Subcentimetre thyroid nodules: Sonographic features associated with malignancy

Abstract

Introduction

The aim of this study was to identify the ultrasound features which are associated with malignancy in subcentimetre thyroid nodules.

Methods

This retrospective study included 454 thyroid nodules <10 mm in size in 413 patients from 2012 to 2016, which were subjected to fine needle aspiration cytology. Each nodule was classified according to the ultrasound patterns of the 2015 American Thyroid Association guidelines and the high suspicion ultrasound features (solid, hypo-echogenicity, irregular margins, microcalcifications, taller-than-wide, interrupted rim calcifications, and extrathyroidal extension) were identified for evaluation of their diagnostic performance.

Results

Of the American Thyroid Association high suspicion ultrasound features, univariate analysis showed that hypo-echogenicity (sensitivity 81.6% (95% CI 65.7–92.3%), specificity 50.0% (95% CI 43.4–56.6%)), irregular margins (sensitivity 34.2% (95% CI 19.6–51.4%), specificity 92.2% (95% CI 88.0–95.3%)), microcalcifications (sensitivity 23.7% (95% CI 11.4–40.2%), specificity 91.0% (95% CI 86.5–94.3%)), and taller-than-wide (sensitivity 23.7% (95% CI 11.4–40.2%), specificity 92.2% (95% CI 88.0–95.3%)) were significantly associated with a malignant diagnosis. Amongst the above features, subsequent multivariate analysis identified a combination of hypo-echogenicity and irregular margins as significantly associated with malignancy. Our malignancy rates based on American Thyroid Association ultrasound patterns from benign to high suspicion were 0.0, 8.3, 3.9, 15.7, and 40.4%, respectively. The American Thyroid Association high suspicion ultrasound pattern had a sensitivity of 50.0% (95% CI 33.4–66.7%) and specificity of 84.5% (95% CI 79.2–88.9%).

Conclusion

The presence of both hypo-echogenicity and irregular margins was significantly associated with malignancy in subcentimetre thyroid nodules in our study. Hence, subcentimetre nodules which are hypoechoic with irregular margins may warrant further follow-up.

Keywords

Thyroid subcentimetre ultrasound malignancy features

Introduction

Thyroid nodules are common with incidences of up to 60%.¹ This has been increasing in the recent years mirroring the advances of ultrasound imaging and the widespread use of thyroid fine needle aspiration cytology (FNAC).^2,3 However, approximately 50% of this increase in thyroid cancers is due to papillary thyroid microcarcinomas (PTMCs). PTMCs are defined as papillary thyroid carcinomas <1 cm in their maximum dimension. These microcarcinomas are indolent and unlikely to progress to clinically significant cancers.³ Therefore, despite the increased rate of diagnosis, thyroid cancer mortality has remained stable. On the other hand, the costs to patients and healthcare systems have escalated. Thyroid cancer has been predicted to become the third commonest cancer in the United States and would cost about USD 20 billion over 10 years,⁴ if the current practice remains unchanged.

There currently remains controversy regarding the appropriate management of subcentimetre (<1 cm) thyroid nodules. The British Thyroid Association (BTA) guidelines of 2014⁵ recommend FNAC of subcentimetre thyroid nodules with suspicious ultrasound features and extrathyroidal disease or high risk clinical history. Due to the indolent course of PTMCs, the American Thyroid Association (ATA) 2015 guidelines⁶ (Table 1) and American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS)⁷ recommended routine surveillance of subcentimetre (<1 cm) thyroid nodules with high suspicion ultrasound patterns as an alternative to FNAC or surgery.

Table 1.

American Thyroid Association 2015 guidelines for ultrasound risk stratification.

Risk stratification level	Ultrasound features	FNAC size cut-off (largest dimension)
High suspicion	Solid hypoechoic nodule or solid hypoechoic component of a partially cystic nodule with one or more of the following features: irregular margins (infiltrative, micro-lobulated), microcalcifications, taller-than-wide shape, rim calcifications with small extrusive soft tissue component, evidence of extra thyroidal extension	Recommend FNAC at ≥1 cm
Intermediate suspicion	Hypoechoic solid nodule with smooth margins without microcalcifications, extra thyroidal extension, or taller-than-wide shape	Recommend FNAC at ≥1 cm
Low suspicion	Isoechoic or hyperechoic solid nodule, or partially cystic nodule with eccentric solid areas, without microcalcification, irregular margin or extra thyroidal extension, or taller-than-wide shape	Recommend FNAC at ≥1.5 cm
Very low suspicion	Spongiform or partially cystic nodules without any of the sonographic features described in low, intermediate, or high suspicion patterns	Consider FNAC at ≥2 cm
Benign	Purely cystic nodules (no solid component)	No FNAC

FNAC: fine needle aspiration cytology.

Although the current BTA, ATA, and ACR TI-RADS guidelines all apply the same set of ultrasound patterns to thyroid nodules of all sizes, studies have found the ultrasound diagnostic performance to be markedly influenced by nodule size.^8–10 In fact, the individual ultrasound features predictive of malignancy also varied according to nodule size.^11,12

The aim of this study was to assess the diagnostic performance of thyroid nodule ultrasound imaging features in the detection of malignancy in subcentimetre thyroid nodules, and subsequently use multivariate analysis to guide the selection of a set of imaging features which are associated with malignancy.

Methods

Data collection

This retrospective study was approved by our Institutional Review Board, with waiver of informed consent from our patients.

All subcentimetre thyroid nodules, which had an ultrasound study and FNAC, performed at a single centre between January 2012 and December 2016 were identified. For these nodules, the patients’ demographics, thyroid ultrasound images, cytology, and surgical histopathological results were tracked using the electronic medical records and Picture Archiving and Communication System (PACS).

Ultrasound imaging and assessment

The sonographers acquired thyroid ultrasound images using 5–12 MHz high resolution linear transducers (Epiq 7, Philips Healthcare, Bothell, WA; iU22, Philips Healthcare, Bothell, WA; Aplio 500, Canon Medical Systems, Otawara, Japan). The sonographers had between 2 and 12 years of experience in thyroid ultrasound imaging.

The images from these ultrasound examinations were retrieved from PACS and reviewed. Each examination was assessed independently by two experienced radiologists (with 8 and 12 years of experience each) who were blinded to the subsequent follow-up, cytologic or surgical histopathological diagnosis. Any disagreement in the assessment was resolved by consensus.

The size of each nodule was measured in at least two dimensions. The nodules were assessed for the presence or absence of the high suspicion ultrasound features as described by the ATA 2015 guidelines (Figure 1). Each nodule was then classified according to the ATA 2015 thyroid nodule management guidelines, into ‘High suspicion’, ‘Intermediate suspicion’, ‘Low suspicion’, ‘Very low suspicion’, or ‘Benign’ based on their ultrasound patterns.

Figure 1.

Suspicious sonographic features of thyroid nodules <1 cm in diameter: (a) Solid, (b) hypoechoic, (c) irregular margin, (d) microcalcifications, (e) taller-than-wide, (f) rim calcifications with small extrusive soft tissue component, and (g) extra thyroidal extension.

Ultrasound-guided FNAC

The thyroid nodules were percutaneously aspirated under ultrasound guidance using a 21 or 23 gauge needle. The aspirate was smeared onto glass slides and fixed in 95% alcohol for cytopathologist interpretation off-site. The remaining aspirate was processed in CytoLyt (Hologic Inc., Marlborough, MA). Upon request by the radiologist, on-site cytologytechnologist assistance was available to perform adequacy assessment on Papanicolaou stained slides on nodules that were deemed technically challenging.

The Bethesda System for Reporting Thyroid Cytopathology¹³ was used to interpret and categorize the cytology results (Bethesda classification (BC) I: non-diagnostic; BC II: benign; BC III: atypia of undetermined significance/follicular lesion of undetermined significance; BC IV: follicular neoplasm/suspicious for follicular neoplasm; BC V: suspicious for malignancy; BC VI: malignant).

Statistical analysis

Continuous variables were presented either as mean (minimum–maximum) or median (interquartile range). Categorical variables were presented as frequency (percentage), and data comparisons were made using Fisher’s exact test. For the purpose of calculating the diagnostic performance, the presence of each ultrasound feature or combination of ultrasound features was compared against the final diagnosis for each thyroid nodule.

A malignant final diagnosis was defined as a malignant surgical histopathology. A benign final diagnosis was defined as either (a) benign surgical histopathology, (b) BC II cytological result for at least two FNACs on the same thyroid lesion, or (3) BC II cytological result stable on follow-up ultrasound for at least one year.⁸

Univariate analysis was performed to evaluate the sensitivity, specificity, positive predictive value, negative predictive value, odds ratio with 95% confidence interval, and p-value.

Multivariate logistic regression was performed on ultrasound features with p-value < 0.2 in univariate logistic regression. Stepwise selection was used to identify the high suspicion ultrasound features which were associated with a malignant diagnosis.

Statistical analysis was performed with the SAS 3.6 (Basic Edition) (SAS Institute Inc.) software package.

Results

A total of 469 FNACs were performed on 454 subcentimetre thyroid nodules of 413 patients. There were 342 (82.8%) female patients with a mean age of 58 years (range 15–85 years). The median nodule size was 8 mm. Out of the 454 thyroid nodules, eight could not be classified under the ATA guidelines due to ultrasound features which could not be well assessed.

A total of 38 nodules were identified as malignant from surgical histopathology, while 232 nodules were identified as benign (35 from surgical histopathology, 6 from a repeat BC II cytology following an initial BC II cytology, and 191 from stable follow-up ultrasound >1 year after an initial BC II cytology) (Figure 2). All 38 malignant nodules were identified as PTMCs in histology. The FNAC results and final diagnoses of the nodules are shown in Tables 2 and 3, respectively.

Figure 2.

Study population. BC: Bethesda classification; FNAC: fine needle aspiration cytology.

Table 2.

FNAC result (Bethesda) in each American Thyroid Association (ATA) sonographic pattern group.

	ATA sonographic pattern
BethesdaClassification	High suspicion	Intermediate suspicion	Low suspicion	Very low suspicion	Benign	Total
I	10	27	22	3	0	62
II	40	98	161	22	1	322
III	7	5	6	0	0	18
IV	3	4	3	0	0	10
V	4	4	1	0	0	9
VI	11	10	4	0	0	25
Total	75	148	197	25	1	446

FNAC: fine needle aspiration cytology.

Table 3.

Final diagnoses in each American Thyroid Association (ATA) sonographic pattern group.

	ATA ultrasound pattern
	High suspicion	Intermediate suspicion	Low suspicion	Very low suspicion	Benign	Total
Surgery (%)	26 (55.3)	21 (25.3)	24 (18.9)	2 (16.7)	0 (0.0)	73 (27.0)
Malignant	19	13	5	1	0	38
Benign	7	8	19	1	0	35
Stable follow-up (%)	20 (42.6)	59 (71.1)	101 (79.5)	10 (83.3)	1 (100.0)	191 (70.7)
Repeated BC II (%)	1 (2.1)	3 (3.6)	2 (1.6)	0 (0.0)	0 (0.0)	6 (2.2)
Total	47	83	127	12	1	270

BC: Bethesda classification.

The diagnostic performances of the individual ultrasound features are summarized in Table 4. Amongst high suspicion sonographic features, the solid and hypoechoic features had a high sensitivity but poor specificity, while irregular margins, microcalcifications, taller-than-wide, rim calcifications, and extrathyroidal extension show high specificity but poor sensitivity.

Table 4.

Diagnostic performance of suspicious ultrasound features and American Thyroid Association (ATA) high suspicion ultrasound pattern using univariate and multivariate analysis.

Ultrasound feature	Sensitivity, % (95% CI) (N=38)	Specificity, % (95% CI) (N=232)	PPV, % (95% CI)	NPV, % (95% CI)	PLR (95% CI)	NLR (95% CI)	OR^a (95% CI)	p-value
Univariate
Solid/solid component of partially cystic nodule	97.4 (86.2–99.9)	4.3 (2.1–7.8)	14.3 (13.6–15.0)	90.9 (56.9–98.7)	1.02 (0.96–1.08)	0.61 (0.08–4.63)	1.18 (0.19–7.28)	1.0000
Hypoechoic	81.6 (65.7–92.3)	50.0 (43.4–56.6)	21.1 (18.0–24.6)	94.3 (89.3–97.0)	1.63 (1.34–1.99)	0.37 (0.19–0.73)	4.20 (1.81–9.73)	0.0003
Irregular margins	34.2 (19.6–51.4)	92.2 (88.0–95.3)	41.9 (27.9–57.5)	89.5 (87.2–91.5)	4.41 (2.36–8.24)	0.71 (0.57–0.90)	6.14 (2.70–13.98)	<0.0001
Microcalcifications	23.7 (11.4–40.2)	91.0 (86.5–94.3)	30.0 (17.5–46.4)	87.9 (85.8–89.7)	2.62 (1.30–5.28)	0.84 (0.70–1.01)	3.17 (1.33–7.54)	0.0211
Taller-than-wide	23.7 (11.4–40.2)	92.2 (88.0–95.3)	33.3 (19.5 50.7)	88.1 (86.0–89.8)	3.05 (1.48–6.29)	0.83 (0.69–0.99)	3.73 (1.54–9.05)	0.0060
Rim calcifications with small extrusive soft tissue component	0.0 (0.0–9.3)	99.6 (97.6–100.0)	0.0 (–)	85.9 (85.7–86.0)	0.00 (–)	1.00 (1.00–1.01)	2.04 (0.02–187.29)	1.0000
Extra thyroidal extension	0.0 (0.0–9.3)	99.1 (96.9–99.9)	0.0 (–)	85.8 (85.7–86.0)	0.00 (–)	1.01 (1.00–1.02)	1.20 (0.03–50.14)	1.0000
ATA High suspicion	50.0 (33.4–66.7)	84.5 (79.2–88.9)	34.6 (25.4–45.0)	91.2 (88.2–93.4)	3.22 (2.08–4.99)	0.59 (0.43–0.82)	5.38 (2.61–11.11)	–
Multivariate
Hypoechoic	–	–	–	–	–	–	3.19 (1.30–7.85)	0.0113
Irregular margins	–	–	–	–	–	–	4.17 (1.76–9.86)	0.0012

CI: confidence interval; NLR: negative likelihood ratio; NPV: negative predictive value; OR: odds ratio; PLR: positive likelihood ratio; PPV: positive predictive value.

Unadjusted odds ratio for univariate analysis. Adjusted odds ratio for multivariate analysis

Univariate logistic analysis for these individual ultrasound features showed that hypoechoic nodule, irregular margins, microcalcifications, and taller-than-wide were significantly associated with a malignant diagnosis.

Amongst these features, subsequent multivariate analysis (Table 4) using stepwise selection identified that hypoechoic and irregular margins were significantly associated with a malignant outcome.

The ATA high suspicion pattern had a sensitivity of 50.0% and specificity of 84.5%. The malignancy rates corresponding to each ATA ultrasound risk pattern were 0.0, 8.3, 3.9, 15.7, and 40.4% for benign, low, very low, intermediate, and high suspicion patterns, respectively.

Discussion

In our univariate analysis, we found that the ultrasonic features of solid, hypoechoic, irregular margins, microcalcifications, and taller-than-wide were significant predictors of malignancy in subcentimetre thyroid nodules. Features such as interrupted rim calcifications and extrathyroidal extension were poor predictors of malignancy in subcentimetre thyroid nodules. These features may be rarely observed in our cohort due to the small size of subcentimetre nodules, in addition to technical challenges in establishing the features reliably. For example, only nodules in subcapsular locations can potentially exhibit extrathyroidal extension. Even for subcapsular nodules, the distance of extension would be small given our maximum nodule size of 1 cm, which may be further reduced by compression of the thyroid during ultrasound examination.

Remonti et al.¹⁴ reported similar findings in a meta-analysis of all nodule sizes. Using univariate analysis, they found irregular margins, microcalcifications, taller-than-wide, and absence of elasticity to be associated with a higher risk of malignancy. However, they also noted that none of the individual features had a clinically relevant positive likelihood ratio (>10). Sharma et al.¹⁵ found the most suspicious ultrasound features in subcentimetre thyroid nodules to be posterior acoustic shadowing, rim calcifications, entirely solid structure, many diffuse calcifications, and taller-than-wide configuration. Univariate analysis by Berker et al.¹¹ showed hypo-echogenicity, microcalcifications, and a relatively round shape (long axis/short axis < 1.5) to be associated with malignancy in subcentimetre nodules. Univariate logistic regression analysis by Papini et al.¹⁶ found only blurred margins, intranodular vascularization, and microcalcifications, but not hypo-echogenicity, to be independent risk factors significantly associated with malignancy amongst nodules 8–15 mm.

Our results show that the presence of a solid nodule was 97.4% sensitive for malignancy, but the specificity was only 4.3%. Similarly, a hypoechoic appearance on ultrasound was also sensitive (81.6%) but not specific (50.0%) for malignancy. The poor specificity can be attributed to the high incidence of solid nodules and hypoechoic appearance amongst both malignant and benign lesions. These results are in agreement with Berker et al.¹¹ (solid: sensitivity 100.0%, specificity 7.6%; hypoechoic: sensitivity 72.7%, specificity 51.6%) and Sharma et al.¹⁵ (solid: sensitivity 100.0%, specificity 31.7%; hypoechoic: sensitivity 93.8%, specificity 21.8%) for subcentimetre nodules, as well as Papini et al.¹⁶ (hypoechoic: sensitivity 87.1%, specificity 43.4%) for nodules 8–15 mm in size.

Recent studies have found osteopontin to be overexpressed in papillary thyroid carcinomas¹⁷ and are significantly related to the formation of psammoma bodies and microcalcifications.^18,19 Histopathologically, microcalcifications seen on ultrasound are due to psammoma bodies, which are possibly formed through the extra-cellular accumulation and calcification of precursor substances released by neoplastic cells in papillary thyroid carcinomas.²⁰ Osteopontin expression has also been associated with increased rates of lymph node metastasis.^17–19 Oh et al.²¹ found calcifications in PTMCs to be associated with a higher rate of lymph node metastasis (38.6% versus 23.2%, p < 0.05). However, Moon et al.⁹ found microcalcifications to be less diagnostic of malignancy in subcentimetre nodules (36.6% sensitivity; 89.7% specificity) compared to supracentimetre nodules (51.4% sensitivity; 91.6% specificity). Papini et al.¹⁶ showed that amongst nodules 8–15 mm in size, microcalcifications had a sensitivity of 29.0% and specificity of 95.0%. Our results are similar, with a sensitivity of 23.7% and specificity of 91% for microcalcifications in subcentimetre nodules. In addition, microcalcifications were not identified as a significant feature in our multivariate analysis.

In the multivariate analysis on nodules <15 mm by Lyshchik et al.,¹² irregular margins, subcapsular location, and type III vascularization were reported to be predictive of malignancy. They also found that hypoechoic appearance was a useful predictor in nodules >15 mm but not <15 mm. Berker et al.¹¹ found that hypoechoic nodules with microcalcifications or a relatively round shape (long axis/short axis < 1.5) were good predictors of malignancy in subcentimetre nodules. By pairing hypo-echogenicity with various other ultrasound features, Papini et al.¹⁶ showed that a combination of hypo-echogenicity and blurred margins has a high sensitivity and specificity of 74.2 and 87.8%. Our multivariate analysis showed that only a hypoechoic appearance and irregular margins were significantly associated with malignancy after adjusting for confounders. These differences in the suspicious ultrasound features between subcentimetre and supracentimetre nodules may have contributed to the higher malignancy rates in subcentimetre nodules reported in previous studies.^11,22,23

Given that only hypo-echogenicity and irregular margins were identified as significant features and features of interrupted rim calcification and extrathyroidal extension were rarely observed in our study, subcentimetre nodules may be biased towards a less suspicious classification when evaluated using scoring based systems such as ACR TI-RADS. Subsequently, the subcentimetre nodules for which additional workup may be warranted may not be investigated further under the current ACR TI-RADS guidelines.

The diagnostic performance of the ATA high suspicion ultrasound pattern in subcentimetre thyroid nodules has been reported by Xu et al.⁸ to have a sensitivity of 80.5% and a specificity of 63.7%. Using similar diagnostic criteria, Moon et al.²⁴ reported a sensitivity of 91.4% and a specificity of 67.2% amongst subcentimetre thyroid nodules. However, our results revealed a lower sensitivity (50.0%) but higher specificity (84.5%) using the ATA high suspicion ultrasound pattern.

The malignancy rates of benign, very low, low, intermediate, and high suspicion ultrasound patterns in the ATA guidelines were estimated to be <1, <3, 5–10, 10–20, and >70–90%, respectively. In our study, the corresponding malignancy rates were 0.0, 8.3, 3.9, 15.7, and 40.4%. The malignancy rate of the high suspicion ultrasound patterns in our subcentimetre nodules was lower compared to the estimated risk of malignancy by ATA.

The FNAC rate of low suspicion ATA pattern nodules was relatively high in our study (44.2%). This may be due to majority of the FNACs being performed before the ATA 2015 guidelines were released. Hence the threshold for sampling subcentimetre nodules was lower. In addition, patients’ concerns and anxiety about the lesions, despite being subcentimetre, may have driven these lesions to be taken up for intervention despite recommendations. This observation of patients’ preference for intervention is also evident in previous studies such as Ito et al.²⁵ and Chua et al.²² where up to 80 and 76.6% of patients opted for surgical intervention despite their lack of highly suspicious features or that current guidelines recommended against aggressive intervention of these subcentimetre lesions.

Cytotechnologists were present in 9.7% of FNACs with a diagnostic yield of 91.6% compared to 85.5% without cytotechnologists. Our centre achieved an overall diagnostic yield of 86.1% amongst subcentimetre FNAC. This yield is similar to the results achieved by Bo et al.²³ (80.7%) and Berker et al.¹¹ (83.4%) amongst nodules <1 cm.

The optimal management of PTMCs is still an issue of contention, with recent longitudinal studies indicating that ultrasound surveillance is a viable option due to the indolent nature of these malignancies. The latest BTA, ATA, and ACR TI-RADS guidelines also vary in their recommended management of subcentimetre thyroid nodules. An understanding of the natural history of PTMCs, together with a cost-effectiveness study would be required to further evaluate the various proposed management strategies. Although our study does not directly address the management of PTMCs, we note based on our results that current ultrasound imaging criteria may be biased in the evaluation of subcentimetre thyroid nodules, and fail to identify malignant nodules which warrant further follow-up.

There are several limitations to our study. First, this was a retrospective study. Our study defined the final diagnosis based on a combination of ultrasound follow-up and repeat cytology results in addition to surgical histopathology. Nodules with BC II cytology without surgery or follow-up were excluded from the study. Nodules with other cytology results were also excluded if they did not have subsequent surgical histopathology results. Selection bias cannot be excluded due to the exclusion criteria. Second, inter-observer variability of ultrasound interpretation between the readers cannot be excluded. Third, all interpretations were based on static ultrasound images that were captured previously, which may result in misclassification of the thyroid nodules.

In conclusion, our results indicate that amongst the ATA high suspicion ultrasound features, the presence of both hypo-echogenicity and irregular margins was significantly associated with malignancy in subcentimetre thyroid nodules. Hence, subcentimetre thyroid nodules which are hypoechoic with irregular margins may require follow-up either through FNAC or routine ultrasound surveillance. Subcentimetre nodules for which additional workup may be warranted may not be investigated further under the current ACR TI-RADS guidelines.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

Ethics approval

SingHealth Centralised Institutional Review Board (Institution: SingHealth, reference number: 2014/238/C).

Guarantor

KD.

Contributors

KD, BST, JMEC, SKS, NV, CLH, RAB, and PRS were involved in protocol development, gaining ethical approval. Ultrasound image interpretation was performed by SKS and PRS. RG and TKW performed the data analysis. TKW wrote the first draft of the manuscript. TKW, KD, PRS, and SKS wrote the final version of the manuscript. All authors reviewed and edited the manuscript and approved the final version of the manuscript.

Acknowledgments

We are grateful to the sonographers and interventional radiologists from Singapore General Hospital for performing the thyroid ultrasound and fine needle aspiration cytology.

ORCID iDs

Tian K Woon

Karthikeyan Damodharan

References

Tan

Gharib

Thyroid incidentalomas: management approaches to nonpalpable nodules discovered incidentally on thyroid imaging.

Ann Intern Med 1997; 126: 226–231.

Ahn

Kim

Welch

HG.

Korea’s thyroid-cancer “epidemic” – screening and overdiagnosis. N Engl J Med 2014; 371: 1765–1767.

Davies

Welch

HG.

Current thyroid cancer trends in the United States.

JAMA Otolaryngol Head Neck Surg 2014; 140: 317–322.

Aschebrook-Kilfoy

Schechter

Shih

, et al. The clinical and economic burden of a sustained increase in thyroid cancer incidence. Cancer Epidemiol Biomarkers Prev 2013; 22: 1252–1259.

Perros

Boelaert

Colley

, et al. Guidelines for the management of thyroid cancer. Clin Endocrinol 2014; 81: 1–122.

Haugen

Alexander

Bible

, 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: 1–133.

Tessler

Middleton

Grant

, et al. ACR thyroid imaging, reporting and data system (TI-RADS): white paper of the ACR TI-RADS committee. J Am Coll Radiol 2017; 14: 587–595.

, et al. Thyroid nodule sizes influence the diagnostic performance of ACR TI-RADS and ultrasound patterns of 2015 ATA guidelines: a multicenter retrospective study. Sci Rep 2017; 7: 43183.

Moon

Jung

Lee

, et al. Benign and malignant thyroid nodules: US differentiation–multicenter retrospective study. Radiology 2008; 247: 762–770.

10.

Baek

, et al. Diagnostic performance of practice guidelines for thyroid nodules: thyroid nodule size versus biopsy rates. Radiology 2019; 291: 92–99.

11.

Berker

Aydin

Ustun

, et al. The value of fine-needle aspiration biopsy in subcentimeter thyroid nodules. Thyroid 2008; 18: 603–608.

12.

Lyshchik

Drozd

Demidchik

, et al. Diagnosis of thyroid cancer in children: value of gray-scale and power Doppler US. Radiology 2005; 235: 604–613.

13.

Cibas

Ali

SZ.

The Bethesda system for reporting thyroid cytopathology.

Thyroid 2009; 19: 1159–1165.

14.

Remonti

Kramer

Leitao

, et al. Thyroid ultrasound features and risk of carcinoma: a systematic review and meta-analysis of observational studies. Thyroid 2015; 25: 538–550.

15.

Sharma

Gabriel

Nemcek

, et al. Subcentimeter thyroid nodules: utility of sonographic characterization and ultrasound-guided needle biopsy. AJR Am J Roentgenol 2011; 197: W1123–W1128.

16.

Papini

Guglielmi

Bianchini

, 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.

17.

Guarino

Faviana

Salvatore

, et al. Osteopontin is overexpressed in human papillary thyroid carcinomas and enhances thyroid carcinoma cell invasiveness. J Clin Endocrinol Metab 2005; 90: 5270–5278.

18.

Tunio

Hirota

Nomura

, et al. Possible relation of osteopontin to development of psammoma bodies in human papillary thyroid cancer. Arch Pathol Lab Med 1998; 122: 1087–1090.

19.

Sun

Fang

Dong

, et al. Correlation between osteopontin messenger RNA expression and microcalcification shown on sonography in papillary thyroid carcinoma. J Ultrasound Med 2011; 30: 765–771.

20.

Das

Sheikh

George

, et al. Papillary thyroid carcinoma: evidence for intracytoplasmic formation of precursor substance for calcification and its release from well-preserved neoplastic cells. Diagn Cytopathol 2008; 36: 809–812.

21.

Chung

Song

, et al. The pattern and significance of the calcifications of papillary thyroid microcarcinoma presented in preoperative neck ultrasonography. Ann Surg Treat Res 2014; 86: 115–121.

22.

Chua

Tang

Lim

, et al. Should we perform fine needle aspiration cytology of subcentimetre thyroid nodules? A retrospective review of local practice. Ultrasound 2019; 27: 64–68.

23.

Ahn

Lee

, et al. Malignancy rate in sonographically suspicious thyroid nodules of less than a centimeter in size does not decrease with decreasing size. J Korean Med Sci 2011; 26: 237–242.

24.

Moon

Son

Kim

, et al. The diagnostic values of ultrasound and ultrasound-guided fine needle aspiration in subcentimeter-sized thyroid nodules. Ann Surg Oncol 2012; 19: 52–59.

25.

Ito

Uruno

Nakano

, et al. An observation trial without surgical treatment in patients with papillary microcarcinoma of the thyroid. Thyroid 2003; 13: 381–387.