- Email: [email protected]
Comparison between Thin-Slice 3-D Volumetric Ultrasound and Conventional Ultrasound in the Differentiation of Benign and Malignant Thyroid Lesions
Ultrasound in Med. & Biol., Vol. 41, No. 12, pp. 3096–3101, 2015 Copyright Ó 2015 World Federation for Ultrasound in Medicine & Biology Printed in the USA. All rights reserved 0301-5629/$ - see front matter
http://dx.doi.org/10.1016/j.ultrasmedbio.2015.06.022
d
Original Contribution COMPARISON BETWEEN THIN-SLICE 3-D VOLUMETRIC ULTRASOUND AND CONVENTIONAL ULTRASOUND IN THE DIFFERENTIATION OF BENIGN AND MALIGNANT THYROID LESIONS WEN-BO LI,* BO ZHANG,* QING-LI ZHU,* YU-XIN JIANG,* JIAN SUN,y MENG YANG,* and JIAN-CHU LI* * Department of Ultrasound, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Dongcheng District, Beijing, China; and y Department of Pathology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Dongcheng District, Beijing, China (Received 26 February 2015; revised 29 May 2015; in final form 25 June 2015)
Abstract—We explored the efficacy of thin-slice volumetric 3-D ultrasound (3-DUS) in distinguishing between benign and malignant thyroid nodules. A total of 103 thyroid nodules were evaluated prospectively using 3-D gray-scale ultrasonography. The shape, margin, halo and potential capsular invasion of the nodules were compared with the findings of conventional 2-D ultrasound (2-DUS). Of the 103 thyroid nodules, there were 50 pathologically confirmed benign lesions and 53 malignant lesions (51.5%). Shape irregularity, ill-defined margins and capsular invasion provided sensitivities of 90.0%, 47.2% and 39.6% and specificities of 88.0%, 84.0% and 100%, respectively, for the malignant lesions. The diagnosis of thyroid cancer was improved in 3-DUS compared with 2-DUS, with a sensitivity of 88.7%, specificity of 90.0%, positive predictive value of 90.4%, negative predictive value of 88.2% and accuracy of 89.3%. The sensitivity of detection for lesions with capsular invasion increased to 39.6% with 3-DUS, more than twice that of 2-DUS. Three-dimensional US is highly accurate in diagnosing thyroid nodules, particularly those with capsular invasion. (E-mail: [email protected]) Ó 2015 World Federation for Ultrasound in Medicine & Biology. Key Words: Thyroid nodule, 3-D ultrasound, Diagnosis.
INTRODUCTION
explored the application of 3-DUS for thyroid diseases; most studies have focused on applications for thyroid/ nodule volume measurement (Andermann et al. 2007; Lyshchik et al. 2004; Ying et al. 2008). We analyzed the 3-D volumetric characteristics of thyroid nodules by performing multiplanar reconstruction of the 3-D volumetric data of the thyroid nodules and explored the diagnostic value of 3-DUS in the differential diagnosis of these lesions by comparing the lesion features determined in this manner with those from 2-D ultrasound (2-DUS).
Ultrasound has an accuracy of 67%–92% in the diagnosis of thyroid nodules (Moon et al. 2007; Stacul et al. 2007; Tae et al. 2007) and remains one of the most frequently used approaches for imaging these lesions. However, ultrasound suffers certain limitations: (i) conventional static images sometimes do not completely reflect the structures of the thyroid nodules; (ii) the procedure is highly dependent on the operator(s); and (iii) systematic reviews and analyses are often not available. By contrast, 3-D ultrasound (3-DUS) can be used to automatically acquire a series of 2-D images containing spatial data. Three-dimensional US avoids the limitations of gray-scale ultrasound by displaying the lesion features via reconstruction. However, few studies to date have
METHODS Patients The experimental protocol was approved by a local institutional review board, and informed consent for the study was obtained from all human patients. Ninety-six consecutive patients with complete pre-operative 3DUS data who underwent surgical operations for thyroid nodules from March 2008 to April 2009 were prospectively enrolled in this study. Two patients withdrew
Address correspondence to: Yu-Xin Jiang, Department of Ultrasound, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Shuaifuyuan 1, Dongcheng District, Beijing 100730, China. E-mail: [email protected] 3096
Thin-slice 3-D US vs. conventional US d W.-B. LI et al.
from the study before surgery; therefore, 103 thyroid nodules in 94 patients were included in the final analysis. Clinical physicians make surgical decisions based on (i) suspicion of malignancy or malignancy diagnosed on fine-needle aspiration (FNA) (n 5 12, 12.8%); (ii) high suspicion of malignancy on US without FNA (n 5 42, 44.7%); (iii) suspected recurrence after resection of thyroid papillary carcinoma (n 5 2, 2.1%); (iv) multiple endocrine neoplasms (n 5 2, 2.1%); (v) metastatic cancer from the thyroid confirmed by pathology (n 5 2, 2.1%); (vi) hyperthyroidism with thyroid nodules (n 5 3, 3.2%); and (vii) retrosternal goiter or compression of the trachea (n 5 31, 33.0%). Of the 103 investigated nodules, 18 (17.5%) were from male patients and 85 (82.5%) were from female patients. The mean age of these patients was 43.6 6 12.2 y (range: 19–71 y). Machine and examinations Sonographic examinations were performed with a Voluson E8 (GE Healthcare, Milwaukee, WI, USA) with an 11 L-D high frequency probe with a frequency range of 6–15 MHz for 2-DUS and a RSP6-16-D 3-D volumetric probe for 3-DUS. The default conditions for thyroid examinations were applied. During examination, the patient was in the supine position, and the anterior area of the neck was fully exposed. Two-dimensional US was performed first, and during this procedure, sagittal, coronal and oblique gray-scale images of the thyroid and its surrounding tissues and lymph nodes were obtained. The dynamic range, depth, gain and focus were adjusted to obtain the optimal image. The shape, border, margin, halo and potential capsular invasion of the thyroid nodules were recorded. Three-dimensional US was then initiated. The patient was asked to hold his or her breath and to avoid swallowing. Three-dimensional volumetric US was performed using the built-in 3-D grayscale mode with the RSP6-16-D 3-D volumetric probe. The 3-D sample frame was set to include the index lesion and the maximum amount of adjacent tissues, approximately 3–4 cm in depth and 4 cm in width. The sweep angle was 15–29 based on the nodule size. Then the probe was stabilized and initial automatic volume data were acquired. The digital images acquired with conventional 2-DUS and 3-DUS were transferred to a personal computer and analyzed using 4-D VIEW PC software (Version 6.2), which provides the same interface as the scanner. Image analysis Volumetric data were analyzed using the VOCAL (virtual organ computer-aided analysis) imaging program (GE Healthcare, Kretz Ultrasound, Zipf, Austria), which enables the automatic calculation of volumetric data. The ‘‘manual mode’’ was applied in our study. The image was
3097
rotated 30 for each analysis, and the profile of each nodule was drawn in six planes to obtain the shape of the nodule. Then, the size of each nodule was obtained by drawing the nodule twice and calculating the average value. Two radiologists with at least 5 y of experience in thyroid sonography, who were blinded to the clinical conditions and pathologic findings of the patients, independently evaluated the 2-DUS and 3-DUS features of the lesions and then summarized the diagnostic results. If the two radiologists did not agree on the diagnosis, the result was discussed until agreement was reached. Multiplanar imaging analysis was applied to the 3-D images, during which three perpendicular sections were used—a horizontal section (plate A), a longitudinal section (plate B) and a coronal section (plate C)—along with a spatial 3-D image of coronal section obtained after these three sections were combined. The various planes could display 3-D images from different sections and in different directions. To evaluate nodule features, a 100% surface smooth algorithm was chosen for volume rendering of 3-D data. The thickness of the slice was adjusted according to lesion size empirically, approximately 15%–25% of the maximal diameter of the nodule. The features of the lesions observed via 2-DUS and 3-DUS included shape (regular or irregular) (Fig. 1), margin (clear or unclear), halo (complete and regular, irregular/incomplete or none) (Fig. 2) and capsular invasion (distant, adjacent or invaded) (Fig. 3). The halo features were defined as follows: a regular slim hypoechoic rim around the thyroid nodule (complete and regular halo); a hypoechoic rim around the thyroid nodule, with varied width or with interruptions (irregular/incomplete halo); or no obvious hypoechoic rim visible around the thyroid nodule (absence of halo). The relationship between the nodule and the capsule was defined as follows: distant 5 thyroid gland tissue visible between the nodule and the capsule; adjacent 5 an unbroken capsule despite the absence of thyroid gland tissue between the nodule and the capsule; capsular invasion 5 continuously interrupted capsule. Pathologic diagnosis Two pathologists with at least 5 y of experience in thyroid pathology independently issued their diagnoses after observing the sections of the surgically removed thyroid lesions. These pathologists were also blinded to the ultrasound results at diagnosis. If the two pathologists did not agree on the diagnosis, the result was further discussed until agreement was reached. Statistical analysis The statistical analysis was performed using SPSS 11.0 software (SPSS, Chicago, IL, USA). Normally distributed data are expressed as the mean 6 standard
3098
Ultrasound in Medicine and Biology
Volume 41, Number 12, 2015
Fig. 1. Papillary thyroid carcinoma in a 34-y-old woman. (a) Two-dimensional gray-scale ultrasonography indicates a hypo-echoic thyroid nodule (maximal diameter 14 mm) with a regular shape, clear margin and calcifications (hollow arrows). (b) Multiplanar reconstruction (MPR) and thin-slice smooth-surface rendering 3-D gray-scale ultrasonography indicate nodule with irregular lobulations (arrows), which were invisible on corresponding 2-D image.
deviation (x 6 SD) and were analyzed using the t-test. Data that were not normally distributed are expressed as the range and were analyzed using the rank-sum test. Count data were analyzed using the c2-test and Fisher’s exact probability test. The level of agreement on the 3D final diagnosis was evaluated by two radiologists. Inter-observer variability was measured using the k index of agreement. p , 0.05 was regarded as the threshold for statistical significance.
RESULTS
group was aged 41.9 6 13.5 y (range: 19–71 y) (p 5 0.136). Of the 103 nodules, 18 were from male patients and 85 were from female patients. Benign lesions had an average size of 5.65 cm3 (range: 0.14– 39.78 cm3), significantly greater than that of the malignant lesions (mean: 4.00 cm3, range: 0.29–18.65 cm3) (p 5 0.048). These 53 malignant nodules included papillary thyroid carcinoma (n 5 48), follicular carcinoma (n 5 1), medullary carcinoma (n 5 3) and lymphoma (n 5 1). The 50 benign nodules included nodular goiter (n 5 40), thyroid adenoma (n 5 8), Hashimoto’s thyroiditis (n 5 1) and fibrous thyroiditis (n 5 1).
General data and pathologic types of thyroid nodules Among the 103 thyroid nodules, there were 50 pathologically confirmed benign lesions and 53 malignant lesions (51.5%). The mean age of the patients was 43.6 6 12.2 y (range: 19–71 y). The benign group was aged 45.5 6 10.6 y (range: 20–68 y), and the malignant
Ultrasonographic features and diagnostic values of 2DUS and 3-DUS for thyroid nodules The ultrasonographic features identified via 2-DUS and 3-DUS for the thyroid nodules are listed in Table 1. For both 2-DUS and 3-DUS, the features ‘‘capsular
Fig. 2. Nodular goiter in a 35-y-old woman. (a) Two-dimensional gray-scale ultrasonography indicates a cystic-solid thyroid nodule (maxmal diameter 26 mm) with regular shape and a clear margin, but no halo (hollow arrows). (b) MPR and thin-slice smooth-surface rendering 3-D gray-scale ultrasonography reveal a complete and regular halo (arrows) prominently. MPR 5 multiplanar reconstruction.
Thin-slice 3-D US vs. conventional US d W.-B. LI et al.
3099
Fig. 3. Papillary thyroid carcinoma in a 69-y-old woman. (a) Two-dimensional gray-scale ultrasonography indicates a typical malignant thyroid nodule (maximal diameter 14 mm), but without evidence of capsular invasion (hollow arrows). (b) MPR and thin-slice smooth-surface rendering 3-D gray-scale ultrasonography reveal that the hyper-echoic capsule was interrupted by hypo-echoic tumor invasion (arrow). MPR 5 multiplanar reconstruction.
invasion,’’ ‘‘irregular shape’’ and ‘‘unclear margin’’ were used in the diagnosis of thyroid cancer. Associated sensitivity, specificity, positive predictive value, negative predictive value and accuracy are summarized in Table 2. The ‘‘irregular shape’’ feature of 3-DUS had the highest sensitivity and specificity in diagnosing thyroid cancer. When the ‘‘complete and regular halo’’ feature of 3DUS was used for the diagnosis of benign thyroid nodules, its sensitivity was 54.0%, specificity 89.1%, positive predictive value 96.4%, negative predictive value 69.3% and accuracy 76.7%. For 3-DUS, the sensitivity was 88.7%, specificity 90.0%, positive predictive value 90.4%, negative predictive value 88.2% and accuracy 89.3%; for 2-DUS, the corresponding measures were 86.8%, 86.0%, 86.8%, 86.0% and 86.4%, respectively. The k index for agree-
Table 1. Ultrasonographic features of 2-DUS and 3-DUS for thyroid nodules 3-DUS Ultrasonographic feature Shape Regular Irregular Capsule Distant Adjacent Invasion Halo Regular and complete Irregular/incomplete None Margin Clear Unclear
ment between two observers was 0.85 for the 3-D grayscale ultrasonographic diagnoses. DISCUSSION Baum and Greenwood first described 3-D ultrasound imaging in 1961 (Baum and Greenwood 1961). By scanning and acquiring a series of 2-D ultrasonographic images with spatial information, 3-DUS can be used for 3-D reconstruction and display. The past two decades have witnessed the rapid development of electronic technology and image processing technology, accompanied by the wide application of 3-DUS in the clinical management of obstetric and gynecologic problems, heart diseases, liver diseases and prostate disorders (Alcazar and Galvan 2009; Armstrong et al. 2013; Correale et al. 2008; Hotta et al. 2007; Mehta et al. 2004; Pretorius et al. 1998; Xu et al. 2003). However, few studies have reported the application of 3-DUS in the management of thyroid disorders.
2-DUS
Table 2. Diagnostic values of 2-DUS and 3-DUS for thyroid nodules
Benign
Malignant
Benign
Malignant
44 6
5 48
45 5
7 46
Ultrasonographic Sensitivity Specificity feature (%) (%)
5 45 0
2 30 21
5 45 0
2 43 8
27 14 9
1 19 33
25 11 14
1 11 41
42 8
28 25
39 11
18 35
Irregular shape 3-DUS 2-DUS Unclear margin 3-DUS 2-DUS Capsular invasion 3-DUS 2-DUS
2-DUS (3-DUS) 5 2-D (3-D) ultrasonography.
PPV (%)
NPV Accuracy (%) (%)
90.6 86.8
88.0 90.0
88.9 90.2
89.8 86.5
89.3 88.3
47.2 66.0
84.0 78.0
75.8 76.1
60.0 68.4
65.0 71.8
61.0 52.6
68.9 56.3
39.6 15.1
100 100
100 100
NPV 5 negative predictive value; PPV 5 positive predictive value; 2-DUS (3-DUS) 5 2-D (3-D) ultrasonography.
3100
Ultrasound in Medicine and Biology
Three-dimensional US volumetric data of thyroid nodules can reveal the morphologic features of lesions on multiple planes. Using three mutually perpendicular sections, we can better evaluate the complicated growth patterns and margins of a nodule and, thus, more accurately identify its overall profile and its anatomic relationships with neighboring structures while avoiding excessive dependence on operators and missing information because of non-standard sections (Fernandez et al. 2004; Munoz Agel and Varas Lorenzo 2005). Additionally, volumetric data can be retrieved and reviewed at any time for any suspicious lesion, thereby providing important information supplementary to 2DUS (Fenster et al. 2001). The volumetric data provided by 3-DUS enable repeated observations of the morphologic features of the nodules from multiple planes and various angles. In our study, the ‘‘irregular shape’’ feature had a sensitivity of 90.6% in the diagnosis of malignant thyroid nodules and was thus the most sensitive sonographic feature. As in our study, Slapa et al. (2011) found that ‘‘a lobulated shape of the nodule in the c-plane’’ had a sensitivity of 94%–100% in the diagnosis of thyroid cancer. In the present study, in one case, the nodule was diagnosed as benign (size 4.0 3 2.6 cm) based on 2-DUS; however, 3-DUS indicated that it had an irregular shape and therefore should be diagnosed as malignant. Post-operative pathology revealed a gray solid nodule 0.9 3 0.6 cm in size, which was confirmed to be a papillary carcinoma of the thyroid. One possible reason for such a difference may be that the morphologic changes in some malignant thyroid nodules are relatively slight and only the local shape of a tumor may become irregular. Because of the complexity of tumor shapes, conventional ultrasound may not be able to (fully) display the lesions and thus may lead to misdiagnoses in some cases. By contrast, the reconstruction of volumetric data enables more detailed and accurate visualization of tumors and thus provides more diagnostic information. In three cases, including one benign lesion and two malignant lesions, the shapes of the thyroid nodules were regular on 2DUS but irregular on 3-DUS. Therefore, benign lesions may also have irregular shapes, similar to those of malignant lesions. However, this irregular characterization of the benign lesion may also be attributable to the rater’s experience. Thyroid cancer manifests as multiple irregular shapes; however, as a result of prolonged growth, benign nodules (e.g., a nodular goiter) may exhibit secondary changes such as bleeding, necrosis and calcification, which can appear as irregular changes. Nevertheless, such changes tend to be relatively slight. Three-dimensional US can improve the visualization of the borders/margins of thyroid nodules and thus provide important information for the differential diag-
Volume 41, Number 12, 2015
nosis of benign and malignant lesions. In the present study, 13 lesions without halos on 2-DUS displayed complete or incomplete halos on 3-DUS (an increase of 12.6%). The ‘‘complete and regular halo’’ feature on 3DUS was a key feature in the diagnosis of benign thyroid nodules, with a specificity of 98.1% and a positive predictive value of 96.4%. Compared with 2-DUS, 3-DUS indicated clear margins in a larger number of nodules. This improvement may be related to the current volumetric imaging data processing mode of 3-DUS, which can provide more homogeneous multiplanar imaging echo patterns, decreased noise and enhanced contrast, thereby improving the visualization of nodular margins (Slapa et al. 2006). However, a clear margin is not a feature that is unique to benign nodules. In the present study, 84% (42/50) of the benign nodules had clear margins on 3-DUS; however, 53% (28/53) of the malignant nodules (including 24 cases of papillary carcinoma, 1 case of follicular carcinoma, 2 cases of medullary carcinoma and 1 case of lymphoma) also displayed clear margins. Malignant thyroid nodules typically have blurred and rough margins, reflecting the infiltration of the malignant tumor cells into the surrounding tissues. However, such infiltration may not be obvious in some malignant nodules; alternatively, fibrous capsules may form. As a result, the nodular margin may appear to be clear on ultrasound. Three-dimensional US can clearly visualize nodular margins and thus allow for accurate evaluation of the relationship between the nodule and capsule. ‘‘Capsular invasion’’ is a key feature for the diagnosis of thyroid cancer using ultrasound. ‘‘Capsular invasion’’ is highly specific and is regarded as an important factor affecting determination of the stage and prognosis of thyroid cancer, as well as the selection of the appropriate surgical procedure (Cooper et al. 2009). However, 2-DUS offers suboptimal sensitivity in the detection of capsular invasion, particularly when this feature is affected by multiple reflection artifacts in the ultrasonic near field (Nelson et al. 2000). As illustrated in our study, the ‘‘capsular invasion’’ feature in 3-DUS exhibited a sensitivity of 39.6% in diagnosing thyroid cancer, which was significantly higher than the corresponding sensitivity in 2-DUS (15.1%). Thus, clear visualization of the relationship between the nodule and capsule using 3-DUS is beneficial in improving the diagnostic accuracy for thyroid cancer and thus can play a role in the staging and prognosis of thyroid cancer. Three-dimensional ultrasound does, however, have certain limitations. For example, some information will be unavoidably missed during 3-D reconstruction, and 3-D artifacts may be produced as a result of involuntary movement or the beating of the carotid artery during data acquisition (Nelson et al. 2000). However, the image quality does not affect the image analysis or diagnosis.
Thin-slice 3-D US vs. conventional US d W.-B. LI et al.
This was a preliminary study of the role of 3-D grayscale ultrasound in the diagnosis of thyroid nodules. On the basis of our limited sample size, the majority of malignant thyroid nodules were found to be papillary carcinoma of the thyroid, and most benign nodules were nodular goiter. Only a small proportion of patients had other pathologic types. However, studies with larger sample sizes are warranted to validate our findings. Acknowledgments—This work is supported by the National Natural Science Foundation of China (81171354).
REFERENCES Alcazar JL, Galvan R. Three-dimensional power Doppler ultrasound scanning for the prediction of endometrial cancer in women with postmenopausal bleeding and thickened endometrium. Am J Obstet Gynecol 2009;200:44.e1–44.e6. Andermann P, Schlogl S, Mader U, Luster M, Lassmann M, Reiners C. Intra- and interobserver variability of thyroid volume measurements in healthy adults by 2-D versus 3-D ultrasound. Nuklearmedizin 2007;46:1–7. Armstrong L, Fleischer A, Andreotti R. Three-dimensional volumetric sonography in gynecology: An overview of clinical applications. Radiol Clin North Am 2013;51:1035–1047. Baum G, Greenwood I. A critique of time–amplitude ultrasonography. Arch Ophthalmol 1961;65:353–365. Cooper DS, Doherty GM, Haugen BR, Kloos RT, Lee SL, Mandel SJ, Mazzaferri EL, McIver B, Pacini F, Schlumberger M, Sherman SI, Steward DL, Tuttle RM. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid 2009;19:1167–1214. Correale M, Ieva R, Di Biase M. Real-time three-dimensional echocardiography: An update. Eur J Intern Med 2008;19:241–248. Fenster A, Downey DB, Cardinal HN. Three-dimensional ultrasound imaging. Phys Med Biol 2001;46:R67–R99. Fernandez LJ, Aguilar A, Pardi S. Three-dimensional ultrasound in small parts: Is it just a nice picture? Ultrasound Q 2004;20: 119–125. Hotta N, Ayada M, Okumura A, Ishikawa T, Kakumu S. Usefulness of live 3-D echocardiography during radiofrequency ablation in a
3101
case of hepatocellular carcinoma. Clin Imaging 2007;31: 283–286. Lyshchik A, Drozd V, Reiners C. Accuracy of three-dimensional ultrasound for thyroid volume measurement in children and adolescents. Thyroid 2004;14:113–120. Mehta SS, Azzouzi AR, Hamdy FC. Three dimensional ultrasound and prostate cancer. World J Urol 2004;22:339–345. Moon HG, Jung EJ, Park ST, Ha WS, Choi SK, Hong SC, Lee YJ, Joo YT, Jeong CY, Choi DS, Ryoo JW. Role of ultrasonography in predicting malignancy in patients with thyroid nodules. World J Surg 2007;31:1410–1416. Munoz Agel F, Varas Lorenzo MJ. Tridimensional (3-D) ultrasonography. Rev Esp Enferm Dig 2005;97:125–134. Nelson TR, Pretorius DH, Hull A, Riccabona M, Sklansky MS, James G. Sources and impact of artifacts on clinical threedimensional ultrasound imaging. Ultrasound Obstet Gynecol 2000;16:374–383. Pretorius DH, Nelson TR, Baergen RN, Pai E, Cantrell C. Imaging of placental vasculature using three-dimensional ultrasound and color power Doppler: A preliminary study. Ultrasound Obstet Gynecol 1998;12:45–49. Slapa RZ, Jakubowski WS, Slowinska-Srzednicka J, Szopinski KT. Advantages and disadvantages of 3-D ultrasound of thyroid nodules including thin slice volume rendering. Thyroid Res 2011;4:1. Slapa RZ, Slowinska-Srzednicka J, Szopinski KT, Jakubowski W. Grayscale three-dimensional sonography of thyroid nodules: Feasibility of the method and preliminary studies. Eur Radiol 2006;16: 428–436. Stacul F, Bertolotto M, De Gobbis F, Calderan L, Cioffi V, Romano A, Zanconati F, Cova MA. US, colour-Doppler US and fine-needle aspiration biopsy in the diagnosis of thyroid nodules. Radiol Med 2007;112:751–762. Tae HJ, Lim DJ, Baek KH, Park WC, Lee YS, Choi JE, Lee JM, Kang MI, Cha BY, Son HY, Lee KW, Kang SK. Diagnostic value of ultrasonography to distinguish between benign and malignant lesions in the management of thyroid nodules. Thyroid 2007;17: 461–466. Xu HX, Yin XY, Lu MD, Xie XY, Xu ZF, Liu GJ. Usefulness of threedimensional sonography in procedures of ablation for liver cancers: Initial experience. J Ultrasound Med 2003;22:1239–1247. Ying M, Yung DM, Ho KK. Two-dimensional ultrasound measurement of thyroid gland volume: A new equation with higher correlation with 3-D ultrasound measurement. Ultrasound Med Biol 2008;34: 56–63.
http://dx.doi.org/10.1016/j.ultrasmedbio.2015.06.022
d
Original Contribution COMPARISON BETWEEN THIN-SLICE 3-D VOLUMETRIC ULTRASOUND AND CONVENTIONAL ULTRASOUND IN THE DIFFERENTIATION OF BENIGN AND MALIGNANT THYROID LESIONS WEN-BO LI,* BO ZHANG,* QING-LI ZHU,* YU-XIN JIANG,* JIAN SUN,y MENG YANG,* and JIAN-CHU LI* * Department of Ultrasound, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Dongcheng District, Beijing, China; and y Department of Pathology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Dongcheng District, Beijing, China (Received 26 February 2015; revised 29 May 2015; in final form 25 June 2015)
Abstract—We explored the efficacy of thin-slice volumetric 3-D ultrasound (3-DUS) in distinguishing between benign and malignant thyroid nodules. A total of 103 thyroid nodules were evaluated prospectively using 3-D gray-scale ultrasonography. The shape, margin, halo and potential capsular invasion of the nodules were compared with the findings of conventional 2-D ultrasound (2-DUS). Of the 103 thyroid nodules, there were 50 pathologically confirmed benign lesions and 53 malignant lesions (51.5%). Shape irregularity, ill-defined margins and capsular invasion provided sensitivities of 90.0%, 47.2% and 39.6% and specificities of 88.0%, 84.0% and 100%, respectively, for the malignant lesions. The diagnosis of thyroid cancer was improved in 3-DUS compared with 2-DUS, with a sensitivity of 88.7%, specificity of 90.0%, positive predictive value of 90.4%, negative predictive value of 88.2% and accuracy of 89.3%. The sensitivity of detection for lesions with capsular invasion increased to 39.6% with 3-DUS, more than twice that of 2-DUS. Three-dimensional US is highly accurate in diagnosing thyroid nodules, particularly those with capsular invasion. (E-mail: [email protected]) Ó 2015 World Federation for Ultrasound in Medicine & Biology. Key Words: Thyroid nodule, 3-D ultrasound, Diagnosis.
INTRODUCTION
explored the application of 3-DUS for thyroid diseases; most studies have focused on applications for thyroid/ nodule volume measurement (Andermann et al. 2007; Lyshchik et al. 2004; Ying et al. 2008). We analyzed the 3-D volumetric characteristics of thyroid nodules by performing multiplanar reconstruction of the 3-D volumetric data of the thyroid nodules and explored the diagnostic value of 3-DUS in the differential diagnosis of these lesions by comparing the lesion features determined in this manner with those from 2-D ultrasound (2-DUS).
Ultrasound has an accuracy of 67%–92% in the diagnosis of thyroid nodules (Moon et al. 2007; Stacul et al. 2007; Tae et al. 2007) and remains one of the most frequently used approaches for imaging these lesions. However, ultrasound suffers certain limitations: (i) conventional static images sometimes do not completely reflect the structures of the thyroid nodules; (ii) the procedure is highly dependent on the operator(s); and (iii) systematic reviews and analyses are often not available. By contrast, 3-D ultrasound (3-DUS) can be used to automatically acquire a series of 2-D images containing spatial data. Three-dimensional US avoids the limitations of gray-scale ultrasound by displaying the lesion features via reconstruction. However, few studies to date have
METHODS Patients The experimental protocol was approved by a local institutional review board, and informed consent for the study was obtained from all human patients. Ninety-six consecutive patients with complete pre-operative 3DUS data who underwent surgical operations for thyroid nodules from March 2008 to April 2009 were prospectively enrolled in this study. Two patients withdrew
Address correspondence to: Yu-Xin Jiang, Department of Ultrasound, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences, Shuaifuyuan 1, Dongcheng District, Beijing 100730, China. E-mail: [email protected] 3096
Thin-slice 3-D US vs. conventional US d W.-B. LI et al.
from the study before surgery; therefore, 103 thyroid nodules in 94 patients were included in the final analysis. Clinical physicians make surgical decisions based on (i) suspicion of malignancy or malignancy diagnosed on fine-needle aspiration (FNA) (n 5 12, 12.8%); (ii) high suspicion of malignancy on US without FNA (n 5 42, 44.7%); (iii) suspected recurrence after resection of thyroid papillary carcinoma (n 5 2, 2.1%); (iv) multiple endocrine neoplasms (n 5 2, 2.1%); (v) metastatic cancer from the thyroid confirmed by pathology (n 5 2, 2.1%); (vi) hyperthyroidism with thyroid nodules (n 5 3, 3.2%); and (vii) retrosternal goiter or compression of the trachea (n 5 31, 33.0%). Of the 103 investigated nodules, 18 (17.5%) were from male patients and 85 (82.5%) were from female patients. The mean age of these patients was 43.6 6 12.2 y (range: 19–71 y). Machine and examinations Sonographic examinations were performed with a Voluson E8 (GE Healthcare, Milwaukee, WI, USA) with an 11 L-D high frequency probe with a frequency range of 6–15 MHz for 2-DUS and a RSP6-16-D 3-D volumetric probe for 3-DUS. The default conditions for thyroid examinations were applied. During examination, the patient was in the supine position, and the anterior area of the neck was fully exposed. Two-dimensional US was performed first, and during this procedure, sagittal, coronal and oblique gray-scale images of the thyroid and its surrounding tissues and lymph nodes were obtained. The dynamic range, depth, gain and focus were adjusted to obtain the optimal image. The shape, border, margin, halo and potential capsular invasion of the thyroid nodules were recorded. Three-dimensional US was then initiated. The patient was asked to hold his or her breath and to avoid swallowing. Three-dimensional volumetric US was performed using the built-in 3-D grayscale mode with the RSP6-16-D 3-D volumetric probe. The 3-D sample frame was set to include the index lesion and the maximum amount of adjacent tissues, approximately 3–4 cm in depth and 4 cm in width. The sweep angle was 15–29 based on the nodule size. Then the probe was stabilized and initial automatic volume data were acquired. The digital images acquired with conventional 2-DUS and 3-DUS were transferred to a personal computer and analyzed using 4-D VIEW PC software (Version 6.2), which provides the same interface as the scanner. Image analysis Volumetric data were analyzed using the VOCAL (virtual organ computer-aided analysis) imaging program (GE Healthcare, Kretz Ultrasound, Zipf, Austria), which enables the automatic calculation of volumetric data. The ‘‘manual mode’’ was applied in our study. The image was
3097
rotated 30 for each analysis, and the profile of each nodule was drawn in six planes to obtain the shape of the nodule. Then, the size of each nodule was obtained by drawing the nodule twice and calculating the average value. Two radiologists with at least 5 y of experience in thyroid sonography, who were blinded to the clinical conditions and pathologic findings of the patients, independently evaluated the 2-DUS and 3-DUS features of the lesions and then summarized the diagnostic results. If the two radiologists did not agree on the diagnosis, the result was discussed until agreement was reached. Multiplanar imaging analysis was applied to the 3-D images, during which three perpendicular sections were used—a horizontal section (plate A), a longitudinal section (plate B) and a coronal section (plate C)—along with a spatial 3-D image of coronal section obtained after these three sections were combined. The various planes could display 3-D images from different sections and in different directions. To evaluate nodule features, a 100% surface smooth algorithm was chosen for volume rendering of 3-D data. The thickness of the slice was adjusted according to lesion size empirically, approximately 15%–25% of the maximal diameter of the nodule. The features of the lesions observed via 2-DUS and 3-DUS included shape (regular or irregular) (Fig. 1), margin (clear or unclear), halo (complete and regular, irregular/incomplete or none) (Fig. 2) and capsular invasion (distant, adjacent or invaded) (Fig. 3). The halo features were defined as follows: a regular slim hypoechoic rim around the thyroid nodule (complete and regular halo); a hypoechoic rim around the thyroid nodule, with varied width or with interruptions (irregular/incomplete halo); or no obvious hypoechoic rim visible around the thyroid nodule (absence of halo). The relationship between the nodule and the capsule was defined as follows: distant 5 thyroid gland tissue visible between the nodule and the capsule; adjacent 5 an unbroken capsule despite the absence of thyroid gland tissue between the nodule and the capsule; capsular invasion 5 continuously interrupted capsule. Pathologic diagnosis Two pathologists with at least 5 y of experience in thyroid pathology independently issued their diagnoses after observing the sections of the surgically removed thyroid lesions. These pathologists were also blinded to the ultrasound results at diagnosis. If the two pathologists did not agree on the diagnosis, the result was further discussed until agreement was reached. Statistical analysis The statistical analysis was performed using SPSS 11.0 software (SPSS, Chicago, IL, USA). Normally distributed data are expressed as the mean 6 standard
3098
Ultrasound in Medicine and Biology
Volume 41, Number 12, 2015
Fig. 1. Papillary thyroid carcinoma in a 34-y-old woman. (a) Two-dimensional gray-scale ultrasonography indicates a hypo-echoic thyroid nodule (maximal diameter 14 mm) with a regular shape, clear margin and calcifications (hollow arrows). (b) Multiplanar reconstruction (MPR) and thin-slice smooth-surface rendering 3-D gray-scale ultrasonography indicate nodule with irregular lobulations (arrows), which were invisible on corresponding 2-D image.
deviation (x 6 SD) and were analyzed using the t-test. Data that were not normally distributed are expressed as the range and were analyzed using the rank-sum test. Count data were analyzed using the c2-test and Fisher’s exact probability test. The level of agreement on the 3D final diagnosis was evaluated by two radiologists. Inter-observer variability was measured using the k index of agreement. p , 0.05 was regarded as the threshold for statistical significance.
RESULTS
group was aged 41.9 6 13.5 y (range: 19–71 y) (p 5 0.136). Of the 103 nodules, 18 were from male patients and 85 were from female patients. Benign lesions had an average size of 5.65 cm3 (range: 0.14– 39.78 cm3), significantly greater than that of the malignant lesions (mean: 4.00 cm3, range: 0.29–18.65 cm3) (p 5 0.048). These 53 malignant nodules included papillary thyroid carcinoma (n 5 48), follicular carcinoma (n 5 1), medullary carcinoma (n 5 3) and lymphoma (n 5 1). The 50 benign nodules included nodular goiter (n 5 40), thyroid adenoma (n 5 8), Hashimoto’s thyroiditis (n 5 1) and fibrous thyroiditis (n 5 1).
General data and pathologic types of thyroid nodules Among the 103 thyroid nodules, there were 50 pathologically confirmed benign lesions and 53 malignant lesions (51.5%). The mean age of the patients was 43.6 6 12.2 y (range: 19–71 y). The benign group was aged 45.5 6 10.6 y (range: 20–68 y), and the malignant
Ultrasonographic features and diagnostic values of 2DUS and 3-DUS for thyroid nodules The ultrasonographic features identified via 2-DUS and 3-DUS for the thyroid nodules are listed in Table 1. For both 2-DUS and 3-DUS, the features ‘‘capsular
Fig. 2. Nodular goiter in a 35-y-old woman. (a) Two-dimensional gray-scale ultrasonography indicates a cystic-solid thyroid nodule (maxmal diameter 26 mm) with regular shape and a clear margin, but no halo (hollow arrows). (b) MPR and thin-slice smooth-surface rendering 3-D gray-scale ultrasonography reveal a complete and regular halo (arrows) prominently. MPR 5 multiplanar reconstruction.
Thin-slice 3-D US vs. conventional US d W.-B. LI et al.
3099
Fig. 3. Papillary thyroid carcinoma in a 69-y-old woman. (a) Two-dimensional gray-scale ultrasonography indicates a typical malignant thyroid nodule (maximal diameter 14 mm), but without evidence of capsular invasion (hollow arrows). (b) MPR and thin-slice smooth-surface rendering 3-D gray-scale ultrasonography reveal that the hyper-echoic capsule was interrupted by hypo-echoic tumor invasion (arrow). MPR 5 multiplanar reconstruction.
invasion,’’ ‘‘irregular shape’’ and ‘‘unclear margin’’ were used in the diagnosis of thyroid cancer. Associated sensitivity, specificity, positive predictive value, negative predictive value and accuracy are summarized in Table 2. The ‘‘irregular shape’’ feature of 3-DUS had the highest sensitivity and specificity in diagnosing thyroid cancer. When the ‘‘complete and regular halo’’ feature of 3DUS was used for the diagnosis of benign thyroid nodules, its sensitivity was 54.0%, specificity 89.1%, positive predictive value 96.4%, negative predictive value 69.3% and accuracy 76.7%. For 3-DUS, the sensitivity was 88.7%, specificity 90.0%, positive predictive value 90.4%, negative predictive value 88.2% and accuracy 89.3%; for 2-DUS, the corresponding measures were 86.8%, 86.0%, 86.8%, 86.0% and 86.4%, respectively. The k index for agree-
Table 1. Ultrasonographic features of 2-DUS and 3-DUS for thyroid nodules 3-DUS Ultrasonographic feature Shape Regular Irregular Capsule Distant Adjacent Invasion Halo Regular and complete Irregular/incomplete None Margin Clear Unclear
ment between two observers was 0.85 for the 3-D grayscale ultrasonographic diagnoses. DISCUSSION Baum and Greenwood first described 3-D ultrasound imaging in 1961 (Baum and Greenwood 1961). By scanning and acquiring a series of 2-D ultrasonographic images with spatial information, 3-DUS can be used for 3-D reconstruction and display. The past two decades have witnessed the rapid development of electronic technology and image processing technology, accompanied by the wide application of 3-DUS in the clinical management of obstetric and gynecologic problems, heart diseases, liver diseases and prostate disorders (Alcazar and Galvan 2009; Armstrong et al. 2013; Correale et al. 2008; Hotta et al. 2007; Mehta et al. 2004; Pretorius et al. 1998; Xu et al. 2003). However, few studies have reported the application of 3-DUS in the management of thyroid disorders.
2-DUS
Table 2. Diagnostic values of 2-DUS and 3-DUS for thyroid nodules
Benign
Malignant
Benign
Malignant
44 6
5 48
45 5
7 46
Ultrasonographic Sensitivity Specificity feature (%) (%)
5 45 0
2 30 21
5 45 0
2 43 8
27 14 9
1 19 33
25 11 14
1 11 41
42 8
28 25
39 11
18 35
Irregular shape 3-DUS 2-DUS Unclear margin 3-DUS 2-DUS Capsular invasion 3-DUS 2-DUS
2-DUS (3-DUS) 5 2-D (3-D) ultrasonography.
PPV (%)
NPV Accuracy (%) (%)
90.6 86.8
88.0 90.0
88.9 90.2
89.8 86.5
89.3 88.3
47.2 66.0
84.0 78.0
75.8 76.1
60.0 68.4
65.0 71.8
61.0 52.6
68.9 56.3
39.6 15.1
100 100
100 100
NPV 5 negative predictive value; PPV 5 positive predictive value; 2-DUS (3-DUS) 5 2-D (3-D) ultrasonography.
3100
Ultrasound in Medicine and Biology
Three-dimensional US volumetric data of thyroid nodules can reveal the morphologic features of lesions on multiple planes. Using three mutually perpendicular sections, we can better evaluate the complicated growth patterns and margins of a nodule and, thus, more accurately identify its overall profile and its anatomic relationships with neighboring structures while avoiding excessive dependence on operators and missing information because of non-standard sections (Fernandez et al. 2004; Munoz Agel and Varas Lorenzo 2005). Additionally, volumetric data can be retrieved and reviewed at any time for any suspicious lesion, thereby providing important information supplementary to 2DUS (Fenster et al. 2001). The volumetric data provided by 3-DUS enable repeated observations of the morphologic features of the nodules from multiple planes and various angles. In our study, the ‘‘irregular shape’’ feature had a sensitivity of 90.6% in the diagnosis of malignant thyroid nodules and was thus the most sensitive sonographic feature. As in our study, Slapa et al. (2011) found that ‘‘a lobulated shape of the nodule in the c-plane’’ had a sensitivity of 94%–100% in the diagnosis of thyroid cancer. In the present study, in one case, the nodule was diagnosed as benign (size 4.0 3 2.6 cm) based on 2-DUS; however, 3-DUS indicated that it had an irregular shape and therefore should be diagnosed as malignant. Post-operative pathology revealed a gray solid nodule 0.9 3 0.6 cm in size, which was confirmed to be a papillary carcinoma of the thyroid. One possible reason for such a difference may be that the morphologic changes in some malignant thyroid nodules are relatively slight and only the local shape of a tumor may become irregular. Because of the complexity of tumor shapes, conventional ultrasound may not be able to (fully) display the lesions and thus may lead to misdiagnoses in some cases. By contrast, the reconstruction of volumetric data enables more detailed and accurate visualization of tumors and thus provides more diagnostic information. In three cases, including one benign lesion and two malignant lesions, the shapes of the thyroid nodules were regular on 2DUS but irregular on 3-DUS. Therefore, benign lesions may also have irregular shapes, similar to those of malignant lesions. However, this irregular characterization of the benign lesion may also be attributable to the rater’s experience. Thyroid cancer manifests as multiple irregular shapes; however, as a result of prolonged growth, benign nodules (e.g., a nodular goiter) may exhibit secondary changes such as bleeding, necrosis and calcification, which can appear as irregular changes. Nevertheless, such changes tend to be relatively slight. Three-dimensional US can improve the visualization of the borders/margins of thyroid nodules and thus provide important information for the differential diag-
Volume 41, Number 12, 2015
nosis of benign and malignant lesions. In the present study, 13 lesions without halos on 2-DUS displayed complete or incomplete halos on 3-DUS (an increase of 12.6%). The ‘‘complete and regular halo’’ feature on 3DUS was a key feature in the diagnosis of benign thyroid nodules, with a specificity of 98.1% and a positive predictive value of 96.4%. Compared with 2-DUS, 3-DUS indicated clear margins in a larger number of nodules. This improvement may be related to the current volumetric imaging data processing mode of 3-DUS, which can provide more homogeneous multiplanar imaging echo patterns, decreased noise and enhanced contrast, thereby improving the visualization of nodular margins (Slapa et al. 2006). However, a clear margin is not a feature that is unique to benign nodules. In the present study, 84% (42/50) of the benign nodules had clear margins on 3-DUS; however, 53% (28/53) of the malignant nodules (including 24 cases of papillary carcinoma, 1 case of follicular carcinoma, 2 cases of medullary carcinoma and 1 case of lymphoma) also displayed clear margins. Malignant thyroid nodules typically have blurred and rough margins, reflecting the infiltration of the malignant tumor cells into the surrounding tissues. However, such infiltration may not be obvious in some malignant nodules; alternatively, fibrous capsules may form. As a result, the nodular margin may appear to be clear on ultrasound. Three-dimensional US can clearly visualize nodular margins and thus allow for accurate evaluation of the relationship between the nodule and capsule. ‘‘Capsular invasion’’ is a key feature for the diagnosis of thyroid cancer using ultrasound. ‘‘Capsular invasion’’ is highly specific and is regarded as an important factor affecting determination of the stage and prognosis of thyroid cancer, as well as the selection of the appropriate surgical procedure (Cooper et al. 2009). However, 2-DUS offers suboptimal sensitivity in the detection of capsular invasion, particularly when this feature is affected by multiple reflection artifacts in the ultrasonic near field (Nelson et al. 2000). As illustrated in our study, the ‘‘capsular invasion’’ feature in 3-DUS exhibited a sensitivity of 39.6% in diagnosing thyroid cancer, which was significantly higher than the corresponding sensitivity in 2-DUS (15.1%). Thus, clear visualization of the relationship between the nodule and capsule using 3-DUS is beneficial in improving the diagnostic accuracy for thyroid cancer and thus can play a role in the staging and prognosis of thyroid cancer. Three-dimensional ultrasound does, however, have certain limitations. For example, some information will be unavoidably missed during 3-D reconstruction, and 3-D artifacts may be produced as a result of involuntary movement or the beating of the carotid artery during data acquisition (Nelson et al. 2000). However, the image quality does not affect the image analysis or diagnosis.
Thin-slice 3-D US vs. conventional US d W.-B. LI et al.
This was a preliminary study of the role of 3-D grayscale ultrasound in the diagnosis of thyroid nodules. On the basis of our limited sample size, the majority of malignant thyroid nodules were found to be papillary carcinoma of the thyroid, and most benign nodules were nodular goiter. Only a small proportion of patients had other pathologic types. However, studies with larger sample sizes are warranted to validate our findings. Acknowledgments—This work is supported by the National Natural Science Foundation of China (81171354).
REFERENCES Alcazar JL, Galvan R. Three-dimensional power Doppler ultrasound scanning for the prediction of endometrial cancer in women with postmenopausal bleeding and thickened endometrium. Am J Obstet Gynecol 2009;200:44.e1–44.e6. Andermann P, Schlogl S, Mader U, Luster M, Lassmann M, Reiners C. Intra- and interobserver variability of thyroid volume measurements in healthy adults by 2-D versus 3-D ultrasound. Nuklearmedizin 2007;46:1–7. Armstrong L, Fleischer A, Andreotti R. Three-dimensional volumetric sonography in gynecology: An overview of clinical applications. Radiol Clin North Am 2013;51:1035–1047. Baum G, Greenwood I. A critique of time–amplitude ultrasonography. Arch Ophthalmol 1961;65:353–365. Cooper DS, Doherty GM, Haugen BR, Kloos RT, Lee SL, Mandel SJ, Mazzaferri EL, McIver B, Pacini F, Schlumberger M, Sherman SI, Steward DL, Tuttle RM. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid 2009;19:1167–1214. Correale M, Ieva R, Di Biase M. Real-time three-dimensional echocardiography: An update. Eur J Intern Med 2008;19:241–248. Fenster A, Downey DB, Cardinal HN. Three-dimensional ultrasound imaging. Phys Med Biol 2001;46:R67–R99. Fernandez LJ, Aguilar A, Pardi S. Three-dimensional ultrasound in small parts: Is it just a nice picture? Ultrasound Q 2004;20: 119–125. Hotta N, Ayada M, Okumura A, Ishikawa T, Kakumu S. Usefulness of live 3-D echocardiography during radiofrequency ablation in a
3101
case of hepatocellular carcinoma. Clin Imaging 2007;31: 283–286. Lyshchik A, Drozd V, Reiners C. Accuracy of three-dimensional ultrasound for thyroid volume measurement in children and adolescents. Thyroid 2004;14:113–120. Mehta SS, Azzouzi AR, Hamdy FC. Three dimensional ultrasound and prostate cancer. World J Urol 2004;22:339–345. Moon HG, Jung EJ, Park ST, Ha WS, Choi SK, Hong SC, Lee YJ, Joo YT, Jeong CY, Choi DS, Ryoo JW. Role of ultrasonography in predicting malignancy in patients with thyroid nodules. World J Surg 2007;31:1410–1416. Munoz Agel F, Varas Lorenzo MJ. Tridimensional (3-D) ultrasonography. Rev Esp Enferm Dig 2005;97:125–134. Nelson TR, Pretorius DH, Hull A, Riccabona M, Sklansky MS, James G. Sources and impact of artifacts on clinical threedimensional ultrasound imaging. Ultrasound Obstet Gynecol 2000;16:374–383. Pretorius DH, Nelson TR, Baergen RN, Pai E, Cantrell C. Imaging of placental vasculature using three-dimensional ultrasound and color power Doppler: A preliminary study. Ultrasound Obstet Gynecol 1998;12:45–49. Slapa RZ, Jakubowski WS, Slowinska-Srzednicka J, Szopinski KT. Advantages and disadvantages of 3-D ultrasound of thyroid nodules including thin slice volume rendering. Thyroid Res 2011;4:1. Slapa RZ, Slowinska-Srzednicka J, Szopinski KT, Jakubowski W. Grayscale three-dimensional sonography of thyroid nodules: Feasibility of the method and preliminary studies. Eur Radiol 2006;16: 428–436. Stacul F, Bertolotto M, De Gobbis F, Calderan L, Cioffi V, Romano A, Zanconati F, Cova MA. US, colour-Doppler US and fine-needle aspiration biopsy in the diagnosis of thyroid nodules. Radiol Med 2007;112:751–762. Tae HJ, Lim DJ, Baek KH, Park WC, Lee YS, Choi JE, Lee JM, Kang MI, Cha BY, Son HY, Lee KW, Kang SK. Diagnostic value of ultrasonography to distinguish between benign and malignant lesions in the management of thyroid nodules. Thyroid 2007;17: 461–466. Xu HX, Yin XY, Lu MD, Xie XY, Xu ZF, Liu GJ. Usefulness of threedimensional sonography in procedures of ablation for liver cancers: Initial experience. J Ultrasound Med 2003;22:1239–1247. Ying M, Yung DM, Ho KK. Two-dimensional ultrasound measurement of thyroid gland volume: A new equation with higher correlation with 3-D ultrasound measurement. Ultrasound Med Biol 2008;34: 56–63.