Endocrine Journal 2013, 60 (2), 155-160 Or i g i n a l Relationship between patterns of calcification in thyroid nodules and histopathologic findings Bu Kyung Kim 1), Young Sik Choi 1), Hye Jung Kwon 1), Jun Seop Lee 1), Jae Joon Heo 1), You Jin Han 1), Yo-Han Park 1) and Jeong Hoon Kim 2) 1) Department of Internal Medicine, Kosin University College of Medicine, Busan, Korea 2) Department of General Surgery, Kosin University College of Medicine, Busan, Korea Abstract. Various patterns of calcification have been detected in benign and malignant thyroid nodules on ultrasonography (US). Microcalcification has been found to be highly associated with papillary thyroid carcinoma. However, other patterns of calcification have unclear clinical significance. The aim of this study was to evaluate which pattern of calcification could be predictive of malignancy. A total of 1,431 thyroid nodules of 1,078 patients who received preoperative ultrasound examinations and subsequently underwent thyroidectomy were retrospectively reviewed. The types of calcification were defined as follows: microcalcification, annular-like peripheral calcification, crescent-like peripheral calcification, intranodular coarse calcification, and calcified spot. Of these 1,431 nodules, 1,305 (91.1%) were thyroid carcinomas and 126 (8.9%) were benign nodules. Calcifications were detected in 38.6% (552/1,431) of all nodules. Calcifications were found in 40.2% (524/1,305) of malignant nodules and 22.2% (28/126) of benign nodules. Of the 524 malignant nodules with calcification, microcalcification was the most common pattern, occurring in 42.9% (225/524), and annular type was the least common pattern, occurring in 5.9% (31/524). Among the calcification types, only microcalcification and intranodular had a significant association with malignancy (p = 0.001, 0.035), with OR values of 3.5 (95% CI, 1.6-7.7) and 2.4 (95% CI, 1.1-5.6). Though using the patterns of calcification alone for predicting malignancy had limitation, microcalcification and intranodular calcification were significantly associated with malignancy. Key words: Thyroid nodule, Calcification Presence of a thyroid nodule is a common clinical problem seen in endocrinology. High-resolution ultrasound has detected thyroid nodules in approximately 19-67% of the general adult population [1]. Differentiating between benign and malignant thyroid nodules is an important clinical concern. Ultrasound (US) features and US-guided fine needle aspiration (FNA) are useful for differentiating between benign and malignant thyroid nodules. However, several limitations exist in both methods to differentiate benign from malignant nodules. One of those limitations is the pattern of calcification. Calcification was found in approximately 19.8-32.1% of thyroid nodules [2, 3] and was detected in both benign and malignant nodules. Various patterns of calcification are seen, including egg- Submitted Aug. 12, 2012; Accepted Sep. 13, 2012 as EJ12-0294 Released online in J-STAGE as advance publication Oct. 6, 2012 Correspondence to: Young Sik Choi, M.D., Ph.D. Department of Internal Medicine, Kosin University College of Medicine, 34 Amnam-dong, Seo-gu, Busan, 602-702, Korea. E-mail: yschoi@kosinmed.or.kr shell or rim-like peripheral calcifications, intranodular coarse dense calcifications, calcified spot, and microcalcifications [4, 5]. Microcalcification is known to be highly associated with papillary thyroid carcinoma and its histologic type is easy decided by US-guided FNA. However, for other patterns of calcification, it is difficult to perform US-guided FNA that nodules and also their clinical significance is uncertain. In the present study, we classified non-microcalcification patterns into four types of calcification. Then, we analyzed five patterns of calcification of thyroid nodules detected on US to evaluate which pattern of calcification was associated with malignancy. Materials and Methods Subjects The records of 1078 consecutive patients who received preoperative ultrasound examinations and subsequently underwent thyroidectomy for nodular thyroid disease between January 2008 and July 2011 The Japan Endocrine Society
156 Kim et al. The patterns of calcification were classified into five categories: (1) microcalcification (fine stippling psammoma body); (2) annular-like peripheral calcification; (3) crescent-like peripheral calcification; (4) intra nodular coarse calcification; and (5) a calcified spot (Figs. 1, 2). Definitions of each were as follows: (1) microcalcification was defined as multiple punctate bright echoes with or without acoustic shadowing, (2) annular-like peripheral calcification was defined as coarse calcifications located around the nodule, (3) crescent-like peripheral calcification consisted of crescentat the Kosin University Gospel Hospital were retrospectively reviewed. Subjects were 171 male and 907 female patients (with a male-to-female ratio of 5.3:1), with a mean age of 47.2 ± 11.1 years old (range 16-82 years old). The reason for the decision to perform surgery varied and ranged from risk of malignancy to cosmetic concerns. Patients charts were reviewed for clinical findings and preoperative diagnostic investigations. All pathology reports were reviewed for data regarding the final diagnosis and the presence of intrathyroidal calcification. This study was approved by the institutional review board (IRB) of Kosin University Gospel Hospital (approval number: 11-87). Assessment of thyroid nodule calcification All patients underwent thyroid US before surgery. Sonographic examination was performed by two radiologists with real-time scanners (Philips Healthcare IU 22, Bothell, WA, USA) equipped with a 5-to12-MHz linear array transducer. All the ultrasound reports and images were reviewed by only one endocrinologist who specializes in thyroid pathology. The size and number of thyroid nodules and the presence and pattern of calcification were recorded for each patient. Fig. 1 Transverse ultrasound image of thyroid nodules. Ultrasound showing 0.8 cm nodule with microcalcifications. Fig. 2 Transverse ultrasound images of non-microcalcification thyroid nodules. (A) Annular-type calcification, (B) crescent-type calcification, (C) Intranodular calcification, and (D) calcified spot
Calcification in thyroid nodules 157 shaped large coarse calcifications, without the ability to identify the tumor by its acoustic shadowing, (4) intra nodular coarse calcification consisted of a hyperechoic structure (larger than 2 mm) with acoustic shadowing, and (5) a calcified spot was a single coarse calcification without visualization of the surrounding tumor. Each calcification type on US was compared with the corresponding histological findings. Statistical analysis The data were analyzed in terms of frequency. Chisquare test was performed for each calcification type, and the sensitivity, specificity and positive predictive value was determined from a two-by-two table. To make up for the small number of cases, Fisher s exact test was also used. The value of p < 0.05 was considered as statistically significant. Statistical analysis was performed using SPSS, version 18.0 (SPSS Inc., Chicago, IL, USA). Results A total of 1,431 thyroid nodules of 1,078 patients were studied, and they had a mean diameter of 1.2 ± 0.8 cm (range 0.1 cm - 6.0 cm). Of these 1,431 nodules, 1,305 (91.1%) were thyroid carcinomas and 126 (8.9%) were benign nodules. The majority of malignant nodules were papillary carcinoma (94.7% [1,236/1,305]). Among the benign nodules, nodular hyperplasia was the most common (57.9% [73/126]). The types of histology are summarized in Table 1. Calcifications were detected in 38.6% (552/1,431) of all nodules. Calcifications were found in 40.2% (524/1,305) of malignant nodules and 22.2% (28/126) of benign nodules. Malignancy was significantly more common in nodules with calcifications as compared with nodules without calcifications (94.9% [524/552] vs. 88.8% [781/879]; p=0.00). Microcalcification was found in 17.2% (225/1,305) of malignant nodules and 5.5% (7/126) of benign nodules. The calcification patterns observed in the less common types of malignant nodules and those in benign thyroid nodules are summarized in Table 2. Fig. 3 shows the configuration according to the types of calcification. Of the 524 malignant nodules with calcification, microcalcification was the predominant pattern, occurring in 42.9% (225/524); in contrast, the annular type was the least common, only occurring in 5.9% (31/524). Of the 28 benign nodules with cal- Table 1 Histology of 1,431 thyroid nodules of 1,078 patients Type of Histology Number (%) Malignant Total 1305 (100) Papillary carcinoma 1236 (94.7) Follicular variant PTC 40 (3.1) Columnar cell variant PTC 2 (0.2) Solid variant PTC 8 (0.6) Tall cell variant PTC 6 (0.4) Follicular carcinoma 10 (0.8) Medullary carcinoma 3 (0.2) Benign Total 126 (100) Nodular hyperplasia 73 (57.9) Follicular adenoma 36 (28.6) Hashimoto s thyroiditis 5 (4.0) Hürthle cell adenoma 11 (8.7) Hyalinizing trabecular tumor 1 (0.8) PTC, papillary thyroid carcinoma Table 2 Ultrasound patterns of calcification and incidence of malignancy in 1,431 excised thyroid nodules Type of calcification Malignant (%) n = 1305 Benign (%) n = 126 No calcification 781 (59.8) 98 (77.8) Calcification 524 (40.2) 28 (22.2) Microcalcification 225 (17.2) 7 (5.6) Annular 31 (2.4) 5 (4.0) Crescent 58 (4.4) 4 (3.2) Calcified spot 70 (5.4) 6 (4.7) Intra-nodular 140 (10.7) 6 (4.7) Fig. 3 The percentage of each type of calcification in malignant and benign nodules
158 Kim et al. Table 3 Sensitivity, specificity, positive predictive value and odds ratio according to calcification type Type of calcification Sensitivity (%) Specificity (%) Positive predictive value (%) Odds ratio (95% CI) p-value Total calcification 40.2 77.8 94.9 2.3 (1.5-3.6) 0.000 Microcalcification 17.2 94.4 97.0 3.5 (1.6-7.7) 0.001 Annular 2.4 96.0 86.1 0.6 (0.2-1.5) 0.240 Crescent 4.4 96.8 93.5 1.4 (0.5-4.0) 0.649 Calcified spot 5.4 95.2 92.1 1.1 (0.5-2.7) 0.774 Intranodular 10.7 95.3 95.9 2.4 (1.1-5.6) 0.035 Annular, annular like peripheral calcification; Crescent, crescent-like peripheral calcification; Intranodular, intranodular coarse calcification OR determined by chi-square test or Fisher s exact test. cification, each calcification type were distributed into similar portion. The overall sensitivity of any calcification was 40.2% and each calcification type had a low sensitivity (microcalcification 17.2%, annular 2.4%, crescent 4.4%, calcified spot 5.4%, and intra nodular 10.7%). The specificity of total calcification, microcalcification, annular, crescent, calcified spot and intranodular was 77.8%, 94.4%, 96.0%, 96.0%, 96.8%, 95.2% and 95.3%, respectively; the positive predictive value was 94.9%, 97.0%, 86.1%, 93.5%, 92.1% and 95.9%, respectively (Table 3). Chi-square test indicated that total calcification differed significantly between the benign and malignant groups (p < 0.001), with an odds ratio value of 2.3 (95% CI, 1.5-3.6). Among the calcification types, only microcalcification and intranodular were significantly associated with malignancy (p = 0.001, 0.035), with OR values of 3.5 (95% CI, 1.6-7.7) and 2.4 (95% CI, 1.1-5.6), respectively. On the other hand, the annular type had a relatively low OR of 0.6 (95% CI, 0.2-1.5), although it was not significant (Table 3). Discussion In our study, thyroid nodules detected on US were classified into five patterns of calcification. Our results showed that ultrasonographic calcification of thyroid nodules was statistically significant for predicting malignant nodules. According to the calcification pattern, microcalcification and intranodular calcification were significantly associated with malignancy. Ultrasonography can detect thyroid nodules at an earlier stage, so it is currently used as the initial investigation for evaluation of thyroid nodules [6]. However, there is no reliable classification for distinguishing malignant from benign. Calcification of a thyroid nod- ule is one of the most important US findings in distinguishing malignant from benign nodules. Several reports have noted that calcification seems to be more common in malignant than benign thyroid nodules [7-11]. The incidence of calcification is 26-54% in malignant and 8-32% in benign lesions. Our study showed similar results that calcification was noted in 40.2% of malignant and in 22.2% of benign thyroid nodules. The patterns of calcification also can be used to predict malignancy. Microcalcification is already known to be the most reliable indicator of malignancy, because microcalcification represents the histological pattern of psammoma bodies [12]. Psammoma bodies are typically seen in papillary thyroid carcinoma. But, ultrasonographic microcalcification is not completely explain psammoma bodies, because there are several calcification types histologically [13, 14]. Wang et al. [4] reported microcalcification was present in 38.2% (26 of 68) of malignant nodules with calcification. Taki et al. [5] reported the proportion of microcalcification in malignant nodules with calcification was 29.0% (9 of 31). In this study, microcalcification occupied 42.9% (225 of 524) of malignant thyroid nodules with calcification. It is considered more reliable than previous studies because of the large sample size. On the other hand, the proportion of microcalcification in benign nodules with calcification was not dominant, 7 out of 28 cases (Fig. 2). This is the first study that classified non-microcalcification patterns into four groups, annular, crescent, calcified spot and intranodular. In particular, the crescent type of calcification was unique to this study. It is well-known that annular or egg-shell -type calcifications are generally thought to be associated with benign nodules. However, these types have been
Calcification in thyroid nodules 159 reported not only in cases of papillary carcinoma [15] and but also in undifferentiated carcinoma [16]. In our study, 5.9% of malignant thyroid nodules had the annular type of calcifications. However, this type was the least frequent and the odds ratio was less than 1 (Table 3) and not significant. In our experience, the clinical feature of the crescent type is similar with the annular type. This type is dense and non-diagnostic in fineneedle aspiration cytology (FNAC). We thought the crescent type was also more likely to be benign than malignant. Crescent type was the second least frequent, occurring in 11.1% of malignant thyroid nodules with calcification. Coarse dystrophic calcification has been reported to be unhelpful in the diagnosis of malignancy [5, 12, 17]. However, these results may be due to the crude classification of non-microcalcification types. Except for crescent and calcified spot types from coarse dystrophic calcification, the intranodular type of calcification was a significant indicator of malignant thyroid nodules with an odds ratio of 2.4 (95% CI 1.1-5.6, p=0.035). One of the most interesting findings was that there was no dominant pattern in benign thyroid nodules with calcification. Khoo et al. [9] showed that 91.7% of benign nodules for which patients underwent surgery were multinodular goiters. The incidence of calcification increased with the duration of the presence of the goiter [7, 8, 18, 19]. Such calcifications have a low risk of malignancy [20]. In this study, the most common reasons for operating for benign nodules were the presence of a multinodular goiter and suspicion of malignancy on the FNAC. From our results, we can speculate that there are no clear indicators of malignancy in multinodular goiter. The present study has several limitations. The most important limitation is that the significant difference in number of malignant and benign patients. This is because of unlikeness of surgery of benign nodules. In this context, one more limitation is raised. The results of the present study were not from all incidentalomas of the thyroid. Almost all cases with microcalcification underwent surgery; however, only a few cases of other patterns of calcification underwent surgery. Therefore, the sensitivity and specificity of microcalcification seem to be reliable, but the results of other patterns may not be given the smaller sample size. Despite these limitations, this study has several strengths. First, it includes the largest number of cases in the literature to date, a total of 1,431 thyroid nodules. Second, this is the first report which classified microcalcifications into five patterns with detailed descriptions of each type and examined their associations with malignancy. In conclusion, if the five types of calcification pattern which were presented in this study would be used in routine examination of thyroid nodules, it may aid in the diagnosis of malignancy. Not only microcalcification but also intranodular calcification should be considered to be suspicious for malignancy. Extended studies for all incidentalomas of the thyroid are needed in the future. Acknowledgments This study was supported by a grant from Kosin University College of Medicine (2010). References 1. Hegedus L (2004) Clinical practice. The thyroid nodule. N Engl J Med 351:1764-1771. 2. Lu Z, Mu Y, Zhu H, Luo Y, Kong Q, et al. (2011) Clinical value of using ultrasound to assess calcification patterns in thyroid nodules. World J Surg 35:122-127. 3. Chen G, Zhu XQ, Zou X, Yao J, Liang JX, et al. (2009) Retrospective analysis of thyroid nodules by clinical and pathological characteristics, and ultrasonographically detected calcification correlated to thyroid carcinoma in South China. Eur Surg Res 42:137-142. 4. Wang N, Xu Y, Ge C, Guo R, Guo K (2006) Association of sonographically detected calcification with thyroid carcinoma. Head Neck 28:1077-1083. 5. Taki S, Terahata S, Yamashita R, Kinuya K, Nobata K, et al. (2004) Thyroid calcifications: sonographic patterns and incidence of cancer. Clin Imaging 28:368-371. 6. Seiberling KA, Dutra JC, Grant T, Bajramovic S (2004) Role of intrathyroidal calcifications detected on ultrasound as a marker of malignancy. Laryngoscope 114:1753-1757. 7. Hayashi N, Tamaki N, Yamamoto K, Senda M, Yonekura Y, et al. (1986) Real-time ultrasonography of thyroid nodules. Acta Radiol Diagn (Stockh) 27:403-408. 8. Watters DA, Ahuja AT, Evans RM, Chick W, King WW, et al. (1992) Role of ultrasound in the management of thyroid nodules. Am J Surg 164:654-657. 9. Khoo ML, Asa SL, Witterick IJ, Freeman JL (2002)
160 Kim et al. Thyroid calcification and its association with thyroid carcinoma. Head Neck 24:651-655. 10. Kakkos SK, Scopa CD, Chalmoukis AK, Karachalios DA, Spiliotis JD, et al. (2000) Relative risk of cancer in sonographically detected thyroid nodules with calcifications. J Clin Ultrasound 28:347-352. 11. Koike E, Noguchi S, Yamashita H, Murakami T, Ohshima A, et al. (2001) Ultrasonographic characteristics of thyroid nodules: prediction of malignancy. Arch Surg 136:334-337. 12. Takashima S, Fukuda H, Nomura N, Kishimoto H, Kim T, et al. (1995) Thyroid nodules: re-evaluation with ultrasound. J Clin Ultrasound 23:179-184. 13. Das DK, Mallik MK, Haji BE, Ahmed MS, Al-Shama a M, et al. (2004) Psammoma body and its precursors in papillary thyroid carcinoma: a study by fine-needle aspiration cytology. Diagn Cytopathol 31:380-386. 14. Bai Y, Zhou G, Nakamura M, Ozaki T, Mori I, et al. (2009) Survival impact of psammoma body, stromal calcification, and bone formation in papillary thyroid carcinoma. Mod Pathol 22:887-894. 15. Park CH, Rothermel FJ, Judge DM (1976) Unusual calcification in mixed papillary and follicular carcinoma of the thyroid gland. Radiology 119:554. 16. Vescini F, Di Gaetano P, Vigna E, Pascoli A, Cacciari M (2000) Anaplastic thyroid carcinoma in a 49 yearold woman with a long-standing goiter. A case report. Minerva Endocrinol 25:81-83. 17. Tessler FN, Tublin ME (1999) Thyroid sonography: current applications and future directions. AJR Am J Roentgenol 173:437-443. 18. Komolafe F (1981) Radiological patterns and significance of thyroid calcification. Clin Radiol 32:571-575. 19. Katz JF, Kane RA, Reyes J, Clarke MP, Hill TC (1984) Thyroid nodules: sonographic-pathologic correlation. Radiology 151:741-745. 20. Kuma K, Matsuzuka F, Kobayashi A, Hirai K, Morita S, et al. (1992) Outcome of long standing solitary thyroid nodules. World J Surg 16:583-587; discussion 587-588.