Compressibility of Thyroid Masses: A Sonographic Sign Differentiating Benign From Malignant Lesions?

Transcription

1 Neuroradiology/Head and Neck Imaging Original Research Seo et al. Ultrasound of Thyroid Masses Neuroradiology/Head and Neck Imaging Original Research Young Lan Seo 1,2 Dae Young Yoon 1,2 Soo Jeong Yoon 1 Kyoung Ja Lim 1 Eun Joo Yun 1 Chul Soon Choi 1 Sang Hoon Bae 1 Seo YL, Yoon DY, Yoon SJ, et al. Keywords: compressibility, sonography, thyroid cancer, thyroid nodule DOI: /AJR Received January 5, 2011; accepted after revision June 1, Department of Radiology, Kangdong Seong-Sim Hospital, Hallym University College of Medicine, Seoul, Republic of Korea. 2 Department of Radiology, Ilsong Memorial Institute of Head and Neck Cancer, Kangdong Seong-Sim Hospital, Hallym University College of Medicine, 445 Gil-dong Kangdong-Gu, Seoul, , Republic of Korea. Address correspondence to D. Y. Yoon (evee0914@chollian.net). AJR 2012; 198: X/12/ American Roentgen Ray Society Compressibility of Thyroid Masses: A Sonographic Sign Differentiating Benign From Malignant Lesions? OBJECTIVE. The purpose of this study was to assess the manual compressibility of thyroid masses with an ultrasound probe and to determine whether this ultrasound feature can be used to differentiate benign from malignant thyroid lesions. SUBJECTS AND METHODS. We prospectively compared images obtained during compression with an ultrasound probe and noncompressed ultrasound images of 180 pathologically proven thyroid masses (51 malignant, 129 benign) smaller than 2 cm in 169 patients (127 women, 42 men; mean age, 51.2 years). The size (anteroposterior and transverse dimensions) and shape (ratio of anteroposterior to transverse dimension) of the selected lesions were measured on both noncompressed and compressed ultrasound images at a computer workstation, and the compressibility (anteroposterior-to-transverse ratio on noncompressed images minus anteroposterior-to-transverse ratio on compressed images) was calculated. Compressibility was analyzed to determine its association with histopathologic results (benign versus malignant) and the characteristics of the thyroid mass (involved lobe, location in lobe, halo, and composition). The area under the receiver operating characteristic curve was used as an indicator of performance. RESULTS. The mean anteroposterior-to-transverse ratio of a thyroid mass on compressed ultrasound images was significantly lower than that on noncompressed images (0.78 ± 0.28 vs 0.92 ± 0.30; p < 0.001). The compressibility of masses was greater for benign than for malignant lesions (0.19 ± 0.16 vs 0.05 ± 0.12; p < 0.001). No statistically significant association was identified between compressibility and other characteristics of a lesion. The area under the receiver operating characteristic curve for compressibility of thyroid masses was On the basis of a cutoff value for malignancy of compressibility less than 0.10, the sensitivity, specificity, and accuracy were 72.5%, 72.9%, and 72.8%. CONCLUSION. Compressibility with an ultrasound probe is a useful criterion for differentiating benign from malignant lesions of the thyroid. S onography is widely accepted as the modality of choice for evaluating thyroid masses because of its noninvasive nature, accessibility, and high spatial resolution. In addition to its use in detecting thyroid masses, ultrasound is useful in differentiating the benign or malignant nature of the mass. Previous studies [1 4] have identified several features predictive of malignancy, including the presence of microcalcifications, hypoechogenicity, irregular margins, absence of a halo, predominantly solid composition, and intranodular vascularity. In 2002, Kim et al. [1] described a new ultrasound sign of thyroid malignancy. They observed that an anteroposterior-to-transverse diameter ratio greater than 1 of a nodule was highly predictive of malignancy and called this finding a shape that was taller than wide. Thereafter, a number of studies [4 7] confirmed that a thyroid lesion with an anteroposterior-to-transverse ratio 1 is a reliable ultrasound criterion for malignancy. More recently, Yoon et al. [8] compared the differences in shape of thyroid masses seen on both ultrasound and CT images and suggested that the mechanism of the tallerthan-wide sign is that compared with benign masses, malignant lesions exhibit no or minimal compressibility when an ultrasound probe is applied. Several studies have been conducted to evaluate the usefulness of ultrasound images obtained during compression in the differentiation of breast tumors [9, 10] and musculoskeletal soft-tissue lesions [11]. To our knowledge, 434 AJR:198, February 2012

2 Ultrasound of Thyroid Masses however, no reports on manual compressibility of thyroid masses with an ultrasound probe have been published. The purpose of this prospective study was to evaluate the diagnostic accuracy of manual compressibility as a new ultrasound criterion for differentiation of benign from malignant thyroid masses. Subjects and Methods The entire study protocol was approved by our institutional review board, and informed consent was obtained from all included patients. Patient Sample The subjects recruited for this prospective study were 196 consecutively registered patients with 210 thyroid nodules who underwent both thyroid ultrasound and fine-needle aspiration (FNA) biopsy or surgery. The study was conducted during the 10-month-period January October The indications for FNA in this study were nodule larger than 10 mm, presence of spiculated margins, hypoechoic appearance, and presence of microcalcifications or macrocalcifications. Because we attempted to assess the shape of thyroid masses, only thyroid masses smaller than 2.0 cm in maximum diameter were included. Larger lesions were excluded because they are more likely to invade adjacent structures, which would influence their shape and conspicuity. A total of 30 nodules were excluded for the following reasons: poorly defined border of the mass on ultrasound images (n = 11), diffuse enlargement of thyroid (diffuse or multinodular goiter) (n = 10), and inadequate sample at FNA (n = 9). The final study sample was 169 patients (127 women, 42 men; mean age, 51.2 ± 12.4 years; range, years) with 180 thyroid masses (a single mass in 158 patients and two masses [one in each lobe] A Fig year-old woman with benign thyroid lesion (adenomatous goiter). A = common carotid artery, V = internal jugular vein. A, Transverse noncompressed ultrasound image shows 8-mm isoechoic nodule (arrow) with hypoechoic rim in right lobe of thyroid. Anteroposterior-to-transverse diameter ratio of nodule is Artifact (arrowheads) is present at both ends of image owing to air gap between probe and skin. B, Compressed ultrasound image shows nodule (arrow) that is taller than wide in A is transformed into lesion that is wider than tall. Anteroposterior-to-transverse diameter ratio changes from 1.05 to 0.59 (compressibility, 0.46). Internal jugular vein collapses with application of pressure by ultrasound probe. in 11 patients). In patients with multiple thyroid masses, the largest mass in each lobe was chosen for evaluation. In all cases, the final diagnosis of benign (n = 129) or malignant (n = 51) mass was determined pathologically in specimens obtained by FNA (n = 126), thyroidectomy (n = 8), or both (n = 46). All malignant thyroid tumors were found to be papillary carcinoma at pathologic examination. Ultrasound Examination Ultrasound examinations were performed by two radiologists with more than 5 years of experience in thyroid ultrasound and ultrasound-guided FNA (HDI 5000 or iu 22 ultrasound unit, Philips Healthcare; or Acuson Sequoia 512 unit, Siemens Healthcare) with an 8 15 MHz linear array transducer. For noncompressed ultrasound imaging, the probe was placed lightly on the anterior aspect of the neck above the examined lesion with a large amount of ultrasound transmission gel (Supersonic, Sung Heung Medical) to create a standoff pad. Absence of external pressure was proved by the presence of noise at both ends of the probe due to the presence of an air gap between the probe and the skin (Figs. 1 and 2). For compression ultrasound images, pressure was applied to the same area with the transducer, stopping when the patient reported discomfort. Only static sonograms were obtained in this series; cine clips were not made. Data Collection and Analysis The ultrasound images were independently evaluated at a PACS workstation (π-viewstar, Infinitt) by two radiologists. All images were presented randomly in a blinded manner. Each examination was allocated a study number known only to the study coordinator. The readers were not informed of the measurements made by the other investigator but were informed about the location of the mass evaluated with both compressed and noncompressed ultrasound in patients with multiple thyroid masses. Each reader chose representative transverse compressed and noncompressed ultrasound images of each thyroid mass so that the maximal area of the mass appeared on images. The readers independently assessed the following parameters on both compressed and noncompressed ultrasound images: anteroposterior dimension, transverse dimension, and ratio of anteroposterior to transverse dimension. All quantitative measurements were performed at the workstation with electronic calipers after ap propriate magnification. The anteroposterior and transverse diameters of the lesions were mea sured once by each reader, and the mean value was calculated from the two independent measurements. Anteroposterior dimension of the mass was defined as the diameter in the axis perpendicular to the anterior surface of the thyroid. Transverse dimension was defined as the diameter in the axis perpendicular to the diameter used for measurement of the anteroposterior dimension. The reviewers also assessed the following characteristics of thyroid masses on compressed ultrasound images: location in the gland (right lobe, left lobe, isthmus), location in right or left lobe (upper third, middle third, lower; anterior third, middle third, posterior third), peripheral halo (present, absent), and lesion composition (cystic, cystic portion > 75%; mixed, cystic portion, 25 75%; solid, cystic portion < 25%). If there were disagreements in qualitative assessment, the reviewers discussed them and reached final decisions by consensus. We compared the anteroposterior and transverse dimensions and anteroposterior-to-transverse ratio of thyroid masses between compressed and noncompressed ultrasound images. In addition, the compressibility of each lesion (anteroposterior-totransverse ratio on noncompressed images minus B AJR:198, February

3 Seo et al. TABLE 1: Size and Shape of Thyroid Masses: Differences on Compressed and Noncompressed Ultrasound Images Parameter Compressed Ultrasound Noncompressed Ultrasound p Anteroposterior dimension (mm) 7.62 ± ± 3.56 < Transverse dimension (mm) ± ± 4.00 < Ratio of anteroposterior to transverse dimension 0.78 ± ± 0.30 < Note Data are mean ± SD. the ratio on compressed images) was calculated to determine the association between compressibility and histopathologic result (benign vs malignant) and the various gray-scale characteristics (location in the gland, location in the lobe, halo, and composition). The performance of ultrasound compressibility as a means of differentiating benign and malignant thyroid lesions was evaluated with receiver operating characteristic (ROC) curve analysis. Area under the ROC curve was used as an indicator of performance. Using ROC curves, we identified an optimal compressibility cutoff value of 0.10 for differentiating benign from malignant masses. Statistical analysis was performed with the paired Student t test or one-way analysis of variance test for continuous variables and the chisquare test for categoric variables. A value of p < 0.05 was considered statistically significant. ROC curve analysis was conducted with Stata/SE 11.0 software (StatCorp). Statistical analyses other than ROC analysis were performed with SPSS version 12.0 for Microsoft Windows software (SPSS). Results Table 1 shows the mean anteroposterior and transverse dimensions and anteroposterior-to-transverse diameter ratios of all thyroid masses measured on both compressed and noncompressed ultrasound images. The A Fig year-old woman with malignant thyroid lesion (papillary carcinoma). A = common carotid artery, V = internal jugular vein. A, Transverse noncompressed ultrasound image shows 6-mm hypoechoic nodule (large arrow) with microcalcifications in left lobe of thyroid. Anteroposterior-totransverse diameter ratio of nodule is Artifact from air gap (arrowheads) is evident. Small arrow indicates small intrathyroid vessel. B, Compressed ultrasound image shows anteroposterior-to-transverse diameter ratio (1.24) of nodule (arrow) changes little with probe compression (compressibility, 0.05). Internal jugular vein collapses with probe compression. Small intrathyroid vessel has collapsed. transverse dimension of the thyroid masses on compressed ultrasound images was significantly larger than that on noncompressed images. The anteroposterior dimension and anteroposterior-to-transverse ratio of the thyroid masses on compressed ultrasound images were significantly smaller than those on noncompressed images. In addition, the anteroposterior-to-transverse ratio of 33 thyroid masses decreased from greater than 1.0 (taller) to 1.0 or less (wide) with probe compression (Fig. 1). The mean compressibility of all the thyroid masses was 0.15 ± The compressibility of benign masses (0.19 ± 0.16) was significantly greater than that of malignant masses (0.05 ± 0.12; p < 0.001) (Figs. 1 and 2). In contrast, none of the ultrasound characteristics of the thyroid mass (location, location in lobe, halo, composition) significantly correlated with compressibility (Table 2). The ROC curve for differentiating benign from malignant thyroid lesions with compressibility is shown in Figure 3. The area under the curve of thyroid mass compressibility was 0.78 (95% CI, ). With a calculated compressibility value less than 0.10 considered suspicious for malignancy, the sensitivity, specificity, positive and negative predictive values, and accuracy were 72.5% (37/51), 72.9% (94/129), 51.4% (37/72), 87.0% (94/108), and 72.8% (131/180). Discussion After the first description of the taller-thanwide sign by Kim et al. [1], several studies of ultrasound showed that the sign is useful for differentiating malignant from benign masses, either as the sole criterion or in combination with other ultrasound features [4 7]. In those studies [1, 4 7], the presence of the taller-than-wide sign was highly specific ( %) but not sensitive ( %). Few studies have addressed the exact cause of the taller-than-wide sign. Kim et al. [1] postulated that benign nodules grow in normal tissue planes, whereas malignant nodules (taller than wide) grow across normal tissue planes. This hypothesis is similar to those in breast ultrasound studies in which it was found that malignant breast masses are more likely to be taller than wide [12 15]. To our knowledge, however, there is no literature supporting the presence of such tissue planes in the thyroid. Yoon et al. [8] compared the shape of thyroid masses evident on both ultrasound and CT images of 77 patients with 90 pathologically proven thyroid masses. In that study, the anteroposterior dimension of the thyroid measured with B 436 AJR:198, February 2012

4 Ultrasound of Thyroid Masses TABLE 2: Compressibility According to Characteristics and Pathologic Result of Thyroid Mass Ratio of Anteroposterior to Transverse Dimension Characteristic Compressed Ultrasound Noncompressed Ultrasound Compressibility Location Right lobe (n = 91) 0.85 ± ± ± 0.19 Isthmus (n = 7) 0.53 ± ± ± 0.04 Left lobe (n = 82) 0.72 ± ± ± 0.13 Anteroposterior location in lobe Anterior third (n = 65) 0.72 ± ± ± Middle third (n = 63) 0.78 ± ± ± 0.18 Posterior third (n = 45) 0.89 ± ± ± 0.19 Superoinferior location in lobe Upper third (n = 38) 0.85 ± ± ± 0.18 Middle third (n = 74) 0.79 ± ± ± 0.16 Lower third (n = 61) 0.75 ± ± ± 0.17 Peripheral halo Present (n = 59) 0.68 ± ± ± 0.16 Absent (n = 121) 0.82 ± ± ± 0.16 Composition Cystic (n = 12) 0.74 ± ± ± 0.15 Mixed (n = 37) 0.64 ± ± ± 0.14 Solid (n = 131) 0.82 ± ± ± 0.17 Pathologic result < Benign (n = 129) 0.69 ± ± ± 0.16 Malignant (n = 51) 0.99 ± ± ± 0.12 Note Data are the mean ± SD. ultrasound was significantly lower than that recorded with CT, suggesting the pressure of the transducer during ultrasound examination compressed the lesion in the anteroposterior dimension. In addition, the study showed that the ratio of anteroposterior to transverse dimension of the thyroid masses measured with ultrasound was significantly lower than that recorded with CT. The differences were significantly greater for benign than for malignant masses. On the basis of the results, the investigators hypothesized that the mechanism of the tallerthan-wide sign is that malignant lesions exhibit no or minimal compressibility from the transducer during ultrasound examination. To assess the ultrasound compressibility of thyroid masses, we directly compared the anteroposterior-to-transverse ratios of thyroid masses measured on compressed and noncompressed ultrasound images. The results of our study suggest that pressure during ultrasound examination of a thyroid can compress a thyroid mass. One might hypothesize that compared with malignant tumors, benign tumors have greater compressibility because benign tumors generally tend to be softer and have less infiltration to surrounding tissue, whereas malignant tumors tend to be more rigid and have more infiltration into surrounding tissue. The foregoing hypothesis can be applied to cystic masses. In general, cystic masses are softer and therefore may be more compressible than mixed or solid masses. In addition, thyroid masses in the isthmus may be less susceptible to compression than those in the right or left lobe, probably because patients are less tolerant of compression directly over the trachea. These trends did not reach statistical significance in this study. Further studies on a larger scale are needed for detailed evaluation of the correlation between compressibility and the characteristics of thyroid masses. The characteristics noted on gray-scale compressed ultrasound images had a sensitivity of 51.4% and specificity of 87.0% for the diagnosis of malignancy. These values are similar to those reported in previous studies [1, 4 7]. Lesion shape, however, differed between compressed and noncompressed ultrasound images in our series, altered by probe compression. We believe that the shape of a thyroid mass at ultrasound examination must be interpreted with consideration to the degree of probe compression and the consistency of the mass (in other words, the compressibility). Ultrasound elastography depicts tissue stiffness by depicting tissue strain from external compression. With this method, strain images (elastograms) are produced by measurement of the local displacement induced by ultrasound probe compression [16, 17]. Although it is not yet routinely used in clinical practice, elastography has been found useful in the differential diagnosis of breast and prostate tumors [18, 19] and thyroid tumors [20 22]. Many studies [20 22] have shown that ultrasound elastography is useful in the differential diagnosis of thyroid nodules, but the studies have been conducted with different criteria. Lyshchik et al. [20] found that a strain index greater than 4 was strongly associated with thyroid cancer, having 96% specificity and 82% sensitivity. Using the subjective scoring of p AJR:198, February

5 Seo et al. Sensitivity Specificity Fig. 3 Graph shows receiver operating characteristic curve for differentiating benign (n = 129) and malignant (n = 51) thyroid lesions. Area under receiver operating characteristic curve is 0.78 (95% CI, ). tissue stiffness with elastography, Rago et al. [21] found sensitivity of 97% and specificity of 100% for differentiation of thyroid masses. Dighe et al. [22] reported that a critical value of an 18.0 thyroid stiffness index corresponds to sensitivity of 87.8% and specificity of 77.5% for prediction of papillary carcinoma. Although ultrasound elastography is useful for evaluating thyroid masses, there are several limitations to this technique. First, the diagnostic performance of elastography is directly related to the skill of the sonographer or sonologist performing the study and to the specifications of the image reconstruction algorithm used [21]. Out-of-plane motion of the lesion examined and artifacts caused by carotid arterial pulsation during external compression can also interfere with elastographic assessment [21, 22]. In contrast, manual ultrasound compressibility imaging (compressed and noncompressed ultrasound images) is easily performed during routine thyroid ultrasound, requiring no more than 1 2 minutes of additional examination time, and can be performed without the machine upgrades and software programs necessary for elastography. There were several limitations to our study. First, the amount of pressure applied by the ultrasound transducer was subjective (not objectively quantified) in this study. Therefore, the accuracy of our method may depend on the degree of probe compression. Second, the inclusion of only patients who had undergone ultrasound and subsequent FNA biopsy or surgery might have introduced bias because patients with obviously benign findings at ultrasound usually did not undergo FNA. Therefore, a substantial number of benignappearing thyroid masses were excluded. Finally, most benign conditions were diagnosed only on the basis of FNA biopsy result. Patients with benign FNA results did not undergo follow-up ultrasound evaluation. Because ultrasound-guided FNA can be associated with a 5% false-negative diagnosis rate, some cases of cancer might have been misclassified as benign lesions in our study. Conclusion The results of this study indicate that the compressibility of a thyroid mass by an ultrasound probe is a useful ultrasound criterion for differentiating benign from malignant lesions. Acknowledgment We thank Young-Su Ju, director, Department of Occupational and Environmental Medicine, Clinical Research Coordinating Center, Hallym University Sacred Heart Hospital, for statistical assistance. 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: 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: 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: 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: Cappelli C, Castellano M, Pirola I, et al. Thyroid nodule shape suggests malignancy. Eur J Endocrinol 2006; 155: Moon WJ, Jung SL, Lee JH, et al. Benign and malignant thyroid nodules: US differentiation multicenter retrospective study. Radiology 2008; 247: 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: Yoon SJ, Yoon DY, Chang SK, et al. Taller-thanwide sign of thyroid malignancy: comparison between ultrasound and CT. AJR 2010; 194:1343 [web]w420 W Park SY, Oh KK, Kim EK, Son EJ, Chung WH. Sonographic findings of breast hamartoma: emphasis on compressibility. Yonsei Med J 2003; 44: Moon WK, Chang RF, Chen CJ, Chen DR, Chen WL. Solid breast masses: classification with computer-aided analysis of continuous US images obtained with probe compression. Radiology 2005; 236: Khoury V, Cardinal E. Tenomalacia : a new sonographic sign of tendinopathy? Eur Radiol 2009; 19: 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: Fornage BD, Sneige N, Faroux MJ, Andry E. Sonographic appearance and ultrasound-guided fine-needle aspiration biopsy of breast carcinomas smaller than 1 cm3. J Ultrasound Med 1990; 9: Fornage BD, Lorigan JG, Andry E. Fibroadenoma of the breast: sonographic appearance. Radiology 1989; 172: Venta LA, Dudiak CM, Salomon CG, Flisak ME. Sonographic evaluation of the breast. Radio- Graphics 1994; 14: Ophir J, Alam SK, Garra B, et al. Elastography: ultrasonic estimation and imaging of the elastic properties of tissues. Proc Inst Mech Eng H 1999; 213: Gao L, Parker KJ, Lerner RM, Levinson SF. Imaging of the elastic properties of tissue: a review. Ultrasound Med Biol 1996; 22: Garra BS, Cespedes EI, Ophir J, et al. Elastography of breast lesions: initial clinical results. Radiology 1997; 202: Cochlin DL, Ganatra RH, Griffiths DF. Elastography in the detection of prostatic cancer. Clin Radiol 2002; 57: Lyshchik A, Higashi T, Asato R, et al. Thyroid gland tumor diagnosis at US elastography. Radiology 2005; 237: Rago T, Santini F, Scutari M, et al. Elastography: new developments in ultrasound for predicting malignancy in thyroid nodules. J Clin Endocrinol Metab 2007; 92: Dighe M, Bae U, Richardson ML, Dubinsky TJ, Minoshima S, Kim Y. Differential diagnosis of thyroid nodules with US elastography using carotid artery pulsation. Radiology 2008; 248: AJR:198, February 2012