Neuroradiology/Head and Neck Imaging Original Research Yoon et al. Taller-Than-Wide Sign of Thyroid Malignancy Neuroradiology/Head and Neck Imaging Original Research Soo Jeong Yoon 1 Dae Young Yoon 1,2 Suk Ki Chang 1,2 Young Lan Seo 1,2 Eun Joo Yun 1 Chul Soon Choi 1 Sang Hoon Bae 1 Yoon SJ, Yoon DY, Chang SK, et al. Keywords: comparative study, CT, sonography, thyroid cancer, thyroid nodule DOI:10.2214/AJR.09.3376 Received July 28, 2009; accepted after revision November 1, 2009. 1 Department of Radiology, Kangdong Seong-Sim Hospital, Hallym University College of Medicine, 445 Gil-dong Kangdong-Gu, Seoul 134-701, Republic of Korea. Address correspondence to D. Y. Yoon (evee0914@chollian.net). 2 Ilsong Memorial Institute of Head and Neck Cancer, Kangdong Seong-Sim Hospital, Hallym University College of Medicine, Seoul, Republic of Korea. WEB This is a Web exclusive article. AJR 2010; 194:W420 W424 0361 803X/10/1945 W420 American Roentgen Ray Society Taller-Than-Wide Sign of Thyroid Malignancy: Comparison Between Ultrasound and CT OBJECTIVE. The objective of our study was to investigate the mechanism of the tallerthan-wide sign that is, an anteroposterior dimension to transverse dimension ratio of 1 on ultrasound. MATERIALS AND METHODS. Ultrasound and CT images of 90 pathologically proven thyroid masses (57 malignant and 33 benign) smaller than 2 cm in 77 patients (mean age, 45 years) were retrospectively reviewed. Two readers assessed the anteroposterior and transverse dimensions of the mass, anteroposterior transverse ratio of the mass, anteroposterior dimension of the ipsilateral thyroid lobe, and the position of the common carotid artery (CCA) relative to the thyroid lobe. In addition, the difference in the anteroposterior transverse ratio of the mass between ultrasound and CT was correlated with the ultrasound characteristics of the thyroid mass (i.e., maximal diameter, location, location within lobe, and composition), histopathologic results, and ultrasound operator. RESULTS. The mean (± SD) anteroposterior transverse ratio of the thyroid masses on ultrasound was significantly lower than that on CT (0.97 ± 0.34 vs 1.07 ± 0.28, respectively; p < 0.001), and the differences were significantly greater in benign masses than malignant masses, in masses located at the anterior or mid third of the lobe than those located at the posterior third, and in cystic masses than mixed or solid masses. There were statistically significant differences between the two techniques with regard to the anteroposterior dimension of the ipsilateral thyroid lobe and the position of the CCA, suggesting the effect of probe compression. CONCLUSION. The mechanism of the taller-than-wide sign is no or minimal compressibility of a thyroid mass by the ultrasound probe, which occurs more frequently in malignant masses than in benign masses. C urrently, ultrasound is used as the technique of choice for evaluating thyroid masses because of its noninvasive nature, accessibility, and high spatial resolution. In addition to its use for characterizing thyroid masses, ultrasound is useful in differentiating benign from malignant masses. Several ultrasound features that have been identified in previous studies as being suggestive of malignancy include the presence of fine or coarse calcifications, hypoechogenicity, irregular margins, absence of a halo, predominantly solid composition, and intranodular vascularity [1 4]. In 2002, Kim et al. [1] described a new sign of thyroid cancer observed on ultrasound. They observed a nodule with an anteroposterior dimension transverse dimension ratio of > 1 as a highly predictive sign for malignancy and described this finding as a shape that was more tall than it was wide [1]. Thereafter, a number of ultrasound studies confirmed that a thyroid lesion with an anteroposterior transverse ratio of > 1 or 1 is a reliable criterion for malignancy [4 7] (Table 1). To our knowledge, however, the true mechanism of this ultrasound finding has not yet been established. We hypothesized that the mechanism of the taller-than-wide sign is no or minimal compressibility of a malignant thyroid mass by the ultrasound probe, whereas compressibility is seen more frequently in benign than malignant masses. The purpose of this study was to validate this hypothesis by comparing lesion shape between ultrasound and CT images obtained at the same anatomic level. Materials and Methods The entire study protocol was approved by our institutional review board, and informed consent was waived because the data were evaluated ret- W420 AJR:194, May 2010
Taller-Than-Wide Sign of Thyroid Malignancy rospectively after patient-identifying information had been removed. Patient Population A search of the hospital information system database for the period from January 2006 to June 2009 identified 99 patients with pathologically proven thyroid masses who underwent both preoperative ultrasound and CT. Because we attempted to assess the shape of the thyroid masses, the radiology reports were read and only patients with a thyroid mass smaller than 2.0 cm in maximum diameter were included. Of those cases, 23 patients were excluded for the following reasons: poorly defined border of the mass on CT or ultrasound (n = 7); diffuse enlargement of the thyroid gland (i.e., diffuse or multinodular goiter) (n = 6); and time lapse between ultrasound and CT examinations of more than 2 weeks (n = 10). This yielded a final study population of 76 patients (62 females and 14 males; mean age, 49.4 ± 12.6 years; range, 16 74 years) with 90 thyroid masses (a single mass in 62 patients and two masses [one in each lobe] in 14 patients). In patients with multiple thyroid masses, the largest mass in each lobe was chosen for evaluation. In all cases, the final diagnosis of a benign (n = 33) or malignant (n = 57) mass was determined pathologically by fine-needle aspiration (FNA) (n = 9), thyroidectomy (n = 31), or both (n = 50). Cases with benign FNA results had ultrasound follow-up for more than 1 year to rule out changes in the volume or imaging features of the nodule. All malignant thyroid tumors were papillary carcinomas on pathologic examination. Ultrasound and CT Examinations The ultrasound examinations were performed by two radiologists on an HDI 5000 or iu22 unit (Philips Healthcare) or an Acuson Sequoia 512 unit (Siemens Healthcare) using a 8 15-MHz linear-array transducer. All CT scans were obtained using a 16-MDCT scanner (MX8000 Infinite Detector Technology, Philips Healthcare) with the following parameters: 3-mm section thickness, pitch of 1.5, 4 1.5 mm collimation, 120 kv and 200 mas, and 45-second scan delay. The scanning range was planned in a craniocaudal direction from the level of the maxillary sinus to the tracheal bifurcation including the entire thyroid mass (mean coverage, 250 mm). Each patient received 100 ml of nonionic contrast material (iohexol [Omnipaque 300, GE Healthcare]) through an 18-gauge needle placed in a peripheral arm vein at a rate of 2 ml/s. Axial images were reconstructed at 3-mm increments on a 512 512 matrix. All patients in the series underwent ultrasound and CT within 2 weeks, with a mean interval between examinations of 5.0 days (range, 0 14 days). Thirty-five patients underwent ultrasound and CT on the same day. Data Collection and Analysis The ultrasound and CT images were independently evaluated on a PACS workstation (π-viewstar, Infinitt) by two radiologists. All images were presented without patient-identifying information in a random fashion. Each examination was allocated a study number that was known only to the study coordinator. Both readers were blinded to the findings of the other technique and assessment of the other investigator, but they were informed about the location of the histologically proven mass in patients with multiple thyroid masses. Each reader chose a representative transverse ultrasound image (and a corresponding level on CT) for each thyroid mass so the maximal area of the mass appeared on the images and independently assessed the following parameters: mass size (anteroposterior and transverse dimensions), mass shape (anteroposterior transverse ratio), anteroposterior dimension of the ipsilateral thyroid lobe, and position of the common carotid artery (CCA) relative to the ipsilateral thyroid lobe. The CCA was noted to be located at the level of the anterior third of the thyroid lobe or above, at the mid third, or at the posterior third or below. Reviewers then assessed the characteristics of the thyroid masses on the basis of the ultrasound findings: size (maximal diameter < 1 cm or 1 2 cm); location (right lobe, left lobe, or isthmus); location within lobe (right or left; upper third, mid third, or lower third; anterior third, mid third, or posterior third); and composition (cystic [cystic portion > 75%], mixed [cystic portion = 25 75%], or solid [cystic portion < 25%]). All quantitative measurements were performed on the workstation using electronic calipers after appropriate magnification; measures were made by each of the two readers, and the two independent measurements were then averaged. The anteroposterior dimension of the mass was defined as the diameter on the axis perpendicular to the anterior surface of the thyroid gland. The transverse dimension of the mass was defined as the diameter on the axis perpendicular to the diameter used for measurement of the anteroposterior dimension. If there were disagreements in qualitative assessments, the reviewers discussed and reached a final decision by consensus. To minimize learning bias, the ultrasound and CT images were reviewed in different randomized orders and were reviewed during two sessions separated by a 4-week interval. For each case, we compared the anteroposterior and transverse dimensions of the thyroid mass, anteroposterior transverse ratio of the thyroid mass, anteroposterior dimension of the ipsilateral thyroid lobe, and position of the CCA relative to the ipsilateral thyroid lobe between ultrasound and CT images. In addition, differences between the anteroposterior transverse ratios of the same thyroid mass ([anteroposterior transverse ratio on CT] [anteroposterior transverse ratio on ultrasound]) were analyzed to determine whether the differences were associated with characteristics of the thyroid mass (size, location, location within lobe, and composition), histopathologic result (benign vs malignant), or ultrasound operator. Statistical analysis was assessed with the paired Student s t test or one-way analysis of variance test TABLE 1: Reported Diagnostic Index for Taller-Than-Wide Sign of Malignant Thyroid Nodules First Author [Reference No.] Sensitivity Specificity PPV NPV Accuracy Comments Kim [1] 32.7 92.5 66.7 74.8 73.5 Nonpalpable solid nodules Iannuccilli [4] 44.1 72.2 60.0 57.8 58.6 Nodules > 1.0 cm Cappelli [5] 76.0 60.0 8.3 98.0 60.7 Moon [6] 40.0 91.4 77.4 67.4 69.6 Kim [7] 24.1 100.0 100.0 56.2 68.6 Nonpalpable solid nodules Note PPV = positive predictive value, NPV = negative predictive value. TABLE 2: Size and Shape of Thyroid Masses: Differences Between Ultrasound and CT Parameters Ultrasound CT p Anteroposterior dimension (mm) 9.17 ± 3.85 9.68 ± 4.11 < 0.005 Transverse dimension (mm) 10.04 ± 4.07 9.36 ± 3.80 < 0.001 Anteroposterior dimension transverse dimension ratio 0.97 ± 0.34 1.07 ± 0.28 < 0.001 Note All data except p values are presented as mean ± SD. AJR:194, May 2010 W421
Yoon et al. for continuous variables and the chi-square test for categoric variables. A p value of < 0.05 was considered to indicate a significant difference. Statistical analysis was performed using commercially available software (SPSS version 12.0, SPSS) for Microsoft Windows. A Fig. 1 61-year-old woman with benign nodular hyperplasia. A, CT scan shows hypodense nodule (arrow) in right lobe of thyroid gland with anteroposterior transverse ratio of 1.30. c = common carotid artery, v = internal jugular vein. B, Transverse ultrasound image corresponding to A shows hypoechoic nodule (arrow) with decrease in anteroposterior transverse ratio (0.80). Thyroid mass with shape taller than wide on CT is compressed and transformed into wide shape on ultrasound. Common carotid artery (c) is located at level of posterior third of thyroid lobe on CT (A) and has moved to level of anterior third of thyroid lobe on ultrasound (B). Internal jugular vein is completely collapsed on ultrasound. Results The values of the size and shape (anteroposterior transverse ratio) of the thyroid masses measured on ultrasound and CT are shown in Table 2. The mean anteroposterior transverse ratio of the thyroid masses based on ultrasound measurements was significantly lower than that based on CT measurements (0.97 ± 0.34 vs 1.07 ± 0.28, respectively; p < 0.001). In 10% (9/90) of the thyroid masses, the shape of the masses changed from an anteroposterior transverse ratio of 1.0 to an anteroposterior transverse ratio of < 1.0 on ultrasound examination compared with CT (Figs. 1 and 2). The differences in anteroposterior transverse ratios between ultrasound and CT were significantly greater in benign than malignant masses (p < 0.05), in masses located at the anterior or mid third than at the posterior third (p < 0.05), and in cystic masses than in mixed or solid masses (p < 0.05). The differences in anteroposterior transverse ratios of the thyroid masses in the right lobe or isthmus tended to be higher than those in the left lobe, although the p value was not statistically significant (p = 0.054). There was no relationship between the differences in anteroposterior transverse ratios and the size of the thyroid masses, location of the thyroid masses within the lobe (upper, mid, or lower), and ultrasound operator (Table 3). In addition, the mean anteroposterior dimension of the thyroid gland based on ultrasound measurements was significantly lower B A Fig. 2 36-year-old woman with papillary carcinoma. A, CT scan shows hypodense nodule (arrow) in right lobe of thyroid gland with anteroposterior transverse ratio of 1.08. c = common carotid artery, v = internal jugular vein. B, Transverse ultrasound image corresponding to A shows hypoechoic nodule (arrow) with slight decrease in anteroposterior transverse ratio (0.74). There is also difference between ultrasound and CT in positions of common carotid artery (c) relative to thyroid gland. B W422 AJR:194, May 2010
Taller-Than-Wide Sign of Thyroid Malignancy TABLE 3: Differences of Anteroposterior Dimension Transverse Dimension Ratio Between Ultrasound and CT According to Characteristics and Pathologic Result of Thyroid Mass and Ultrasound Operator Characteristic Anteroposterior Transverse Ratio Ultrasound than that based on CT measurements (15.08 ± 3.84 mm vs 18.60 ± 4.90 mm, respectively; p < 0.001). When the position of the CCA relative to the adjacent thyroid lobe was compared, the CCA had moved to a more anterior position on ultrasound examination as compared with CT (Figs. 1 and 2) in 49 (58.3%) of 84 masses and the CCA was located in the same position in the remaining 35 (41.7%) (p < 0.001) (Table 4). CT Difference Between CT and Ultrasound Ratios a Size 0.594 < 1.0 cm (n = 44) 0.92 ± 0.30 1.05 ± 0.25 0.08 ± 0.25 1.0 2.0 cm (n = 46) 0.97 ± 0.37 1.07 ± 0.32 0.11 ± 0.19 Location 0.054 Right lobe (n = 48) 0.97 ± 0.33 1.10 ± 0.26 0.13 ± 0.21 Isthmus (n = 6) 0.47 ± 0.05 0.68 ± 0.19 0.22 ± 0.18 Left lobe (n = 36) 1.05 ± 0.32 1.08 ± 0.29 0.03 ± 0.23 Location within lobe b Anterior third (n = 39) 0.97 ± 0.36 1.07 ± 0.29 0.10 ± 0.21 Mid third (n = 20) 1.00 ± 0.29 1.17 ± 0.29 0.17 ± 0.25 < 0.05 Posterior third (n = 25) 1.08 ± 0.29 1.07 ± 0.23 0.00 ± 0.20 Location within lobe b Upper third (n = 24) 1.06 ± 0.30 1.08 ± 0.23 0.02 ± 0.18 Mid third (n = 36) 0.99 ± 0.36 1.12 ± 0.30 0.14 ± 0.27 0.151 Lower third (n = 24) 0.98 ± 0.30 1.06 ± 0.27 0.08 ± 0.18 Composition < 0.05 Cystic (n = 11) 0.73 ± 0.45 1.00 ± 0.43 0.28 ± 0.22 Mixed (n = 14) 0.89 ± 0.32 0.98 ± 0.25 0.09 ± 0.12 Solid (n = 65) 1.03 ± 0.31 1.10 ± 0.26 0.07 ± 0.23 Pathologic result < 0.05 Benign (n = 33) 0.87 ± 0.35 1.05 ± 0.31 0.18 ± 0.21 Malignant (n = 57) 1.03 ± 0.32 1.08 ± 0.27 0.05 ± 0.22 Ultrasound operator 0.857 Reviewer 1 (n = 55) 0.96 ± 0.33 1.05 ± 0.31 0.09 ± 0.22 Reviewer 2 (n = 35) 0.98 ± 0.37 1.09 ± 0.25 0.10 ± 0.24 Note All data except p values are presented as mean ± SD. a Difference = [(ratio on CT) (ratio on ultrasound)]. b Isthmic lesions were excluded from analysis. Discussion A number of investigators have documented that the taller-than-wide sign of the thyroid mass, as the sole criterion or in combination with other ultrasound features, is useful for differentiating malignant from benign masses [1 7]. In their studies, the presence of the taller-than-wide sign had a relatively high specificity ranging from 60.0% to 100.0%, but the sensitivity was low, ranging from 24.1% to 76.0% [1, 2, 5 7] (Table 1). To our knowledge, this study is the first to investigate the mechanism of the taller-thanwide sign. The results of our study showed that the mean anteroposterior transverse ratio of thyroid masses obtained with ultrasound measurements was significantly lower than that obtained with CT measurements. Kim et al. [1], who initially described the taller-than-wide sign, postulated that benign nodules grow parallel to normal tissue planes, whereas malignant nodules (taller than wide) grow across normal tissue planes (they did p not mention a reference article). This hypothesis is similar to those previously reported in breast ultrasound studies: It has been generally accepted that nodules in the breast that are taller than wide are more likely to be malignant [8 10]. To our knowledge, however, there are no studies in the literature about the presence of a tissue plane in the thyroid gland. Furthermore, a recent ultrasound study [11] revealed that the taller-than-wide sign was also observed in cases of local tumor recurrence on the thyroid bed after thyroidectomy (47.4% sensitivity and 100.0% specificity). The results of our study showed that the anteroposterior dimension of the thyroid gland measured on ultrasound was significantly lower than that with measured on CT. This difference suggests that pressure during ultrasound on a thyroid mass may compress normal thyroid tissue as well as mass. In addition, when the ipsilateral CCA was used as a landmark to compare ultrasound and CT images, the CCA had moved to a more anterior position on ultrasound examination in 58.3% of the cases. This finding may be explained by the posterior displacement of the thyroid gland and confirms the presence of pressure being applied by the ultrasound probe. Our results also showed that differences in anteroposterior transverse ratios between ultrasound and CT were significantly associated with the pathologic result, composition, and location within the lobe of the thyroid masses. One might hypothesize that compared with malignant thyroid tumors, benign tumors had greater compressibility because benign tumors generally tend to be softer than malignant masses and to have less infiltration to surrounding tissue than malignant masses. This hypothesis can also be applied to cystic masses: In general, cystic masses are softer and therefore may be more compressible than mixed or solid masses. Furthermore, the compressibility of thyroid masses seems to be dependent on their location. Thyroid masses in the anterior portion of the lobe are thought to be more susceptible to compression than those in the posterior portion. In this study, the differences in anteroposterior transverse ratios between ultrasound and CT were higher in masses in the right lobe and isthmus than in masses in the left lobe, although no significant difference was reached (p = 0.054). These findings can be explained by the influence of the esophagus between the thyroid gland and vertebral body on the compressibility of thyroid masses. AJR:194, May 2010 W423
Yoon et al. TABLE 4: Anteroposterior Dimension of Thyroid Gland and Location of the Common Carotid Artery Relative to the Thyroid Gland: Differences Between Ultrasound and CT Parameters Ultrasound CT p Anteroposterior dimension (mm) of thyroid gland, 15.08 ± 3.84 18.60 ± 4.90 < 0.001 mean ± SD Location of common carotid artery a, no. < 0.001 Anterior third or above 22 (26.2) 2 (2.4) Mid third 53 (63.1) 41 (48.8) Posterior third or below 9 (10.7) 41 (48.8) a Isthmic lesions were excluded from analysis. Many previous studies have shown that the taller-than-wide sign is very specific for differentiating malignant thyroid nodules from benign ones. However, a review of the ultrasound and CT results in our series revealed that the shape of the thyroid mass can be changed by probe compression, and this change in shape was more prominent in benign lesions, cystic lesions, and lesions located at the anterior third of the thyroid gland. Therefore, we believe that the shape of the thyroid mass on ultrasound examination must be interpreted while considering the degree of probe compression and consistency of the mass in other words, the compressibility of the mass. There were some limitations of our study. First, the inclusion of only patients who had undergone both ultrasound and CT may have introduced a bias because patients with obviously benign findings at ultrasound and FNA usually did not undergo CT. Therefore, a substantial number of benign thyroid masses may have been excluded inadvertently. Second, although we excluded several thyroid masses that did not have a clearly defined border on CT, it cannot be ruled out that the CT measurements of small lesions may be influenced by partial volume effects or beam-hardening artifacts arising from high attenuation of the thyroid tissue. Finally, the amount of probe compression was not objectively quantified. Further prospective studies with quantitative assessment of the pressure applied by the ultrasound probe are needed in the future. In conclusion, the mean anteroposterior dimension transverse dimension ratio of thyroid masses based on ultrasound measurements was significantly lower than that based on CT measurements. Our results indicate that the mechanism of the taller-thanwide sign is no or minimal compressibility of a thyroid mass by the ultrasound probe, which can occur more frequently in malignant than benign masses. References 1. 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 2002; 178:687 691 2. 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 3. 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 4. Iannuccilli JD, Cronan JJ, Monchik JM. Risk for malignancy of thyroid nodules as assessed by sonographic criteria: the need for biopsy. J Ultrasound Med 2004; 23:1455 1464 5. Cappelli C, Castellano M, Pirola I, et al. Thyroid nodule shape suggests malignancy. Eur J Endocrinol 2006; 155:27 31 6. Moon WJ, Jung SL, Lee JH, et al. Benign and malignant thyroid nodules: US differentiation multicenter retrospective study. Radiology 2008; 247: 762 770 7. Kim JY, Lee CH, Kim SY, et al. Radiologic and pathologic findings of nonpalpable thyroid carcinomas detected by ultrasonography in a medical screening center. J Ultrasound Med 2008; 27: 215 223 8. Stavros AT, Thickman D, Rapp CL, Dennis MA, Parker SH, Sisney GA. Solid breast nodules: use of sonography to distinguish between benign and malignant lesions. Radiology 1995; 196:123 134 9. Fornage BD, Sneige N, Faroux MJ, Andry E. Sonographic appearance and ultrasound-guided fine-needle aspiration biopsy of breast carcinomas smaller than 1 cm 3. J Ultrasound Med 1990; 9:559 568 10. Fornage BD, Lorigan JG, Andry E. Fibroadenoma of the breast: sonographic appearance. Radiology 1989; 172:671 675 11. Lee JH, Lee HK, Lee DH, et al. Ultrasonographic findings of a newly detected nodule on the thyroid bed in postoperative patients for thyroid carcinoma: correlation with the results of ultrasonography-guided fine-needle aspiration biopsy. Clin Imaging 2007; 31:109 113 W424 AJR:194, May 2010