Neuroradiology/Head and Neck Imaging Original Research Virmani and Hammond Sonographic Patterns of enign Thyroid Nodules Neuroradiology/Head and Neck Imaging Original Research Vivek Virmani 1 Ian Hammond Virmani V, Hammond I Keywords: benign thyroid nodules, Hashimoto thyroiditis, sonographic patterns, thyroid nodules, ultrasound DOI:10.2214/JR.10.5363 Received July 20, 2010; accepted without revision ugust 23, 2010. 1 oth authors: Department of Diagnostic Imaging, The Ottawa Hospital, University of Ottawa, 501 Smyth Rd, Ottawa, ON K1H 8L6, Canada. ddress correspondence to V. Virmani (drvivekvirmani@rediffmail.com). JR 2011; 196:891 895 0361 803X/11/1964 891 merican Roentgen Ray Society Sonographic Patterns of enign Thyroid Nodules: Verification at Our Institution OJECTIVE. The objective of our study was to apply the sonographic criteria for the diagnosis of benign thyroid nodules described by onavita et al. to our patient referral population. MTERILS ND METHODS. We performed a retrospective review of the sonographic features of 811 thyroid nodules that had undergone fine-needle aspiration biopsy in our radiology department from July 2005 to July 2009 and that had conclusive cytologic results. We determined the percentage of nodules that exhibited strictly the four benign sonographic patterns described by onavita et al. and the accuracy of those patterns in the prediction of benign cytologic findings. RESULTS. Sixty-six nodules (8.1%) showed a spongiform pattern and all were benign; 28 nodules (3.5%) showed the pattern described as cyst with colloid clot and all were benign; 14 nodules (1.7%) showed a giraffe pattern and all were proven to be Hashimoto thyroiditis; and eight nodules (1%) showed the white knight pattern and all were Hashimoto thyroiditis. Of the 121 nodules in our series with malignant cytologic features, none exhibited any of these four benign patterns. CONCLUSION. The four sonographic patterns described by onavita et al., when applied strictly, were 100% specific for benignity. However in our referral population, the application of these patterns would have eliminated the need for only 14% of biopsies, in contrast with 61% of biopsies in the population studied by onavita et al. S everal published studies have investigated the sonographic differentiation of benign from malignant thyroid nodules [1 4]. lthough no single criterion can reliably distinguish between a malignant nodule and a benign nodule, a combination of features, such as solid texture, hypoechogenicity, and intrinsic microcalcification in a nodule, predicts that the nodule is significantly more likely to be malignant than is a nodule without those features. Reading et al. [5] proposed a classical pattern approach to diagnose thyroid nodules and noted several patterns typical of benign thyroid lesions [5]. onavita et al. [6] expanded on that work by describing four characteristic morphologic patterns that in their sample of 500 nodules were 100% specific for benignity: spongiform configuration; cyst with colloid clot; giraffe pattern; and diffuse hyperechogenicity, which they refer to as white knight. In their series, identification of the nodules with those patterns would have obviated 61% of thyroid biopsies. The purpose of our study was to test the reliability of these four sono- graphic patterns for the identification of benign thyroid nodules in our radiologic practice. Materials and Methods Data Collection We conducted a retrospective review of the pathology records of 950 thyroid nodules that underwent ultrasound-guided fine-needle aspiration biopsy (FN) in the radiology department of a single institution between July 2005 and July 2009. The biopsies were performed by three experienced radiologists, each with a minimum of 20 years in radiologic practice, or on occasion by a radiology resident under the direct supervision of one of the three radiologists. Most aspirations were performed with a 25-gauge needle using capillary action, and there were at least three passes per nodule. The reports with the FN results from the 950 nodules were classified initially into one of four cytologic categories: category 1, insufficient cells; category 2, benign (i.e., colloid nodules, adenomatous hyperplasia, and thyroiditis); category 3, indeterminate (i.e., follicular lesions and Hürthle cell neoplasms); and category 4, malignant. The 127 nodules in category 1 were eliminated from the study. Nodules in category JR:196, pril 2011 891
Virmani and Hammond 3 were admitted to the study group only if there was surgical confirmation of their nature and ultimately classified as benign or malignant. This excluded another 12 nodules. The ultrasound images of the remaining eligible 811 nodules were then analyzed. The study had institutional review board approval. Image Interpretation The ultrasound images were reviewed in consensus by an attending staff radiologist with more than 30 years of ultrasound experience and a fellow in radiology with 5 years of ultrasound experience. The objective of the consensus review was to select all nodules that could be assigned to any one of the four definitely benign morphologic patterns described by onavita et al. [6] and to then correlate those patterns with the corresponding pathology results. In cases in which the two observers disagreed, the opinion of the senior radiologist was used. The following sonographic criteria of onavita et al. [6] were applied: The spongiform pattern is an aggregation of multiple linear microcystic components in the nodule giving a honeycomb appearance or puff pastry appearance. The nodule TLE 1: Distribution of enign and Malignant Nodules in Our Study Group and of the Four Sonographic Patterns Described by onavita et al. [6] Diagnosis Total Pattern 1: Spongiform No. of Thyroid Nodules Pattern 2: Cyst With Colloid Clot Pattern 3: Giraffe Pattern 4: White Knight enign 690 66 28 14 8 Malignant 121 0 0 0 0 C Fig. 1 Sonographic criteria for benign thyroid nodules described by onavita et al. [6]., Pattern 1. Ultrasound scan of 37-year-old man with multinodular goiter shows spongiform nodule. Similarity of nodule to water-filled sponge or puff pastry is evident., Pattern 2. Ultrasound scan of 49-year-old man with colloid cyst shows cyst with colloid clot. When cystic portion of nodule is subtracted, spongiform nodule, or type 1 nodule, remains. C, Pattern 3. Ultrasound scan of 28-year-old woman with Hashimoto thyroiditis shows nodule with hyperechoic areas separated by hypoechoic bands, much like pattern of a giraffe s hide. D, Pattern 4. Ultrasound scan of 27-year-old woman with Hashimoto thyroiditis shows homogeneous hyperechoic nodule, which onavita et al. referred to as white knight. is avascular or isovascular in relation to the rest of the gland (Fig. 1). The cyst with a colloid clot is a cystic nodule containing a retracted clot or avascular colloid plug so that if the cystic portion were subtracted, a spongiform nodule would remain (Fig. 1). The giraffe pattern is characterized by rounded or ovoid areas of hyperechogenicity separated by thin linear areas of hypoechogenicity that appear similar to the two-tone blocklike coloring of a giraffe (Fig. 1C). The fourth pattern is a homogeneous hyperechoic nodule, which was named a white knight (Fig. 1D). Statistical nalysis The sensitivity and specificity of these four patterns for the diagnosis of benignity were evaluated. Results The 811 nodules in our study were from 661 patients (552 females and 109 males; mean age, 46.5 years; age range, 13 88 years). The nodules ranged in size from 4.8 to 72 mm D 892 JR:196, pril 2011
Sonographic Patterns of enign Thyroid Nodules (mean size, 16.2 ± 10.8 mm [SD]). The distribution of the benign and malignant nodules across the four patterns is illustrated in Table 1. Sixty-six of the 811 nodules (8.1%) showed a spongiform pattern. Of those nodules, 60 were benign on cytologic analysis and the remaining six showed indeterminate cytologic findings but were confirmed to be histologically benign after surgical resection. Twenty-eight of the 811 nodules (3.5%) showed the cyst with a colloid clot pattern: 24 were benign on cytology and four were benign on postoperative histologic analysis. Fourteen nodules (1.7%) exhibited the giraffe pattern, of which 13 had a cytologic diagnosis of Hashimoto thyroiditis. The remaining nodule, which had indeterminate cytology results, was surgically proven Hashimoto thyroiditis. Eight nodules (1%) were classified with the white knight pattern and all were diagnosed as Hashimoto thyroiditis on cytology. None of the 121 malignant nodules exhibited any of the proposed benign sonographic patterns. Discussion The widespread dissemination of modern medical imaging in particular, high-resolution ultrasound has resulted in our ability to detect thyroid nodules in at least 50% of Fig. 2 Potential pitfalls in applying sonographic criteria described by onavita et al. [6]., 47-year-old man with papillary carcinoma. Ultrasound scan shows complex cystic nodule (cursors) that can be mistaken for colloid clot pattern if criteria are not strictly applied., 52-year-old woman with papillary carcinoma. Ultrasound scan shows solid nodule (cursors) in cyst that resembles colloid clot on gray-scale imaging. C, Doppler ultrasound image of nodule shown in reveals intense vascularity of nodule. This finding excludes colloid clot pattern, which must be isovascular or avascular. the general adult population, compared with a detection rate of 4 8% by traditional clinical palpation [7]. Several sets of guidelines have been published to direct the management of the large number of nodules so discovered [8, 9]. The latest merican Thyroid ssociation guidelines recommend FN of nodules larger than 5 mm in diameter in patients at high risk for thyroid cancer and, depending on their sonographic features, FN of nodules more than 1 1.5 cm in all other patients [8]. Given the resource implications of performing FNs on such a large cohort of patients, the ability to reliably identify even a small percentage of nodules that can be left alone would be of significant benefit [10, 11]. Fig. 3 39-year-old woman with papillary carcinoma. Ultrasound scan shows homogeneous hyperechoic nodule (cursors) with macrocalcification within it. This nodule resembles white knight, but presence of calcification distinguishes it from that benign pattern. Thus, we were eager to field test the results of onavita et al. [6] in our institution, where we currently perform FN of approximately 600 thyroid nodules annually in the radiology department. s onavita et al. did, we reviewed a selection of cases in which both pathologic reports and ultrasound images were available and our results are in agreement with theirs; application of the sonographic criteria they described was 100% specific for the detection of benign thyroid nodules. In accordance with their observations, we noted that the spongiform pattern was the most common of the four leave-alone patterns and that the giraffe pattern and the white knight, although relatively uncommon, C JR:196, pril 2011 893
Virmani and Hammond Fig. 4 26-year-old woman with papillary carcinoma on background of Hashimoto thyroiditis., Ultrasound scan shows giraffe pattern that is typical for Hashimoto thyroiditis., Caudal sections of ultrasound scan show hypoechoic nodule on background of Hashimoto thyroiditis. Nodule was surgically proven papillary carcinoma. were indicative of Hashimoto thyroiditis. None of the malignant nodules in our series displayed any one of the four patterns. lthough we did not attempt to measure interobserver variability, we did note three potential pitfalls in the application of these sonographic patterns. First, we identified a partially cystic papillary carcinoma that could be confused with a cyst with colloid clot and another cystic lesion with a hypervascular peripheral clot that could potentially be mistaken for the colloid clot pattern on gray-scale imaging (Fig. 2). Second, we encountered one white knight variant in that it contained a nidus of central calcification that was also a papillary cancer (Fig. 3). Third, we observed one patient with a nodule showing the giraffe pattern in association with a solid hypoechoic nodule, the latter of which was proven to be papillary cancer on a background of Hashimoto disease (Fig. 4). It has been reported that there is an increased incidence of papillary cancer in women with Hashimoto disease [12]. Thus, we emphasize that the sonographic criteria reported by onavita et al. [6] must be interpreted strictly if biopsy is to be avoided. This guideline is particularly true with the colloid clot pattern. s an incidental finding, we observed a proven benign colloid nodule evolve from a spongiform nodule into a cyst with colloid clot, suggesting that these two patterns are part of a continuum (Fig. 5). The major discrepancy between our series and that of onavita et al. [6] is the relatively low sensitivity of 14% for the detection of benign nodules compared with their sensitivity of 61%. lthough we analyzed 811 nodules compared with their 500 nodules, it seems more likely that differences in our respective populations or referral patterns are responsible for this discrepancy than is a difference in sample size. In our sample, there was a 14.9% incidence of thyroid carcinoma; in their study sample, the incidence of carcinoma was 3.9%, rising to 7.9% with the inclusion of follicular and Hürthle cell neoplasms, which are not always malignant when examined histologically. The malignancy rate in thyroid nodules is reported in various series as 5 15% [3]. Endocrinologists in our region do not perform thyroid ultrasound or ultrasound-guided FN, whereas this practice is becoming more widespread in the United States [13]. It is possible that many malignant nodules were prescreened and thereby removed from those ultimately referred to onavita et al. for inclusion in the study group. n alternate or additional explanation might be that there is a greater prevalence of thyroid carcinoma in our local population than in their population. The limitations of our study, which we share with onavita et al. [6], include its retrospective nature and the consensus analysis of sonograms by only two radiologists. s did onavita et al., we relied on cytology for the diagnosis of the benign and obviously malignant nodules, but we required histologic proof for indeterminate nodules following an aspiration cytology classification used in a medical audit in England [14]. We recommend strongly that other workers wishing to apply these sonographic criteria determine the validity of the patterns in their local referral population and assess their personal confidence in recognizing the four patterns. Fig. 5 55-year-old woman with benign thyroid nodule., Ultrasound scan shows solid nodule with spongiform appearance., Ultrasound scan of same nodule obtained 4 months after shows cystic change within nodule that gives nodule colloid cyst appearance. 894 JR:196, pril 2011
Sonographic Patterns of enign Thyroid Nodules cknowledgments We thank Donald lair and Margaret Fraser-Hill for their technical expertise in performing many of the biopsies. nodules: US differentiation multicenter retrospective study. Radiology 2008; 247:762 770 5. Reading CC, Charboneau JW, Hay ID, Sebo TJ. Sonography of thyroid nodules: a classical pat- Society of Radiologists in Ultrasound. Management of thyroid nodules detected at US: Society of Radiologists in Ultrasound consensus conference statement. Radiology 2005; 237:794 800 References 1. Papini E, Guglielmi R, ianchini, 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 2. 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 3. Frates MC, enson C, Doubilet PM, et al. Prevalence and distribution of carcinomas in patients with solitary and multiple thyroid nodules on sonography. J Clin Endocrinol Metab 2006; 91: 3411 3417 4. Moon W-J, Jung SL, Lee JH, et al.; Thyroid Study Group; Korean Society of Neuro- and Head and Neck Radiology. enign and malignant thyroid tern diagnostic approach. Ultrasound Q 2005; 21:157 165 6. onavita J, Mayo J, abb J, et al. Pattern recognition of benign nodules at ultrasound of the thyroid: which nodules can be left alone? JR 2009; 193:207 213 7. Ross DS. Nonpalpable thyroid nodules: managing an epidemic. (editorial) J Clin Endocrinol Metab 2002; 87:1938 1940 8. merican Thyroid ssociation Guidelines Taskforce on Thyroid Nodules and Differentiated Thyroid Cancer; Cooper DS, Doherty GM, Haugen R, et al. Revised merican Thyroid ssociation management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid 2009; 19:1167 1214 [Errata in Thyroid 2010; 20:674 675 and Thyroid 2010; 20:942] 9. Frates MC, enson C, Charboneau JW, et al.; 10. Cronan JJ. Thyroid nodules: is it time to turn off the US machines? Radiology 2008; 247:602 604 11. Hammond I, Lentle C, Odell PF. The pursuit of impalpable thyroid nodules: are we using scarce resources wisely? Can ssoc Radiol J 2010; 61: 98 101 12. Repplinger D, argren, Zhang YW, dler JT, Haymart M, Chen H. Is Hashimoto s thyroiditis a risk factor for papillary thyroid cancer? J Surg Res 2008; 150:49 52 13. Davies L, Welch HG. Increasing incidence of thyroid cancer in the United States, 1973 2002. JM 2006; 295:2164 2167 14. Tabaqchali M, Hanson JM, Johnson SJ, Wadehra V, Lennard TW, Proud G. Thyroid aspiration cytology in Newcastle: a six year cytology/histology correlation study. nn R Coll Surg Engl 2000; 82:149 155 JR:196, pril 2011 895