treat a
thyroid mass surgically.
be malignant. Thyroid nodules are relatively common and can be found in 4 to 7 percent of the population [1-4]. The incidence of solitary thyroid nodules that prove to be malignant on surgical excision ranges from 8 to 33 percent (median 20 percent). Fine-needle aspiration is the initial test used to evaluate a discrete thyroid nodule (Figure 2). Fineneedle aspiration can help classify a thyroid nodule as benign or malignant, as well as differentiate a solid mass from a cystic one. The incidence of inadequate aspiration depends largely on the experience of the person performing fine-needle aspiration, as well as the pathologist. Fine-needle aspiration has a
E V A L U A T I O N OF DISCRETE THYROID NODULE(S)
Fine N e e d l e
/
so,,f \
Malignant
Sur
ry %%
Hot
Follow
A s p i r a t i o n (FNA)
Benign
Cystic
Tc-99m Scan
A s p i r a t e Fluid
Cold
Follow with S u p p r e s s i o n
Ultrasound"
No S m a l l e r
Consider Surgery
216
Follow
Smaller
Follow with S u p p r e s s i o n
Reforms
Ultrasound
J
Purely C y s t i c
Re-aspirate
\
C y s t i c with Solid Component U l t r a s o u n d - g u i d e d FNA of Solid Component
Reforms
Benign
Malignant
Consider Surgery
T r e a t as Benign Solid Lesion
Surgery
F i g u r e 2. Protocol for workup and management of a solitary thyroid nodule. Single asterisk Indicates one prominent nodule in a mul#nodular gland; double asterisk Indicates ultrasonography only If the mass Is difficult to palpate; dagger Indicates computerized tomography or magnetic resonance imaging to assess substernal extension, Invasion, or both. FNA = fine-needle aspiration; T c - g g m = technetlum-ggm.
The American Journal of Surgery
Diagnostic Imaging Techniques in Thyroid Cancer
EVALUATION OF DIFFUSE THYROID ENLARGEMENT
Tc-99m Scan
\
Hot Nodule
Cold Nodule :~
\
FNA (Ultrasound-guided) ~'~
Benign
Follow
FNA (Ultrasound-guided) ~>~
Malignant
Surgery ~
Figure 3. Protocol for the evaluation and
management of a diffusely enlarged thyroid gland. Single asterisk Indicates ultrasonography only ff the mass is difficult to palpate; double asterisk Indicates computerized tomography or magnetic resonance Imaging for large masses to assess substernal extension, Invasion, or both.
hot. The patient should be informed that malignancy occurs in an estimated 10 percent of similar situations, as in benign solid lesion on fine-needle aspiration, but cold on scan. We prefer to treat a patient with a cold lesion that is benign by fine-needle aspiration with a 3 month course of thyroxine (levothyroxine sodium) suppression, with monthly followup to monitor the size of the nodule. Most lesions will not regress, but those that do are almost always benign. Ultrasonography is performed before and after the course of suppression to document size if the mass is not easily palpable. If the mass remains the same size or grows larger, operation is indicated. If a mass is easily palpable and fine needle aspiration has ruled out a cystic lesion, ultrasonography is superfluous. A mass that is benign on fine-needle aspiration and hot on technetium-99m scan may be an autonomous hot nodule and should be followed every 3 to 6 months. The patient can be informed that the likelihood of a malignancy is minimal. We believe that surgical management is usually unnecessary in such cases. On occasion, lesions are seen that are hot on technetium-99m scan but cold on radioiodine scan. These are termed discordant lesions and should be treated as cold nodules. The incidence of malignancy in discordant nodules is 15 percent, and therefore, surgical management should be considered [5]. A nodule that is cystic on fine-needle aspiration
Volume 155, February 1988
Benign
Suppress for 3 Months
( r e p e a t ultrasound) " ~
Smaller
Follow
No Change or Larger
Consider Surgery'f
should be aspirated and then reevaluated monthly for 3 months. If a cyst reforms, ultrasonography can be performed followed by ultrasonographically guided fine-needle aspiration if a solid component is detected. Cystic thyroid nodules are rarely purely cystic and ultrasonography can locate any solid component. Ultrasonographically-guided fine-needle aspiration can improve the yield and accuracy of the technique. If the lesion is indeed purely cystic, repeat aspiration may be attempted. D i f f u s e l y enlarged thyroid gland: Diffuse thyroid enlargement usually represents a benign goiter but may actually hide malignant disease. The initial diagnostic test in the workup of a diffusely enlarged gland is a technetium-99m thyroid scan (Figure 3). If a cold area is isolated but is not palpable on examination, an ultrasonographically-guided fineneedle aspiration can be performed on any irregular region that corresponds with the cold region on the scan. If the fine-needle aspiration findings are benign, the patient is placed on thyroxine suppression for 3 months to shrink the gland and make the cold, nonfunctioning area easier to evaluate. The patient should be examined every month. If after 3 months, the mass has become larger, operation is considered. If the lesion becomes smaller, the patient is followed every 3 months thereafter. Diffusely enlarged glands with a prominent nodule found to be hot on technetium-99m scan are examined every 3 months.
217
Friedman et al
Current Applications of Diagnostic Imaging R a d i o n u c l i d e s c a n s : Thyroid scans remain a first-line diagnostic tool if accurate fine-needle aspiration is unavailable. Scans are also performed to evaluate the functional status of nodules that are benign on fine-needle aspiration. One of the benefits of scanning after fine-needle aspiration is ensuring against false-negative results by further assessing the characteristics of the mass, and the clinician is also protecting himself legally. Also, the results of the scan will help determine the frequency of followup visits for patients on a 3 month course of suppression. Patients with cold nodules are seen approximately every month and patients with hot nodules, every 3 months. If a scan reveals that the nodule is cold and the physician informs the patient that he recommends suppression in spite of a 10 percent chance of malignancy, he has fully informed his patient. Three major types of radionuclide scans are currently available to image the thyroid. Technetium99m is the preferred agent for screening because it has the greatest sensitivity, low radiation exposure, and a short test time of 1 hour compared with 20 to 24 hours for radioiodine. In addition, it is readily available and the cost is relatively low. Technetium99m, which is a d m i n i s t e r e d intravenously, is trapped by the thyroid gland but not organified. Maximal thyroidal uptake is seen in 20 to 40 minutes. The technetium-99m scan can readily identify a nonfunctioning nodule, which does not trap any pertechnetate. A hot or warm nodule seen on early scan (3 minute s ) that becomes cool on late imaging may represent a malignant vascular lesion, which must be ruled out [6]. A hot nodule on late imaging (after 20 minutes) may be a functional nodule but may also represent trapping. A true functional nodule can only be confirmed by radioiodine scan. Radioactive iodine (iodine-131, iodine-125, and iodine-123) has a distinct advantage over all other tracers. Radioiodine is the only agent that is both trapped and organified by the thyroid gland, thus permitting evaluation of the true physiologic features of the thyroid gland and any masses within it. The quality of the image is better than that of technetium-99m. Radioiodine is taken by mouth and requires nearly 24 hours for organification. Iodine123 is the most favorable of the more than 20 radioactive isotopes of iodine. It has a short half-life (13.2 hours), a photon energy that permits good imaging (159 keV), and no particulate emissions [7]. A large r tracer dose can be administered at an acceptable radiation level. Iodine-125 and -131 require much higher radiation levels to yield a comparable image. However, iodine-123 is very expensive and has a short shelf life, rendering it practical only in large institutions where several scans are performed on a daily basis. A nodule that is hot on technetium-99m scan may 218
appear cold on radioiodine scan. Such discordant lesions theoretically trap iodine but do not organify it [7]. The majority of discordant lesions represent benign colloid nodules, follicular adenomas, or thyroiditis, but are nonetheless considered suspicious for Carcinoma, just as other cold nodules. Therefore, nodules that are hot on technetium-99m scan and have enlarged on follow-up examination, should be evaluated with radioiodine scan to verify t h e functional status of the nodule. Some institutions routinely perform iodine-123 scans if a hot nodule is noted on a technetium-99m scan. Indeterminate nodules can be palpated on physical examination, but cannot be visualized On thyroid scan. These account for 6 percent of all solitary palpable nodules. Having similar clinical significance to cold nodules, i n d e t e r m i n a t e nodules should be treated as such [7,8]. Usually such nodules lie anteriorly to a functioning portion Of the thyroid gland and, therefore, are not seen on the scan due to superimposed normal activity [7]. Thallium-201 chloride is a relatively new agent for thyroid scanning. It is readily incorporated into well-perfused cellular lesions and is very sensitive to cancer in nodules that are cold on technetium-99m scan [7]. The disadvantage of thallium-201 is that tumors are indistinguishable from adenomas and thyroiditis [9,10]. Ochi et al [11] found that most malignant tumors retained thallium, whereas benign lesions cleared t h e t r a c e r faster; therefore, delayed thallium-201 scanning improved the specificity. Ikekubo et al [12] found that if technetium-99m and thallium-201 scans are combined, 90.8 percent of differentiated carcinomas will be detected. They concluded that the major advantages of the thallium-201 scan are detection of lymph node involvement or substernal extension; detection of residual, recurrent, or metastatic disease in postthyroidectomy patients whose initial tumor incorporated thallium-201 chloride; and detection of functioning nodules and suppressed normal thyroid with a functioning tumor. Although these studies support the continued use of thallium-201 scanning in large centers for academic purposes, they do not support its use for routine evaluation. Thyroid scintigraphy using technetium-99m (V) dimercaptosuccinic acid has been shown to be specific for medullary carcinoma with no sign of uptake in other forms of thyroid carcinoma [i3]. This new scintographic agent can help determine the extent of medullary thyroid carcinoma and locate its metastasis [13]. U l t r a s o n o g r a p h y : The primary purpose of thyroid echography is to predict the potential for malignancy of a thyroid mass by determining if it is solid, cystic, or multinodular. Ultrasonography can also be used after thyroxine suppressive therapy to evaluate a change in the size of a nodule that is difficult to palpate. There is no need for cessation of thyroid suppression if ultrasonography is used rather than The American Journal of Surgery
Diagnostic Imaging Techniques in Thyroid Cancer
radionuclide scanning. Other uses include guiding fine:needle aspiration, screening patients with prior head and neck irradiation, and detecting recurrent disease. Ultrasonography can be performed with grayscale B-mode scanning or high-resolution real-time ultrasound. Although gray-scale B-mode echography is available in most hospitals and can detect lesions as small as 5 mm, it is not reliable for detecting smaller nonpalpab!e lesions [6]. Abdel-Nabi et al [14] found a 32 percent false-negative rate, with several lesions being missed on gray-scale B-mode u!trasonography but being detected on radionuclide scan-directed operation. High-resolution real-time ultrasonography can detect occult nodules as small as 2 to 6 mm [15]. The incidence of solid cold nodules that are malignant ranges from 15 to 25 percent, which is significantly higher than purely cysti c lesions. Approximately 25 percent of Cystic lesions that have a solid component (mixed or complex) are malignant. Mixed lesions should undergo ultrasonographically-guided fine-needle aspiration from the solid component of the lesion. Cystic thyroid lesions identified by gray-scale Bmode ultrasonography have a relatively high rate of malignancy (7 percent), indicating that many nodules that appear to be purely cystic by this imaging technique actually are not. Indeed, a great majority of lesions that appear to be cystic with this instrumentation are found to contain small amounts of solid tissue on high-resolution real-time ultrasonography. Using high-resolution real-time ultrasonography, Simeone et al [i5] found only one purely cystic nodule in 550 patients evaluated for thyroid masses. On pathologic examination, the majority of these cystic tumors were found to be follicular adenomas with cystic degeneration. High-resolution real-time ultrasonography cap differentiate a purely cystic lesion with a reported incidence of 95 to 100 percent [4]. A ring of light seen surrounding a mass on thyroid ultrasonography is known as the halo sign. Previously associated with benign follicular adenomas, this sign has been found with well-differentiated carcinomas and, in reality, demonstrates a lack of specificity [15,16]. Despite improved sensitivity and higher resolu tion, there is no reliable ultrasonographic criteria on which to base the diagnosis of malignancy. Furthermore, fine-needle aspiration and physical examination can achieve most of the goals of ultrasonog r a p h y at a m u c h lower cost. T h e r e f o r e , ultrasonography must only be used in specific clinical situations to prevent unnecessary testing and misuse of the time and funds of the patient. When Ultrasonography is performed, we suggest high-resolution real-time ultrasonography as opposed to gray-scale B-m0de scanning. Conventional r a d i o g r a p h y : A recent chest roentgenogram should be available to rule out lung Volume 155, February 1988
metastasis and tracheal deviation. A soft-tissue lateral neck roentgen0gram can identify foci of calcification or tracheal compromise but should not be routinely obtained because such nonspecific findings will not alter treatment. If a patient complains of dysphagia or a foreign body sensation, a barium swallow may demonstrate esophageal compression or invasion. Angiography, formerly used in an attempt to differentiate benign from malignant lesions, is impractical due to its invasive nature, equivocal results, and expense [17]. Of conventional roentgenograms, only a recent chest roentgenogram need be routine. Computerized t o m o g r a p h y and magnetic resonance imaging: Computerized tomography and magnetic resonance imaging have similar roles in the evaluation of thyroid disease. Although extremely helpful when indicated, neither should be obtained routinely. Both computerized tomography and magnetic resonance imaging can demonstrate the normal anatomic characteristics of the thyroid and surrounding structures. On computerized tomography scan, thyroid tissue has increased density against surrounding structures because of its high iodine content [18]. Magnetic resonance imaging will image the thyroid giand with a signal intensity distinct from surrounding structures on both T1 and T2 weighted images. Distortion o f normal tissue planes or fatty compartments by tumor extension can be detected by both computerized tomography and magnetic resonance imaging. Computerized t omography and magnetic resonance imaging serve five primary purposes: to determine exact location and degree of invasiveness of large thyroid cancers; to evaluate substernal or retrotracheal extension of large thyroid tumors or goiters (Figures 4 and 5) [47]; to detect regional metastatic disease; to detect preoperatively cervical lymphadenopathy; and to detect local recurrences. Both computerized tomography and magnetic resonance imaging have been shown to correlate well with radionuclide images of the thyroid gland. Silverman et al [18] found regions of decreased density on computerized tomography scan to correspond with areas of decreased tracer activity on radionuclide scanning, especially in patients with multinodular goiter. Kulkarni et al [19] demonstrated areas of decreased uptake on technetium99m thyroid scans that corresponded to regions of increased signal intensity with magnetic resonance imaging on a T2 weighted image. Magnetic resonance imaging has revealed prolonged T1 and T2 relaxation times with solid thyroid nodules, frequently Correlating with findings on radionuclide scan [19]. Both computerized tomography and magnetic resonance have specific advantages for thyroid imaging. Computerized tomography scan is !ess expensive, more accessible, and more familiar to most radiologists; however, magnetic resonance imaging 219
Friedman et al
Figure 4. Thyroid carcinoma. Axial computerized tomography scan shows a round tumor ( t) Involving the right lobe of the thyroid. Note the deformity of the trachea (arrows) adjacent to the mass because of tumor Invasion.
can provide superior soft-tissue contrast and vascular imaging without contrast medium or radiation exposure. Furthermore, the contrast between muscle and tumor or lymphadenopathy provided by magnetic resonance imaging is superior to that of computerized tomography. Magnetic resonance imaging can identify lymphadenopathy, but does not differentiate between hyperplasia and metastatic disease. When the muscle is invaded by tumor, there is a change in signal intensity that further enhances the contrast between muscle and tumor [20]. Streak shoulder artifact is not a problem with magnetic resonance imaging as it is with computerized tomography thyroid imaging. Of particular importance in the examination of the neck and mediastinum is the ability of magnetic resonance imaging to provide coronal and sagittal sections to evaluate mediastinal extension [20]. For these reasons, magnetic resonance imaging is preferred in those cases where computerized tomography or magnetic resonance imaging is indicated. P o s t o p e r a t i v e evaluation: After operation, certain patients will require additional diagnostic tests to rule out residual disease or metastasis. In recent years, the indications for postoperative iodine-131 scans have been liberalized. In general, scans are not needed after the excision o f well-differentiated follicular carcinomas without capsular invasion or solitary papillary thyroid carcinomas less than 1.5 cm. A postoperative iodine-131 scan should be obtained in cases of invasive follicular carcinoma and papillary carcinomas greater than 1.5 cm in size.
220
Figure 5. Thyroid carcinoma. Top, axial view showing substernal extension, 2,000/20 msec, relatively T1 weighted magnetic resonance scan shows a large mass (arrow) arising from the left lobe of the thyroid, t = trachea. Bottom, axial view showing subslernal extension, 2,000 msec, relatively T1 weighted magnetic resonance scan obtained 2.3 cm below that of the top figure, shows the superior medlastinal mass with marked posferior extension (arrow) behind the trachea (T).
Postoperative preparation for iodine-131 total body scan is crucial to obtaining accurate data because high levels of thyroid-stimulating hormone are needed for good visualization of residual primary thyroid cancer or its metastasis [7]. Subtotal, or preferably, total thyroidectomy must be performed to ensure an accurate metastatic survey as appropriate elevations of thyroid-stimulating hor-
The American Journa| of Surgery
Diagnostic Imaging Techniques in Thyroid Cancer
mone will not be obtained if considerable thyroid tissue remains (enough to take up 5 percent of the tracer dose). During the 6 weeks before the scan, all thyroid medication is withheld, and I vJeek prior to the scan, the patient can be placed on a low-iodine diet in an effort to increase visualization [21]. The usual dose of radioactive iodine for a metastatic survey is 2 to 3 mcI of sodium iodine-131 taken orally. Total body imaging is performed at 24 to 72 hours [6,7]. The survey can be repeated whenever there is a suspicion of metastasis or recurrence. Radionuclide scanning can be performed with technetium-99m medronate to detect skeletal metastasis at an earlier stage than with conventional roentgenograms [22]. Systemic staging of thyroid lymphoma can be performed with gallium-67 [6]. These radionuclides are used much less frequently than iodine-131. Iodine-131 scans are not very helpful in determining the presence or extent of local recurrence. In any case where extensive local disease has been resected, postoperative computerized tomography or magnetic resonance imaging is performed as a baseline 4 weeks postoperatively after surgical edema has resolved. Repeat scans should then be performed whenever the question of local recurrence arises. If a tumor has been highly invasive, routine follow-up with computerized tomography or magnetic resonance imaging scanning can be carried out 1 and 5 years postoperatively, even if there is no clinical suspicion of recurrence. Two examples are presented to demonstrate the applications of thyroid imaging. Hyperthyroidism developed in a 56 year old white woman and was treated with radioactive iodine. Hypothyroidism then 'occurred. The patient was controlled with throxine for 2 years with no problems. Right vocal cord paralysis subsequently developed. Examination of the neck by three surgeons revealed no palpable thyroid tissue and no distinct mass. An iodine-131 scan showed essentially no uptake in either lobe, which was consistent with the history of radioactive ablation. A computerized tomography scan clearly showed a 2 by 2 cm mass behind the sternocleidomastoid muscle that was lateral but adjacent to the thyroid gland (Figure 6). At operation, a benign thyroid nodule was found under the sternocleidomastoid muscle, stretching and displacing a nonrecurrent laryngeal nerve [23]. The mass, demonstrated by computerized tomography, was undetectable by physical examination or iodine-131 scan. Another patient presented with hyperthyroidism 14 years after treatment of a nontoxic goiter by subtotal thyroidectomy at the age of 16. Physical examination revealed fullness in the lower neck but no discrete mass. The findings on technetium-99m scanning were consistent with partial thyroidecto-
Volume 155, February 1958
Figure 6. Thyroid adenoma. Postcontrast axial computerized tomography scan shows a mass (arrow) posterior and lateral to the right thyroid lobe. Notice displacement of the vascular structures adjacent to It. There is no evidence of Invasion of the tracheal or paratracheal regions. C -~ left common caroUd artery; E = esophagus;
Figure 7. Technetlum-99m scans showing absence of a portion of the left lobe consistent with previous operation.
my but were otherwise normal (Figure 7). Magnetic resonance imaging clearly showed extensive substernal tumor invading the trachea (Figure 8). At operation, it was confirmed to be a papillary carcinoma.
Summary With the refinement of fine-needle aspiration, the specific applications of thyroid imaging techniques need to be reevaluated for efficiency and cost containment. No thyroid imaging test should be
221
Friedman et al
Figure 8. Thyroid carcinoma. The left lobe is partially gone in all four views. Top left view, 800/25 msec, T~ weighted magnetic resonance scan shows a large tumor mass (NI) with marked compromise of the tracheal lumen (open arrow). Note the common carotid artery (curved arrow). Top right, axial view, 800/25 msec, 7"1 weighted magnetic resonance scan at a different level shows marked enlargement of the thyroid lobes ( straight arrows), deformed trachea (T) and right common carotid artery (curved arrow). Bottom left, coronal view, 2,900/20 msec, 7"2weighted magnetic resonance scan shows thyroid masses (solid arrows) with Invasion of the trachea ( open arrow). Bottom right, coronal view, 2,000/80 msec, T: weighted magnetic resonance scan shows the tumor (solid arrows) and tracheal involvement ( open arrow). Notice that the tumor is slightly hyperintense relative to muscle In T1and becomes very hyperintense in 7"2 weighted magnetic resonance images.
routinely obtained. Radionuclide scanning is most beneficial in evaluating the functional status of thyroid nodules when fine-needle aspiration is inadequate, the findings are benign, or when there is no discrete nodule that is palpated in an enlarged gland. When fine-needle aspiration is unavailable or unreliable, radionuclide scanning becomes a firstline diagnostic tool. Ultrasonography should be used primarily for identifying a solid component of a cystic nodule, determining the size of nodules on thyroxine suppression that are not easily palpable, or for performing guided fine-needle aspiration. Computerized tomography and magnetic resonance imaging both have a definite role in the evaluation
222
of thyroid tumors. Magnetic resonance imaging is superior to computerized tomography for the evaluation of metastatic, retrotracheal, or mediastinal involvement of large thyroid tumors or goiters. Careful selection of the diagnostic techniques will ensure more accurate diagnosis and reduce unnecessary patient costs in the treatment of thyroid cancer. Acknowledgment: We thank D.G. Pavel, MD, from the University of Illinois Hospital, Department of Nuclear Medicine, who reviewed the manuscript and Terri Strorigl, RN, who edited it.
The American Journal of Surgery
Diagnostic Imaging Techniques in Thyroid Cancer
References 1. Ashcraft MW, VanHerle A. Management of thyroid nodules. I. History and physical examination, blood tests, x-ray tests, and ultrasonography. Head Neck Surg 1961; 3: 216-30. 2. Frable WJ. The treatment of thyroid cancer. The role of fine needle aspiration. Arch Otolaryngol Head Neck Surg, 1986; 112: 1200-3. 3. Ashcraft MV, VanHerle AJ. Management of thyroid nodules. I1. Scanning techniques, thyroid suppressive therapy and fine needle aspiration. Head Neck Surg 1981; 3: 297-332. 4. VanHerle AJ, Rich P, Lijung BE, et al. The thyroid nodule. Ann Intern Med 1982; 96: 221-32. 5. Massin JP, Planchon Z, Perez R. The comparison of 99-m Tc pertechnetate and 131-1scanning of thyroid nodules. Clin Nucl Med 1977; 2: 324-33. 6. Noyek AM, Greyson ND, Steinhardt MI, et al. Thyroid tumor imaging. Arch Otolaryngol 1983; 109: 205-24. 7. Freitas JE, Gross MD, Ripley S, Shapiro B. Redionuclide diagnosis and therapy of thyroid cancer. Semin Nucl Med 1985; 15: 106-31. 8. Okerlund MD, Greyson ND, Steinhardt MI, et al. Thyroid tumor imaging. Arch Otolaryngol 1983; 109: 205-24. 9. Tanami N, Hisada K. Clinical experienceof tumor imaging with 201-TI chloride. Clin Nucl Med 1977; 2: 75-81. 10. Makimoto K, Ohmura M, Tamada A, et al. Combined scintiscans in the diagnosis of thyroid carcinomas. Arch Otolaryngol (Stockh) 1985; 419: 189-94. 11. Ochi H, Sawa H, Fukuda T, et al. Thallium-201 chloride thyroid scintigraphy to evaluate benign and/or malignant nodules. Cancer 1982; 50: 236-40. 12. Ikekubo K, Higa T, Hirasa M, et al. Evaluation of radionuclide thyroid echography in the diagnosis of thyroid nodules. Clin
Volume 155, February 1988
Nucl Med 1986; 11: 145-9. 13. Miyauchi A, Endo K, Ohta H, et al. 99m Tc(V)-Dimercaptosuccinic acid scintigraphy for medullary thyroid carcinoma. World J Surg 1986; 10: 640-5. 14. AbdeI-Nabi H, Falko JM, Olsen JO, Freimanis AK. Solitary cold thyroid nodule: cost-effectiveness of ultrasonography. South Med J 1984; 77:1146-8. 15. Simeone JF, Danials GH, Mueller PR, et al. High-resolution real-time sonography of the thyroid. Radiology 1982; 145: 431-5. 16. Propper RA, Skolnick ML, Weinstein BJ, Dekker A. The nonspecificity of the thyroid halo sign. JCU 1980; 8: 129-32. 17. Damascelli B, Casinelli N, Terno G, Dragoni G, Saccozzi R. Second thoughts on the value of selective thyroid angiography. Am J Roentgenol Radium Ther Nucl Med 1972; 114: 822-9. 18. Silverman PM, Newman GE, Korobkin M, Workman JB, Moore AV, Coleman RE. Computed tomography in evaluation of thyroid disease. AJR 1984; 141: 897-902. 19. Kulkarni MV, Sandier MP, Shaft MI, et al. Clinical magnetic resonance imaging with nuclear medicine correlation. J Nucl Med 1985; 26: 944-57. 20. Mafee MF, Rasouli F, Spigos DG, et al. Magnetic resonance imaging in the diagnosis of nonsquamous tumors of the head and neck. Otolaryngol Clin North Am 1986; 19: 523. 21. Maxon HR, Thomas SR, Boehringer A, et al. Low iodine diet in 1-131 ablation of thyroid remnants. Clin Nucl Med 1983; 8: 123-6. 22. Noyek AM. Bone scanning in otolaryngology. Laryngoscope 1979; 89(suppl): 1-87. 23. Friedman M, Toriumi DT, Grybauskas VT. Nonrecurrent laryngeal nerves and their clinical significance. Laryngoscope 1986; 96: 87-9.
223