Receiver operating characteristic (ROC) curves revealed the best cutoff of 3 for the TIRADS and low suspicion of malignancy for the US 2015 ATA guidelines. For nodules without nonspecific US ATA patterns, AUCs, sensitivities, specificities, PPVs and NPVs were 0.934 (95%CI 0.907–0.948), 96.3% (95%CI 93.6–98.1), 85.4% (95%CI 81.1–89.1), 86.9% (95%CI 83.0–90.2), and 95.8% (95%CI 92.8–97.8), respectively, for the TIRADS, and 0.930 (95%CI 0.912–0.952), 92.6% (95%CI 89.2–95.2), 87.3% (95%CI 83.1–90.7), 88.0% (95%CI 84.0–91.2), and 92.1% (95%CI 88.5–94.9), respectively, for the US ATA system (Table 4). For all nodules, AUCs, sensitivities, specificities, PPVs and NPVs were 0.926 (95%CI 0.904–0.944), 96.7% (95%CI 94.4–98.3), 81.5% (95%CI 77.0–85.5), 84.9% (95%CI 81.1–88.2), and 95.9% (95%CI 92.9–97.8) for the TIRADS, and 0.920 (95%CI 0.898–0.939), 93.5% (95%CI 90.4–95.8), 82.4% (95%CI 77.9–86.3), 85.1% (95%CI 81.3–88.4), and 92.1% (95%CI 88.5–94.9) for the US ATA system, respectively. The TIRADS had higher AUC, sensitivity, and NPV (P < .05, respectively) (Table 4).
3.4 Subgroup analysis
For nodules <1 cm, the AUC, sensitivity, specificity, NPV, and PPV were 0.853, 83.0%, 80.3%, 93.0%, and 59.8% for the TIRADS, respectively, and 0.859, 88.4%, 75.0%, 91.8%, and 67.1% for US ATA patterns, respectively. The US ATA patterns had a higher NPV compared with the TIRADS (P < .05). There were no differences in AUC, sensitivity, and specificity between the 2 systems. For nodules of 1 to 2 cm, the AUC, sensitivity, specificity, NPV, and PPV were 0.902, 98.7%, 80.3%, 86.4%, and 98.0% for the TIRADS, respectively, and 0.899, 97.4%, 78.7%, 85.2%, and 96.0% for US ATA patterns, indicating a higher NPV in the TIRADS (P < .05). For nodules >2 cm, the AUC, sensitivity, specificity, NPV, and PPV were 0.941, 87.8%, 91.8%, 72.9%, and 96.7% for the TIRADS, and 0.926, 75.5%, 94.3%, 77.1%, and 93.8% for US ATA patterns. These data indicated that the TIRADS had higher AUC, sensitivity, and NPV (Table 5).
US is used to evaluate the malignancy potential of thyroid nodules and help in biopsy decision. The TIRADS has been clinically used for a few years. A recent meta-analysis reported pooled sensitivity and specificity of the TIRADS in differentiated diagnosis of thyroid nodules to be 0.79 and 0.71, respectively. However, it has not yet been adopted by the ATA guidelines,[7,23] the American Association of Clinical Endocrinologists (AACE), and the British Thyroid Association. In 2016, an US-assisted risk stratification system was proposed by the ATA guidelines; however, more clinical data are required to determine its sensitivity and specificity.
In this study, the US part of the 2015 ATA guidelines patterns were used to assess 515 patients with 708 thyroid nodules. The results showed a malignancy risk of 90.0% for high suspicion, 57.9% for intermediate suspicion, 17.7% for low suspicion, and 69.4% for nonspecific nodules. According to the US 2015 ATA guidelines, the malignancy rate of nodules with high suspicion can reach 70% to 90%, corroborating our findings. Meanwhile, the malignancy rates for intermediate and low suspicion nodules are 10% to 20% and 5% to 10%, respectively, according to the US ATA guidelines, which are much lower compared with those of the current study. This discrepancy might be explained by the fact that nodules obtained postoperatively may lead to selection bias. Indeed, patients with malignancy-suspected US features, including hypoechoic solid composition, microcalcification, and irregular margins, were more likely to undergo surgery, even though their final diagnoses might be benign tumors. Hypoechogenicity is an effective index in predicting the malignancy rate. A recent study reported that about 55% of benign nodules are hypoechoic, with sub-centimeter benign nodules more likely to be hypoechoic. In addition, no malignancy was found in this study for very low suspicion nodules, while 3% was proposed by the US 2015 ATA guidelines.
The malignancy rates were 90.1%, 67.0%, 8.1%, and 0 for TIRADS 5, 4, 3, and 2 nodules, respectively. Malignancy rates were expected to be >60%, 15% to 50%, 3% to 15%, and <3% for TIRADS 5, 4, 3, and 2 tumors, respectively. In the present study, the malignancy rate obtained for TIRADS 4 nodules was overtly higher. This elevated malignancy risk in category 4 nodules can be attributed to a selection bias, as for the US part of the 2015 ATA classification.
As shown by ROC curves, for nodules without nonspecific ones, the TIRADS system had a relatively higher sensitivity, specificity and NPV compared with the US ATA system for malignancy risk stratification. For nodules with nonspecific ones, the TIRADS system also had a relatively higher AUC, sensitivity and NPV compared with the US ATA system. These findings suggested that the TIRADS system could be used for nodules with nonspecific patterns according to the US ATA guidelines. Meanwhile, the TIRADS had a higher diagnostic value compared with US ATA guidelines for all nodules, including those with nonspecific US ATA patterns. In the sub-centimeter group, US ATA patterns had a better NPV (P < .05). For nodules of 1 to 2 cm, the TIRADS had a higher NPV, and slightly but non-significantly higher AUC, sensitivity, specificity, and PPV. For those >2 cm in diameter, the TIRADS had higher AUC, sensitivity, and NPV. Taken together, these findings suggested that the TIRADS is more efficient than the US part of the ATA guidelines in determining the malignancy of larger nodules.
In the ATA guidelines, a nonspecific group was identified with a high malignancy risk (69.4%) according to US. In the latter group, most nodules were >1 cm in diameter (1.68 ± 1.20 cm), with partially cystic composition, or iso- or hyper-echogenicity. Indeed, only 5-26% of partially cystic nodules are malignant, and iso- and hyperechogenicity are more likely to reflect benignity compared with hypoechogenicity.[26–28] Recently, Yoon et al  also reported that nodules not meeting the criteria for any specific pattern in the US ATA guidelines have a relatively high risk of malignancy (18.2%) as they compared malignancy risk stratification of thyroid nodules by the US ATA guidelines and the TIRADS. In the present study, the proportion of nodules assigned to the “unspecified” category by the US ATA guidelines was higher (8.7% vs 3.4%), with a dramatically higher malignancy rate in those nodules (69.4% vs 18.2%). All these nonspecific nodules could be classified as TIRADS 4 with a similar malignancy rate (69.4% vs 67.4%).
US features are very helpful in determining the follow-up procedure for thyroid nodules with benign cytology diagnoses. The TIRADS adopted in this study was proposed by Na et al, combining specific suspicious US features and less specific ones (solidity and echogenicity) in thyroid nodules. It was shown to be useful for risk stratification of thyroid nodules and management decision. Particularly, for cytologically indeterminate thyroid nodules, both the TIRADS and US ATA guidelines allow high-confidence exclusion of malignancy with stringent negativity cut-offs, and high sensitivity may be obtained using the US part of the ATA guidelines.
The limitations of the present study should be addressed. First, the study is retrospective including only patients who had a surgical resection and thus influenced by many other factors to include selection bias and changes in clinical practice over time during data collection. In our study, 51.8% of the nodules was carcinoma which was higher than that of some retrospective studies [31,32] but comparable to those reported studies with similar design. The malignancy rate of our study was higher than that of some retrospective study, such as 37.2% from Han et al  and 39% of Xu et al. An important reason is that in the above 2 studies, 25.9% and 90.1% of the nodules were regarded as benign lesions based on cytology and follow-up US, which may cause false negative results. While in another retrospective study with a similar design, 2544 thyroid nodules in 1758 patients who underwent thyroidectomy were included. Of all the nodules, 863 (33.9%) were benign, whereas 1681 (66.1%) were malignant. The malignancy rate was relatively higher than our study. Secondly, only 2 radiologists retrospectively reviewed the US images and classified the nodules according to the US 2015 ATA guidelines and TIRADS, with possible deviations among investigators. Thirdly, The TIRADS-Na used in this study is a relatively new tool. Therefore the universality of the present study was limit for the small application range. Finally, some patients had multiple nodules, which included nodules <1 cm. Nodules >1 cm may be punctured and, when confirmed as PTC, they undergo surgery. The 2 lobes of the thyroid were removed during surgery and contained the nodules <1 cm. Because the study period was 2013 to 2016 and the ATA guidelines were released in 2015, the understanding of the follow-up observation for thyroid microcarcinoma was not mature at that time, the pathologists suggested that all patients with PTC had to undergo surgery.
In conclusion, both the TIRADS and the US part of the 2015 ATA guidelines have appreciable diagnostic values for thyroid nodules. This study identified thyroid nodules with nonspecific US ATA patterns. The newly proposed TIRADS may be more efficient than the US 2015 ATA guidelines, especially for nodules >2 cm in diameter or those with nonspecific patterns. The identification of nodules with nonspecific US patterns indicates the need for improving the ATA guidelines for risk stratification. The ATA US guideline is used not only for FNA determination, but also for providing a follow-up recommendation for nodules with benign or indeterminate cytology and nodules without FNA. These evaluation and follow-up recommendations remain instructive in the management of nodules with specific patterns. TIRADS is much more used in determining the need of FNA instead of further management. Therefore, a wiser improvement of TIRADS into clinical management of thyroid nodules is required. The best approach is to tailor TIRADS as ATA US guidelines in the aspect of nodules management. The use of US based on these results could be invaluable when combined with clinical features and biopsy results.
Conceptualization: Pingping Xiang, Shuhang Xu, Chao Liu.
Data curation: Pingping Xiang, Xiaoqiu Chu, Guofang Chen, Binbin Liu, Wenbo Ding, Zheng Zeng, Xinping Wu, Jianhua Wang.
Formal analysis: Pingping Xiang, Xiaoqiu Chu, Guofang Chen, Binbin Liu, Wenbo Ding, Zheng Zeng, Xinping Wu, Jianhua Wang, Shuhang Xu, Chao Liu.
Funding acquisition: Shuhang Xu, Chao Liu.
Project administration: Shuhang Xu, Chao Liu.
Writing – original draft: Pingping Xiang.
Writing – review & editing: Xiaoqiu Chu, Guofang Chen, Binbin Liu, Wenbo Ding, Zheng Zeng, Xinping Wu, Jianhua Wang, Shuhang Xu, Chao Liu.
. Nguyen QT, Lee EJ, Huang MG, et al. Diagnosis and treatment of patients with thyroid cancer. Am Health Drug Benefits 2015;8:30–40.
. Wang K, Yang Y, Wu Y, et al. The association between insulin resistance and vascularization of thyroid nodules. J Clin Endocrinol Metab 2015;100:184–92.
. Gharib H, Papini E, Garber JR, et al. American Association of Clinical Endocrinologists, American College of Endocrinology, and Associazione Medici Endocrinologi Medical Guidelines for Clinical Practice for the Diagnosis and Management of Thyroid Nodules—2016 Update. Endocr Pract 2016;22:622–39.
. Guth S, Theune U, Aberle J, et al. Very high prevalence of thyroid nodules detected by high frequency (13MHz) ultrasound
examination. Eur J Clin Invest 2009;39:699–706.
. Burman KD, Wartofsky L. Thyroid nodules. N Engl J Med 2016;374:1294–5.
. Yeh MW, Bauer AJ, Bernet VA, et al. American Thyroid Association statement on preoperative imaging for thyroid cancer surgery. Thyroid 2015;25:3–14.
. Haugen BR, Alexander EK, Bible KC, et al. 2015 American Thyroid Association Management Guidelines for Adult Patients with Thyroid Nodules and Differentiated Thyroid Cancer: The American Thyroid Association Guidelines
Task Force on Thyroid Nodules and Differentiated Thyroid Cancer. Thyroid 2016;26:1–33.
. Mussa A, De Andrea M, Motta M, et al. Predictors of malignancy in children with thyroid nodules. J Pediatr 2015;167:886-892 e881.
. Remonti LR, Kramer CK, Leitao CB, et al. Thyroid ultrasound
features and risk of carcinoma: a systematic review and meta-analysis of observational studies. Thyroid 2015;25:538–50.
. Horvath E, Majlis S, Rossi R, et al. An ultrasonogram reporting system for thyroid nodules stratifying cancer risk for clinical management. J Clin Endocrinol Metab 2009;94:1748–51.
. Na DG, Baek JH, Sung JY, et al. Thyroid imaging reporting and data system
risk stratification of thyroid nodules: categorization based on solidity and echogenicity. Thyroid 2016;26:562–72.
. Russ G, Royer B, Bigorgne C, et al. Prospective evaluation of thyroid imaging reporting and data system
on 4550 nodules with and without elastography. Eur J Endocrinol 2013;168:649–55.
. Kwak JY, Han KH, Yoon JH, et al. Thyroid imaging reporting and data system
for US features of nodules: a step in establishing better stratification of cancer risk. Radiology 2011;260:892–9.
. Grant EG, Tessler FN, Hoang JK, et al. Thyroid ultrasound
reporting lexicon: white paper of the ACR Thyroid Imaging. Reporting and Data System (TIRADS) Committee. J Am Coll Radiol 2015;12:1272–9.
. Yoon JH, Lee HS, Kim EK, et al. Malignancy risk stratification of thyroid nodules: comparison between the thyroid imaging reporting and data system
and the 2014 American Thyroid Association Management Guidelines. Radiology 2016;278:917–24.
. Moon WJ, Baek JH, Jung SL, et al. Ultrasonography and the ultrasound
-based management of thyroid nodules: consensus statement and recommendations. Korean J Radiol 2011;12:1–4.
. Moon WJ, Jung SL, Lee JH, et al. Benign and malignant thyroid nodules: US differentiation--multicenter retrospective study. Radiology 2008;247:762–70.
. Kim BM, Kim MJ, Kim EK, et al. Sonographic differentiation of thyroid nodules with eggshell calcifications. J Ultrasound
. Brito JP, Gionfriddo MR, Al Nofal A, et al. The accuracy of thyroid nodule ultrasound
to predict thyroid cancer: systematic review and meta-analysis. J Clin Endocrinol Metab 2014;99:1253–63.
. Gul K, Ersoy R, Dirikoc A, et al. Ultrasonographic evaluation of thyroid nodules: comparison of ultrasonographic, cytological, and histopathological findings. Endocrine 2009;36:464–72.
. Salmaslioglu A, Erbil Y, Dural C, et al. Predictive value of sonographic features in preoperative evaluation of malignant thyroid nodules in a multinodular goiter. World J Surg 2008;32:1948–54.
. Frates MC, Benson CB, Doubilet PM, et al. Prevalence and distribution of carcinoma in patients with solitary and multiple thyroid nodules on sonography. J Clin Endocrinol Metab 2006;91:3411–7.
. Cooper DS, et al. American Thyroid Association Guidelines
Taskforce on Thyroid N, Differentiated Thyroid C. Revised American Thyroid Association management guidelines for patients with thyroid nodules and differentiated thyroid cancer. Thyroid 2009;19:1167–214.
. Wei X, Li Y, Zhang S, et al. Meta-analysis of thyroid imaging reporting and data system
in the ultrasonographic diagnosis of 10,437 thyroid nodules. Head Neck 2016;38:309–15.
. Perros P, Boelaert K, Colley S, et al. Guidelines for the management of thyroid cancer. Clin Endocrinol (Oxf) 2014;81(Suppl 1):1–22.
. Chan BK, Desser TS, McDougall IR, et al. Common and uncommon sonographic features of papillary thyroid carcinoma. J Ultrasound
. Watters DA, Ahuja AT, Evans RM, et al. Role of ultrasound
in the management of thyroid nodules. Am J Surg 1992;164:654–7.
. Lee MJ, Kim EK, Kwak JY, et al. Partially cystic thyroid nodules on ultrasound
: probability of malignancy and sonographic differentiation. Thyroid 2009;19:341–6.
. Grani G, Lamartina L, Ascoli V, et al. Ultrasonography scoring systems can rule out malignancy in cytologically indeterminate thyroid nodules. Endocrine 2017;57:256–61.
. Trimboli P, Fulciniti F, Zilioli V, et al. Accuracy of international ultrasound
risk stratification systems in thyroid lesions cytologically classified as indeterminate. Diagn Cytopathol 2017;45:113–7.
. Ha SM, Ahn HS, Baek JH, et al. Validation of three scoring risk-stratification models for thyroid nodules. Thyroid 2017;27:1550–7.
. Xu T, Gu JY, Ye XH, et al. Thyroid nodule
sizes influence the diagnostic performance of TIRADS and ultrasound
patterns of 2015 ATA guidelines: a multicenter retrospective study. Sci Rep 2017;7:43183.
. Gao L, Xi X, Jiang Y, et al. Comparison among TIRADS (ACR TI-RADS and KWAK- TI-RADS) and 2015 ATA Guidelines in the diagnostic efficiency of thyroid nodules. Endocrine 2019;[Epub ahead of print].
Keywords:Copyright © 2019 the Author(s). Published by Wolters Kluwer Health, Inc.
American Thyroid Association guidelines; sonographic pattern; Thyroid Imaging Reporting and Data System; thyroid nodule; ultrasound