Common and Uncommon Sonographic Features of Papillary Thyroid Carcinoma

Case Series Common and Uncommon Sonographic Features of Papillary Thyroid Carcinoma Bryan K. Chan, MD, Terry S. Desser, MD, I. Ross McDougall, MD, Ronald J. Weigel, MD, R. Brooke Jeffrey, Jr, MD Objective. To determine the relative frequency of various sonographic findings in papillary carcinoma of the thyroid. Methods. We retrospectively analyzed the sonographic features in 55 patients with proven papillary carcinoma of the thyroid. Sonographic features analyzed were echo texture, cystic change, margin, contour, presence of a peripheral halo, vascularity, and calcification pattern. Features were classified as common ( 35% of cases) or uncommon (<10% of cases). Combinations of features were also analyzed. Results. Common sonographic features of papillary carcinoma included hypoechoic texture (86%), microcalcifications (42%) or no calcifications (47%), well-defined margins (47%), and intrinsic hypervascularity (69%). Uncommon features included hyperechoic or mixed echo texture, cystic elements, irregular margins, hypovascularity, and coarse or peripheral calcifications. Of the 29 lesions that had calcifications, 20 (69%) had microcalcifications; 5 (17%) had coarse calcifications; and 1 had peripheral calcifications. In total, 54% of cases had at least 1 uncommon feature, and 11% had 2 or more uncommon features. Cystic carcinomas were rare and accounted for only 6% of lesions; all had hypervascular solid components. No carcinomas in our series were completely avascular. Conclusions. There is a broad spectrum of sonographic findings in papillary carcinoma of the thyroid. Half of the lesions in this series had at least 1 uncommon sonographic feature. Key words: head and neck neoplasms, sonography; sonography, Doppler studies; sonography, power Doppler studies; thyroid, biopsy; thyroid, neoplasms; thyroid, sonography. Abbreviations FNA, fine-needle aspiration Received April 29, 2003, from the Departments of Radiology (B.K.C., T.S.D., R.B.J.) and Nuclear Medicine (I.R.M.), Stanford University School of Medicine, Stanford, California USA; and Department of Surgery, Thomas Jefferson University, Philadelphia, Pennsylvania USA (R.J.W.). Revision requested June 10, 2003. Revised manuscript accepted for publication July 2, 2003. Address correspondence and reprint requests to Bryan K. Chan, MD, Department of Radiology, H-1307, Stanford University School of Medicine, 300 Pasteur Dr, Stanford, CA 94305. E-mail: xone@stanford.edu. Thyroid nodules are common, but thyroid cancer is rare. Palpable nodules (usually >1.5 cm) are found in approximately 5% of the population. The prevalence of nonpalpable nodules is even higher, occurring in an estimated 40% to 50% of the population. In contrast, the American Cancer Society estimated that there were only 19,500 new cases of thyroid cancer in 2001, representing 1.5% of all new cancers. 1 During the past decade, the use of high-resolution sonography has resulted in high rates of detection of thyroid nodules, but characterization of nodules as either benign or malignant remains problematic because of considerable overlap in sonographic features. 2 2003 by the American Institute of Ultrasound in Medicine J Ultrasound Med 22:1083 1090, 2003 0278-4297/03/$3.50

Sonographic Features of Papillary Thyroid Carcinoma Given this limitation, treatment of patients with palpable thyroid nodules or with large (>1 cm) nodules is usually fine-needle aspiration (FNA) biopsy with sonographic guidance if the nodule is not palpable. 1 Consequently, a large number of patients are referred for sonographically guided FNA of thyroid nodules. Papillary carcinoma accounts for most thyroid malignancies. 3,4 Although there have been multiple studies describing the sonographic features of papillary thyroid carcinoma, few authors have examined all of the reported features of this entity in a single study. In this report, we present the common and uncommon high-resolution sonographic features and feature combinations of the proven papillary thyroid carcinomas encountered in our practice. Materials and Methods We retrospectively reviewed the sonographic characteristics of all cases of papillary thyroid carcinoma imaged over a 7-year period (July 1995 to January 2003) at our institution. Thyroid malignancies of other histologic types, such as lymphoma and follicular carcinoma, were excluded. Real-time sonography of thyroids was performed with Acuson Sequoia and 128XP sonographic scanners (Siemens Medical Solutions, Mountain View, CA) equipped with commercially available 7- to 14-MHz linear probes. Color Doppler imaging and power Doppler imaging were performed with the aforementioned linear array transducers. A total of 55 patients with newly diagnosed papillary thyroid carcinoma were imaged during this period. At least 1 nodule was identified in 54 of 55 patients, with 1 patient having 2 separate nodules (one in each lobe) and 1 patient having lymphadenopathy but no detectible thyroid lesion. Thus, there were a total of 55 thyroid nodules representing papillary carcinoma. All cases were confirmed pathologically by FNA, thyroidectomy, or both. Patients with treated thyroid carcinoma were excluded from this study. Each case of papillary thyroid carcinoma was evaluated for 7 characteristics: echo texture, internal architecture, margin, contour, presence of a halo, vascularity, and calcification pattern. The echo texture of each lesion was classified as hypoechoic, hyperechoic, or mixed hypoechoic and isoechoic in comparison with the background thyroid tissue. The internal architecture was defined as solid, solid with cystic elements, or predominantly cystic. Predominantly cystic lesions were those containing cystic components that constituted more than 50% of the lesion. Margins of lesions were categorized as well defined when clear demarcation with normal thyroid was noted around more than 50% of a nodule and were considered poorly defined when more than 50% of the border of the lesion was not clearly demarcated. Furthermore, the overall shape of the lesion (contour) was classified as either smooth and round or irregular with angulated edges. The presence of a complete or incomplete hypoechoic halo around each lesion was also documented, and the predominating pattern of blood flow was classified as intrinsic to the lesion or perinodular. Color Doppler flow seen within a lesion was defined as intrinsic, whereas flow noted along the immediate margins of the lesion was considered perinodular. The intrinsic flow pattern was further classified as hypervascular or hypovascular with respect to background thyroid tissue. Finally, calcifications were classified as microcalcifications or coarse calcifications with or without peripheral distribution. Sonographic features present in at least 35% of cases were defined as common. Features that were present in less than 10% of cases were considered uncommon. We also looked at combinations of sonographic features to determine which overall appearances were most common. Results The sonographic features of the nodules are given in Table 1. At least 1 nodule was identified and sonographically characterized in 54 of 55 patients. One patient had 2 distinct foci of papillary cancer confirmed by biopsy. Another patient had papillary cancer confirmed by biopsy of a lymph node, with a normal-appearing thyroid gland. This patient was excluded from the remainder of the analysis. Thus, there were 55 distinct thyroid nodules that proved to be cancer on FNA. These 55 nodules constituted the study group for this report. Eighty-six percent of the nodules were hypoechoic lesions (Fig. 1), and only 1 (2%) of 55 was found to be hyperechoic (Fig. 2). Eighty-seven percent of the lesions were solid, with the remainder having cystic elements or predominantly cystic architecture. Of the 7 lesions that were not solid, 4 had small cystic elements, and 3 were predominant- 1084 J Ultrasound Med 22:1083 1090, 2003

Chan et al ly cystic papillary carcinomas (Fig. 3). Lesion margins were almost equally divided between well defined (47%) and poorly defined (53%), with most lesions having round contours. Irregular lesions with sharply defined angulations were noted in 4 (7%) of 55 nodules (Fig. 4). Complete or incomplete hypoechoic halos were identified in 7% of the lesions (Fig. 5). Color Doppler sonography showed that 43 lesions had a primarily intrinsic flow pattern (Fig. 6), including 5 that were hypovascular. Perinodular flow was the dominant pattern identified in 12 lesions, all of which also contained detectable flow within the nodule (Fig. 7). No lesion was entirely avascular. Of the 55 nodules, 26 had no calcifications. Of the 29 lesions with calcifications, 23 (79%) had microcalcifications (Fig. 1); 5 (17%) had coarse, flocculated calcifications (Fig. 8); and 1 (2%) had peripheral calcifications (Fig. 9). One patient had isolated lymphadenopathy without a visible thyroid lesion. Nine common and 11 uncommon sonographic features were identified in papillary carcinoma (Table 2). The common high-resolution sonographic features included a hypoechoic echo texture, either well-defined or poorly defined margins, round contours, and solid structures with evidence of intrinsic hypervascularity. In addition, papillary carcinomas were commonly found both with and without microcalcifications. Uncommon features of the papillary carcinomas in our series included a hyperechoic, mixed, or isoechoic echo texture, cystic elements, nodules in cyst configurations, sharp contours, halos, intrinsic hypovascularity, coarse or peripheral calcifications, and isolated lymphadenopathy. In total, 30 (54%) of 55 cases had at least 1 uncommon feature, including 24 cases (43%) with 1 uncommon feature and 6 (11%) with 2 or more uncommon features. On average, each nodule had 6.3 sonographic features considered common and 0.7 sonographic features considered uncommon. Given the preponderance of nodules that were hypoechoic with a solid internal architecture, it is not surprising that 40 (73%) of the 55 malignant nodules were both solid and hypoechoic. Of the cancers that were predominantly cystic, all contained solid elements with intrinsic vascularity. Similarly, all malignant solid nodules with cystic elements had an intrinsic or a perinodular flow pattern. Table 1. Sonographic Features of Papillary Thyroid Carcinoma Sonographic Findings (n = 55)* No. of Nodules % Echo texture Hypoechoic 47 86 Isoechoic 4 7 Mixed hypoechoic and isoechoic 3 6 Hyperechoic 1 2 Internal architecture Solid 48 87 Solid with cystic elements 4 7 Predominantly cystic 3 6 Margin Poorly defined 29 53 Well defined 26 47 Contour Smooth and round 51 93 Irregular with sharp angulations 4 7 Halo Absence 51 93 Presence 4 7 Vascularity Intrinsic 38 69 Perinodular 12 22 Hypovascular 5 9 Calcifications None 26 47 Microcalcifications 23 42 Coarse 5 9 Peripheral 1 2 *One additional case had isolated lymphadenopathy and was thus excluded from the analysis. Figure 1. Longitudinal sonogram of a typical hypoechoic, well-defined, round lesion with microcalcifications consistent with a papillary carcinoma. J Ultrasound Med 22:1083 1090, 2003 1085

Sonographic Features of Papillary Thyroid Carcinoma Figure 2. Longitudinal sonogram of a hyperechoic papillary carcinoma. Certain combinations of features did not occur, and, as noted above, only 6 nodules had 2 features found to be uncommon. Flow characteristics on color Doppler sonography seemed to be the most useful discriminators, in that papillary carcinomas with uncommon gray scale features were all hypervascular. For example, there were no nodules with halos that lacked either intrinsic or perinodular flow. Similarly, of the 7 papillary cancers with uncommon echo textures (i.e., not hypoechoic), none was hypovascular. Conversely, all hypovascular lesions had otherwise typical features on gray scale imaging; i.e., all were hypoechoic and solid, were round with well-defined margins, and lacked halos. No lesion that was predominantly cystic had solid elements lacking flow. All of our papillary cancers with 1 or more uncommon features on gray scale imaging had either intrinsic or perinodular flow. Discussion Figure 3. Longitudinal power Doppler sonogram of a cystic papillary carcinoma. Figure 4. Transverse sonogram of a papillary carcinoma with irregular margins. Thyroid nodules are common and are found on sonography in up to 50% of the population. 5 As the use of sonography in the cervical region has increased, detection of nonpalpable nodules has become an increasingly difficult but important treatment problem. Although most thyroid incidentalomas are benign, approximately 5% to 6.5% may be malignant. 6 Patients with nonpalpable thyroid nodules are now commonly referred for FNA of the nodules on the basis of the nodules sizes, with 1 cm representing a common threshold for referral. 1 Many studies have shown that the sonographic appearance of papillary carcinoma, the most common malignancy of the thyroid, is quite variable and that there is considerable overlap in features between benign and malignant nodules. Recently, however, Papini et al 6 showed that a treatment strategy of selecting nodules for FNA based only on a size threshold of 1 cm is not as efficient as one that takes the most common sonographic features of thyroid carcinoma into account. In their prospective study of FNA results in patients with nonpalpable thyroid nodules ranging in size from 8 to 15 mm, they found that 87% of the 31 cancers detected were hypoechoic; 77% had irregular margins; 74% had internal vascularity; and 29% had microcalcifications. By comparison, 57% of the 347 benign nodules were hypoechoic; 15% had irregular margins; 19% had 1086 J Ultrasound Med 22:1083 1090, 2003

Chan et al internal vascularity; and 4% had microcalcifications. A size threshold of 10 mm was not helpful because 9% of their cancers were between 8 and 10 mm versus 7% that were between 11 and 15 mm. A treatment strategy that included biopsy of all hypoechoic nodules would have resulted in detection of 87% of the cancers and would have required biopsy of 237 nodules, whereas a strategy based on a 10-mm threshold would have resulted in detection of 61% of cancers and would have required 271 biopsies. Another recent prospective study by Frates et al 7 explored the value of color Doppler flow patterns in predicting malignancy in 209 nodules referred for FNA. They found that solid hypervascular thyroid nodules were more likely to be malignant than partially cystic or nonhypervascular nodules. 7 However, there was significant overlap in color flow patterns, and benign and malignant nodules could not be reliably distinguished. Again, nodule size was not a useful predictor of malignancy, although their nodules were larger (range, 7 57 mm; average, 20.7 mm for benign nodules and 20.4 mm for malignant nodules) than those evaluated by Papini et al. 6 Given the interest in incorporating certain sonographic features into the treatment strategy for nonpalpable thyroid nodules, we sought to delineate the spectrum of appearances of papillary carcinoma of the thyroid and to compare our results with previous reports in the literature. We evaluated a relatively large series of proven cases of papillary thyroid cancer, the most common subtype of thyroid malignancy. To that end, our results generally confirm the notion that most papillary carcinomas of the thyroid are hypoechoic. In our study, 86% of the papillary cancers were hypoechoic; other authors found 77% to 89% of papillary thyroid carcinomas to be hypoechoic. 3,8,9 As in prior studies, we found echogenic papillary cancers to be rare, with only 1 case in our series. Although the precise histologic correlate of this finding is difficult to determine, pathologic examination of this nodule showed extensive areas of hemorrhage. We also confirmed that cystic thyroid cancers are uncommon, although 7 of our 55 nodules had some cystic elements. Nevertheless, we had no avascular cystic papillary thyroid cancers. We observed intrinsic flow in the solid components of all cases of cystic papillary carcinoma (Fig. 3). Unlike prior investigators, we did not find a high association of poorly defined margins in Figure 5. Longitudinal sonogram of an isoechoic papillary carcinoma with a surrounding rim of low echoes consistent with a halo. Figure 6. Longitudinal sonogram showing increased areas of flow within a papillary carcinoma consistent with a hypervascular intrinsic flow pattern. Figure 7. Longitudinal power Doppler sonogram of a papillary carcinoma with a perinodular pattern of vascularity consistent with a color halo. J Ultrasound Med 22:1083 1090, 2003 1087

Sonographic Features of Papillary Thyroid Carcinoma Figure 8. Longitudinal sonogram of a papillary carcinoma with coarse calcifications. papillary cancers. Only 53% of our cases showed poorly defined margins, compared with prior studies, which found poorly defined margins in 60% to 79% of cases. 6,9,10 Some of this difference may be due to the subjectivity inherent in evaluating the delineation of lesion borders. Nevertheless, well-demarcated papillary cancers may be more common than previously supposed. By contrast, the halo sign, a hypoechoic rim surrounding a thyroid nodule, was quite rare in our series, consistent with prior studies, suggesting that the halo sign is relatively uncommon in papillary carcinoma. 3,8,11 13 Fewer than 10% of our cases had either complete or incomplete halos. Figure 9. Longitudinal sonogram of a papillary carcinoma with peripheral eggshell calcifications. Our results also confirm that papillary cancer tends to be hypervascular. Most commonly, we found a pattern of intrinsic hypervascularity (69%) within the malignant nodule rather than perinodular flow (22%). Although the perinodular flow pattern or color halo sign is a less common pattern of vascularity, our findings suggest that it occurs twice as often as previously thought. 14 Although the hypovascular intrinsic flow pattern was uncommon, no lesions were found to be completely avascular. This result differs somewhat from the recent findings of Frates et al 7 because their series included 2 malignant nodules with no detectible intrinsic flow. However, their methods involved only color Doppler sonography, whereas in our series, both color Doppler sonography and power Doppler sonography were used. It is possible that the greater sensitivity to low flow of power Doppler sonography enabled us to detect even very weak signals in hypovascular nodules that might have appeared avascular on color Doppler sonography. Apart from its value in isolation, vascularity assessment was particularly useful in combination with other gray scale features. In our series, the absence of calcifications was nearly as common as their presence. A wide pattern of calcifications on sonography has been noted previously in papillary carcinoma, ranging from microcalcifications to dense calcifications. Microcalcifications are defined as punctate hyperechoic foci without acoustic shadowing, whereas dense calcifications are defined as hyperechoic foci with acoustic shadowing. 15 Histologically, coarse calcifications shown on sonography correlate with irregularly shaped calcifications that measure up to 2 mm and are usually located in fibrous tissue septa rather than epithelia. By contrast, microcalcifications represent psammoma bodies, laminated 10- to 100-µm round crystalline calcific deposits present in the epithelium that are associated with papillary carcinoma. 16,17 Occasionally, inspissated colloid as well as vascular calcifications can mimic the appearance of coarse shadowing calcifications on sonography. 18 Among these calcification patterns, we found that microcalcifications were much more common in papillary cancer, with 23 nodules having microcalcifications versus 5 nodules with coarse calcifications and only 1 with peripheral calcifications. On the whole, our results are consistent with recent studies, which found calcifications in as many as 52% to 68% of thyroid carcinomas. 13,16,19,20 1088 J Ultrasound Med 22:1083 1090, 2003

Chan et al Table 2. Common and Uncommon Sonographic Features of Papillary Thyroid Carcinoma Common (n = 9) Uncommon (n = 11) Hypoechoic Solid Well-defined or poorly defined margins Smooth and round Absence of halo Intrinsic hypervascularity No calcifications or microcalcifications Mixed hypoechoic and isoechoic, isoechoic, or hyperechoic Solid with cystic elements Predominantly cystic Irregular with sharp angulations Presence of halo Intrinsic hypovascularity Coarse or peripheral calcifications Isolated lymphadenopathy Given the multiplicity of individual features seen in papillary carcinoma, it is interesting to note the patterns that did not occur. There were no predominantly cystic nodules that lacked internal vascularity. Similarly, all 4 papillary carcinomas with halos had intrinsic hypervascularity or perinodular flow. All papillary carcinomas that were not hypoechoic had intrinsic or perinodular flow. It is thus unusual for papillary cancer lacking blood flow to also have other features more commonly associated with benign nodules, such as cystic architecture, a halo, or a nonhypoechoic texture. In summary, the typical appearance of papillary carcinoma is a round, solid, hypoechoic mass with internal vascularity and microcalcifications. Uncommon appearances of papillary carcinoma include cystic elements, a hyperechoic texture, intrinsic hypovascularity, sharp angular contours, and metastatic spread without a visible thyroid mass. Uncommon sonographic features were found to occur more often that expected. However, certain appearances were not associated with papillary cancer in our series. These included a hypovascular flow pattern on color Doppler imaging combined with a gray scale appearance of either predominantly cystic architecture, the presence of a halo, or a nonhypoechoic texture. Finally, all carcinomas in our series had at least some detectable flow on color and power Doppler imaging. References 1. Ross DS. Nonpalpable thyroid nodules: managing an epidemic. J Clin Endocrinol Metab 2002; 87: 1938 1940. 2. Tan GH, Gharib H. Thyroid incidentalomas: management approaches to nonpalpable nodules discovered incidentally on thyroid imaging. Ann Intern Med 1997; 126:226 231. 3. Solbiati L, Volterrani L, Rizzatto G, et al. The thyroid gland with low uptake lesions: evaluation by ultrasound. Radiology 1985; 155:187 191. 4. Clark OH, Duh QY. Thyroid cancer. Med Clin North Am 1991; 75:211 234. 5. Mortensen J, Woolner L, Bennett W. Gross and microscopic findings in clinically normal thyroid glands. J Clin Endocrinol Metab 1955; 15:1270 1280. 6. Papini E, Guglielmi R, Bianchini A, et al. Risk of malignancy in nonpalpable thyroid nodules: predictive value of ultrasound and color-doppler features. J Clin Endocrinol Metab 2002; 87:1941 1946. 7. Frates MC, Benson CB, Doubilet PM, Cibas ES, Marqusee E. Can color Doppler sonography aid in the prediction of malignancy of thyroid nodules? J Ultrasound Med 2003; 22:127 131. 8. Watters DA, Ahuja AT, Evans RM, et al. Role of ultrasound in the management of thyroid nodules. Am J Surg 1992; 164:654 657. 9. Lu C, Chang TC, Hsiao YL, Kuo MS. Ultrasonographic findings of papillary thyroid carcinoma and their relation to pathologic changes. J Formos Med Assoc 1994; 93:933 938. 10. Takashima S, Matsuzuka F, Nagareda T, Tomiyama N, Kozuka T. Thyroid nodules associated with Hashimoto thyroiditis: assessment with US. Radiology 1992; 185:125 130. J Ultrasound Med 22:1083 1090, 2003 1089

Sonographic Features of Papillary Thyroid Carcinoma 11. Nassani SN, Bard R. Evaluation of solid thyroid neoplasms by gray scale and real time ultrasonography: the halo sign. Ultrasound Med 1978; 4:323 324. 12. Brkljacic B, Cuk V, Tomic-Brzac H, Bence-Zigman Z, Delic-Brkljacic D, Drinkovic I. Ultrasonic evaluation of benign and malignant nodules in echographically multinodular thyroids. J Clin Ultrasound 1994; 22: 71 76. 13. Koike E, Noguchi S, Yamashita H, Murakami T, Ohshima A, Kawamoto H. Ultrasonographic characteristics of thyroid nodules: prediction of malignancy. Arch Surg 2001; 136:334 337. 14. Shimamoto K, Endo T, Ishigaki T, Sakuma S, Makino N. Thyroid nodules: evaluation with color Doppler ultrasonography. J Ultrasound Med 1993; 12:673 678. 15. Chang TC, Hong CT, Chang SL, Hsieh HC, Liaw KY, How SW. Correlation between sonography and pathology in thyroid diseases. J Formos Med Assoc 1990; 89:777 783. 16. Takashima S, Fukuda H, Nomura N, Kishimoto H, Kim T, Kobayashi T. Thyroid nodules: re-evaluation with ultrasound. J Clin Ultrasound 1995; 23:179 184. 17. Klinck GH, Winship T. Psammoma bodies and thyroid cancer. Cancer 1959; 12:656 662. 18. Katz JF, Kane RA, Reyes J, Clarke MP, Hill TC. Thyroid nodules: sonographic-pathologic correlation. Radiology 1984; 151:741 745. 19. Kakkos SK, Scopa CD, Chalmoukis AK, et al. Relative risk of cancer in sonographically detected thyroid nodules with calcifications. J Clin Ultrasound 2000; 28:347 352. 20. Kim EK, Park CS, Chung WY, et al. New sonographic criteria for recommending fine-needle aspiration biopsy of nonpalpable solid nodules of the thyroid. AJR Am J Roentgenol 2002; 178:687 691. 1090 J Ultrasound Med 22:1083 1090, 2003