The Large Majority of 1520 Patients With Indeterminate Thyroid Nodule at Cytology Have a Favorable Outcome, and a Clinical Risk Score Has a High Negative Predictive Value for a More Cumbersome Cancer Disease

Context:

Clinical management of patients with thyroid nodules indeterminate at fine-needle aspiration (FNA) cytology is still unsettled.

Objective:

Our objective was to establish the clinical outcome of patients with thyroid nodules indeterminate at cytology and to identify the features associated with malignancy.

Design and Patients:

This was a retrospective evaluation of 1520 consecutive patients with indeterminate cytology among 100 065 patients who underwent FNA between January 2000 and December 2010.

Results:

Of 1520 patients, 371 (24.4 %) had thyroid cancer at histology, the follicular variant of papillary cancer being the most frequent histotype, and 342 patients with cancer were free of disease after thyroidectomy and ¹³¹I remnant ablation, whereas 29 needed further treatment because of persistent disease. Among them, only 12 had persistence of disease at the end of follow-up. Atypias at cytology (P = .001), blurred nodule margins (P = .005), and spot microcalcifications (P = .003) at thyroid ultrasound (US) were significantly associated with malignancy. A clinical score including cytology and US characteristics was calculated; the lowest value showed a high negative predictive value (83.9%) for the presence of malignancy and even higher (99.5%) for the presence of a more cumbersome cancer disease, and only 4 of the 29 patients who needed further treatment were included in the group with the lowest risk score.

Conclusions:

Patients with Thy 3 cytology and histology of thyroid cancer had an overall good prognosis. A clinical risk score including the results of cytology and US features is helpful in the management of patients with indeterminate thyroid nodules.

Thyroid nodules are common, particularly in iodine-deficient areas, being palpable in 5% and detectable by thyroid ultrasound (US) in up to 50% of subjects (1–3). Only a minority of thyroid nodules are malignant, and few cause significant health problems. Cytological examination of material obtained by fine-needle aspiration (FNA) is the best single test for differentiating malignant from benign thyroid nodules. However, in 10% to 15% of cases, cytology is not diagnostic; ie, the material is insufficient for diagnosis, and in 5% to 20% (4), it is not possible to discriminate between benign, ie, follicular adenoma, and malignant, ie, follicular variant of papillary thyroid cancer or follicular thyroid cancer (5, 6) (indeterminate cytology). About 20% of nodules with indeterminate cytology are malignant (7, 8).

Clinical data are of little help in predicting malignancy in patients with nodules with indeterminate cytology (7, 8). Similarly, the echographic pattern was reported to be poorly predictive, and only spot microcalcifications were found to be significantly associated with thyroid cancer (7–11). US elastography (USE) has been proposed as a useful tool to differentiate malignant from benign lesions (12–14), but still needs validation. For this reason, surgery is currently recommended in indeterminate nodules to obtain a conclusive diagnosis by histology (15, 16). No study has evaluated the outcome of patients operated on for indeterminate cytology who have a thyroid cancer at histology. Thus, the overall clinical impact of carrying a thyroid nodule with indeterminate cytology is unsettled. To address this point, we analyzed the clinical data of 1520 patients, consecutively observed at our Institution who underwent surgery for nodules with indeterminate cytology (Thy 3), according to the classification of the British Thyroid Association (17), and 371 patients were found to harbor a thyroid cancer at histology. We found that only 29 of 371 patients (7.8%) required further treatment after thyroidectomy and ablative ¹³¹I administration; 12 patients (only 0.79% of the whole series) had a persistent disease at follow-up. Most of them had a less aggressive type of thyroid cancer (follicular variant of papillary thyroid cancer and minimally invasive follicular cancer). A clinical risk score was developed that has a high negative predictive value (NPV) for the chance of carrying a cumbersome cancer disease.

Patients and Methods

Patients

The indication for FNA was a thyroid nodule that was cold at scintiscan and greater than 1 cm or smaller than 1 cm in the presence of suspicious clinical and/or echographic signs (15, 16, 18–20). In multinodular glands, only the dominant nodules and those with suspicious US features underwent FNA. Among the 100 065 patients who underwent FNA at our institution from January 2000 to December 2010, 5446 patients (5.4%) had a Thy 3 cytology; 2746 of these were referred by other endocrinologists only for FNA. The remaining 2700 patients had full evaluation at our institution. Of these, 574 underwent surgery in other institutions. Of the remaining 2126 patients, surgery was not advised in 606 due to 1) small size and no suspicious clinical or US features (n = 220), 2) elderly patients with associated diseases contraindicating surgery (n = 112), or 3) high elasticity score at USE (n = 168), and 106 patients refused thyroidectomy. Thus, our study included 1520 consecutive patients (mean age 46.1 ± 13.2 years) who underwent thyroidectomy and had histological examination at our university hospital.

Thyroid function tests

Serum TSH was measured by a sensitive immunoradiometric assay (Delphia; Pharmacia) (normal values 0.4–3.4 μU/mL). Serum thyroglobulin (Tg) was measured by Immulite 2000 (Diagnostic Products Corp) (normal values 0.73–84 ng/mL) or by ultrasensitive Tg Access immunoassay systems (Beckman Coulter, Inc) (normal values 1.15–130.77 ng/mL). Tg autoantibodies (TgAbs) were measured by Immuno Enzymometric assay (AIA-Pack TgAb; Tosoh) (positive values >30 U/mL).

Thyroid US and USE

Conventional US was performed using a real-time instrument (Esaote SPA, My Lab 70 machine with a linear transducer of 8–13 MHz). The following US parameters were evaluated: nodule echogenicity (hyper-, iso-, or hypoechogenicity with respect to the normal thyroid parenchyma), blurred margins, and spot microcalcifications (hyperechoic spots <2 mm, without acoustic shadowing). Qualitative US elastography was performed as previously described (12), using the same real-time instrument and the same probe as above.

FNA cytology and histology

FNA was performed under US guidance by using a 23-gauge needle attached to a 10-mL syringe. The material was air-dried, stained with Papanicolaou and Giemsa and interpreted by two experienced cytologists (DG, MI). The adequacy of aspirates was defined according to the guidelines of the Papanicolaou Society (21). Results of cytology were reported according to the British Thyroid Association (8), which subdivides the Thy 3 category into 2 subclasses, Thy 3A and Thy 3B, corresponding to the Thy III and Thy IV of the Bethesda classification system (22–24). Thy 3A includes specimens characterized by increased cellularity with a small uniform follicular organization and scant watery colloid, follicular cells with small nuclei, and regularly distributed chromatin. Thy 3B has a higher risk of malignancy, with features not sufficient for inclusion in the Thy 4 category: hypercellularity with absence of colloid, well-defined nests, rarely with overlapping nuclei, follicular cells with an enlarged nuclei, sparse or irregularly distributed chromatin, and rare prominent nucleoli and occasional mitoses. Formalin-fixed, paraffin-embedded tissue obtained after thyroid surgery was stained by hematoxylin and eosin. The histological diagnosis was made by the same pathologist (F.B.) according to the World Health Organization guidelines (25).

Treatment and follow-up of cancer patients

The follow-up was 1 to 11 years (mean 3.1 ± 1.2 years) and included visits at 3, 6, and 12 months after surgery and every 12 months thereafter. At each visit, serum Tg, TgAb, TSH, free T₄, and free T₃ were measured. l-T₄ treatment was titrated to obtain TSH suppression with normal free T₄ and free T₃ until clinical remission of cancer disease was ascertained, and low-normal TSH thereafter. Structural evidence of disease was assessed by neck US. In patients with suspicious lymph nodes or recurrence in the thyroid bed, FNA with measurement of Tg in the washout liquid from the needle was performed. When necessary, the workup included radionuclide scan, computed tomography scan, magnetic resonance, and fluorodeoxyglucose-positron emission tomography.

Patients were considered free of disease if basal serum Tg was <1 ng/mL or stimulated Tg was <2 ng/mL, with negative TgAb (26, 27), and when no morphological evidence of disease was detectable.

Patients were classified as having persistent disease, either biochemical (basal Tg values >1 ng/mL with the Access immunoassay system or stimulated Tg values >2 ng/mL with the Immulite 2000) or structural, when detectable Tg was associated with the presence of remnant tissue (at US or whole-body scan) or local recurrence, metastatic lymph nodes, or distant metastases (28) (Figure 1).

Figure 1.

Follow-up of patients with indeterminate nodule at cytology and thyroid cancer at histology.

Statistical analysis

Proportions and frequencies calculated in different groups of patients were compared using the χ² test and the Fisher exact test, when appropriate. The Kolmogorov-Smirnov test was employed to test the normality of data distribution. Student's t test and Mann-Whitney U test were used to detect differences between groups for normal and skewed variables, respectively. Associations between clinical parameters and malignancy as determined at histology were assessed by univariate and multivariate logistic regression analyses. The Wald test was used to assess the statistical significance of each parameter in the logistic model. A clinical score was computed based on the results of the multivariate logistic model, and the associated risk probability was derived by applying the logistic formula to the calculated risk score: probability = 1/(1 + exp-x). Statistical analyses were performed using the SPSS for Windows version 10.0 software.

Results

Histology

Of 1520 patients, 882 underwent total thyroidectomy at the first operation; 579 had histology of benignity and 303 of malignancy. The remaining 638 patients underwent lobectomy; in 68, completion thyroidectomy was performed for the presence of carcinoma on histology. Thus, overall malignancy was found at histology in 371 of 1520 patients (24.4%): 308 of 371 patients (83.0%) had a papillary thyroid cancer, the majority of which were follicular variant; 6 (1.6%) a follicular cancer; 49 (13.2%) a minimally invasive cancer; and 8 (2.2%) a poorly differentiated cancer (Table 1).

Table 1.

Histology of 1520 Indeterminate Nodules According to the Absence or Presence of Atypias

Carcinoma	Thy 3		Thy 3A, n	Thy 3B, n
	n	%
Papillary	308	83.0	190	118
Follicular	239		155	84
Classic	47		26	21
Oxyphilic	6		2	4
Tall cell	10		6	4
Trabecular	6		1	5
Follicular	55	14.8	32	23
Classic	6		3	3
Minimally invasive	49		29	20
Poorly differentiated	8	2.2	5	3
Benign
Follicular adenoma	899	78.2	753	146
Hyperplastic nodule	138	12	101	37
Lymphocytic Thyroiditis	9	0.8	8	1
Adenoma with foci of papillary carcinoma	103	9	81	22

Histological diagnosis of malignancy was made in 227 of 1170 (19.4%) Thy 3A nodules and in 144 of 350 (41.1%) Thy 3B nodules (P < .0001). Among Thy 3A nodules, 190 were papillary cancer (155 follicular, 26 classic, 2 oxyphilic, 6 tall cell, and 1 trabecular), 3 were follicular cancer, 29 were minimally invasive follicular cancer, and 5 were poorly differentiated cancer. Among Thy 3B nodules, 118 were papillary cancer (84 follicular, 21 classic, 4 oxyphilic, 4 tall cell, and 5 trabecular), 3 were follicular cancer, 20 were minimally invasive follicular cancer, and 3 were poorly differentiated cancer (Table 1).

At histology, 1149 of 1520 indeterminate nodules (75.6%) were benign, the large majority (899 of 1149, 78.2%) being follicular adenomas, 138 (12.0%) hyperplastic nodules, 9 (0.8%) lymphocytic thyroiditis, and 103 (9.0%) follicular adenomas with focus <5 mm of the classic variant of papillary cancer (Table 1).

US elastography

Among the 337 patients who underwent USE, surgery was not advised for 168 because of a high elasticity score and no suspicious features at conventional US. Among the 169 in whom histology was available, a score of 1 describing high elasticity was found in 124 of 138 benign nodules (89.8%) and in only 4 of 31 (12.9%) with a final diagnosis of cancer. Thus, high elasticity at USE was strongly associated with a histological result of benignity (P < .0001), resulting in a high NPV (96.9%).

Clinical parameters, US and cytological features, and probability of thyroid cancer: the risk score

At US, no correlation with malignancy was found for male gender (P = .9), single vs multiple nodularity (P = .8), thyroid volume <20 vs >20 mL (20 mL being the median thyroid volume), nodular volume <3.6 vs >3.6 mL (3.6 mL being the median nodular volume) (P = .8). Age <40 vs >40 years, although associated with malignancy, did not reach the statistical significance. Blurred nodule margins and spot microcalcifications were the US features more significantly associated with malignancy (P = .005 and .003, respectively). Nodule hypoechogenicity, although associated with malignancy, did not reach statistical significance (P = .18) (Table 2). The coexistence of 2 or 3 US features increased the US specificity, decreasing its sensitivity (Table 2).

Table 2.

Echographic Patterns in Indeterminate Nodules According to Histology

	Benign (n = 1149)	Carcinoma (n = 371)	P	Sensitivity, %	Specificity, %
H
No	469	166	.18	55.25	40.81
Yes	680	205
BM
No	873	255	.005	31.26	75.97
Yes	276	116
SM
No	857	255	.003	31.26	74.58
Yes	292	116
H+BM
No	968	293	.01	21.02	84.24
Yes	181	78
H+SM
No	983	256	<.001	30.99	85.55
Yes	166	115
SM+BM
No	1057	322	.002	13.20	91.99
Yes	92	49
H+BM+SM
No	1085	331	<.001	10.78	94.42
Yes	64	40

Abbreviations: BM, blurred margins; H, hypoechogenicity; SM, spot microcalcifications.

By univariate analysis, the presence of Thy 3B cytology, the coexistence of the 3 US features, and age <40 years were significantly associated with cancer (relative risk from 0.6–3.18), whereas nodular volume and gender did not confer a significant risk (Table 3). The relative risk conferred by the Thy 3B cytology and these 3 US features were confirmed at multivariate analysis (Table 3). Therefore, we calculated the probability of cancer according to the formula: P = 1/(1 + exp-x), with x representing the following score: −1.58 + 1.02 (when Thy 3B cytology was present) + 0.66 (when the 3 US features were present) + 0.46 (when age <40 years). Therefore, in absence of the 3 parameters, x = −1.58 and P = .17, whereas in presence of the all 3 parameters, x = 0.56 and P = .64. A total of 735 patients had a risk score of −1.58; this lowest-risk subgroup included 118 patients with cancer. This risk score had a sensitivity of 68.2%, a specificity of 53.7%, and an NPV of 83.9% for the assessment of malignancy (Figure 2).

Figure 2.

Predicted probability of cancer in patients with indeterminate nodule according to the clinical risk score.

Table 3.

Univariate and Multivariate Analysis of Factors Conferring a Risk for Carcinoma in Patients With Indeterminate Nodule

	RR	OR	P
Univariate
Age <40 y	0.6	0.49–0.80	<.001
Gender (F)	1.0	0.76–1.33	.9
3 US featuresa	2.1	1.36–3.25	.001
Thy 3Bb	3.18	2.45–4.12	<.001
NV >3.6 mL	0.9	0.72–1.15	.45
Multivariate
Age <40 y	1.58	1.22–2.06	.001
Gender (F)	1.01	0.74–1.37	.95
3 US features*	1.93	1.23–3.05	.005
Thy 3Bb	2.78	2.11–3.68	<.001
NV >3.6 mL	0.94	0.73–1.22	.64

Abbreviations: F, female; NV, nodular volume; OR, odds ratio; RR, relative risk.

^a

Hypoechogenicity, blurred margins, and spot microcalcifications.

^b

Presence of atypias at cytology.

Clinical outcome of cancer patients and risk score

The stage of tumor at surgery was T1 in 171 o 371 (46.0%), N0 in 356 of 371, (96.0%), and M0 in 368 of 371 (99.1%). Ten patients were not treated with radioiodine, according to the recent guidelines (15, 16, 28), and 332 were free of disease after ¹³¹I remnant ablation. Twenty-nine patients had more cumbersome cancer disease and required further treatment (Figure 1): 18 received 2 and 11 received 3 or more ¹³¹I activities. Of the 18 patients who received 2 activities, 16 were in remission (14 had been treated for remnant tissue and 2 for lymph node metastases); 1 had biochemical and 1 structural persistent disease. Of the 11 patients who received multiple ¹³¹I activities, 1 (who had been treated for remnant tissue) was in remission and 6 had biochemical and 4 structural persistent disease. Therefore, at the end of the follow-up, 359 of 371 patients (96.8%) had no evidence of disease, whereas 7 had biochemical and 5 structural persistent disease (Table 4). The risk score of the 7 patients with biochemical persistent disease ranged from −1.58 to 0.09, with 2 cases in the lowest-risk group. The risk score of the 5 patients with structural persistent cancer disease ranged from −1.28 to −0.46, with 2 cases in the lowest-risk group. In summary, the lowest-risk group included only 4 of the 29 patients who needed further treatments after thyroidectomy and ¹³¹I remnant ablation and 2 of those with structural persistent disease. Thus, the value of −1.58 for the risk score had an NPV of 99.7% for the assessment of the risk of a cumbersome cancer disease.

Table 4.

Clinical Characteristics and Outcome in Patients With Biochemical (A) or Structural (B) Persistent Disease

Patient	Biochemical							Structural
	1	2	3	4	5	6	7	8	9	10	11	12
Age, y	37	46	44	62	18	40	43	62	55	20	64	67
Gender	M	F	F	M	M	M	F	M	F	F	F	M
NV (mL)	6.93	43	0.93	5.7	47	5	4.3	54	25	6	19	53
H	Yes	Yes	Yes	Yes	No	No	Yes	Yes	Yes	Yes	Yes	Yes
BM	Yes	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	Yes	Yes
SM	No	Yes	Yes	Yes	No	No	No	No	Yes	Yes	Yes	No
Cytology Thy	3B	3A	3B	3A	3A	3A	3A	3A	3A	3A	3A	3B
Histotype	TRV	FV	FV	FC	MIFC	FV	TRV	PDC	FV	FV	MIFC	PDC
T	T2	T3	T1m	T3	T3	T3	T1	T4	T3	T3	T3	T3
131I	230	420	230	210	230	130	260	500	1210	1300	700	180 + RT
Tg	5.1a	2.2a	5.3b	14a	1.6b	1.5b	3.2b	1624a	104a	105a	21 790a	996a
TgAb	Neg	Neg	Neg	Neg	Neg	Neg	Neg	Neg	Neg	Neg	Neg	148
Metastasis/remnant	No	No	No	No	No	No	No	Lung	Lung, bone	LN, lung	LN, lung, bone, liver	Remnant
Risk score	−0.10	−1.58	0.09	−0.90	−1.12	−1.12	−1.58	−1.58	−1.58	−0.46	−0.93	−0.56

Abbreviations: BM, blurred margins; F, female; FC, follicular carcinoma; FV, follicular volume; H, hypoechogenicity; LN, lymph nodes; M, male; MIFC, minimally invasive follicular carcinoma; Neg, negative; NV, nodular volume; PDC, poorly differentiated carcinoma; RT, radiotherapy; SM, spot microcalcifications; T, tumor size stadiation; TRV, trabecular variant of papillary carcinoma.

^a

Ultrasensitive.

^b

After recombinant human TSH.

Discussion

Thyroid FNA has been established as the main tool to identify malignant thyroid nodules. It has a high sensitivity, specificity, and accuracy. Its main limitations are nondiagnostic and indeterminate (follicular) specimens, the latter accounting for 5% to 20% of all results (7, 8, 29). In the present as well as in previous reports from our group (7, 8, 29), the prevalence of indeterminate cytology is rather low. This may be explained by the fact that in our series, the vast majority of Thy 3 nodules with a histological diagnosis of benignity were follicular adenomas, whereas hyperplastic nodules accounted for only 12%, being correctly classified as benign (Thy 2). The strict cytological criteria used for indeterminate cytology explain the relatively high rate of malignancy (24.4%) observed in the whole series of patients with Thy 3 cytology as well as in the Thy 3A (19.0%) and Thy 3B (41.0%) subclasses.

In the previous series, both follicular cancer and the follicular variant of papillary thyroid cancer were reported as the final histological diagnosis in patients with Thy 3 cytology (6, 7, 29–32). Follicular cancer lacks the specific cytological features that can differentiate it from follicular adenoma. Similarly, the follicular variant of papillary thyroid cancer may lack the characteristics, namely nuclear grooves, focal nuclear elongation, chromatin clearing, and small eccentric nucleoli that are typical of the classic variant of papillary thyroid cancer (6, 31, 32). Eight of 371 patients were affected by poorly differentiated thyroid cancer that may lack nuclear cell abnormalities that allow a cytological diagnosis of cancer (33, 34). Only 2 of these 8 patients had persistent disease at the follow-up, due to the timely diagnosis and treatment in most of them.

The determination of somatic mutations in Thy 3 specimens has been proposed as a tool to differentiate between benign and malignant nodules. However, molecular analysis of different oncogenes is not helpful in differentiating between follicular adenomas and follicular cancers, due to the lack of specific gene alterations (35–37). In addition, they are more frequently encountered in the classic than in the follicular variant of papillary thyroid cancer (38). Recently, the study of gene expression has been shown to have a high NPV but is not widely available (39). The recent demonstration of the presence of variability in the cytopathologic and histopathologic evaluation of thyroid nodules confirms the limitation of visual microscopic diagnosis (39). Our group has recently shown that high elasticity at USE has a high NPV in the diagnostic assessment of indeterminate nodules (12, 40). Although USE was confirmed to have a high predictive value in this set of patients, we decided not to include it in the risk score, because it was available in a small subgroup of patients only recently included in this study.

Due to the difficulty in distinguishing patients who harbor malignancy among those with indeterminate nodules at cytology, all the main actual guidelines advise surgery in these patients (15, 16). Aims of the present study were 1) to establish the clinical impact of carrying a nodule with indeterminate cytology and 2) to identify the features associated with malignancy and conferring a risk for the presence of a thyroid cancer disease persistent after thyroidectomy and ablative ¹³¹I treatment. To these purposes, we investigated the clinical outcome of 1520 patients with indeterminate cytology, focusing on the 371 who had a cancer at histology. US features previously described as suggestive of malignancy by us (7, 8) and others (9–11) were confirmed to be poorly predictive when taken singularly.

However, due to the larger number of patients included in the present study, blurred margins and spot microcalcifications at US that had shown a borderline value (7–9–11) in our previous reports became significantly associated with malignancy. At univariate and multivariate analysis, Thy 3B cytology, the coexistence of 3 suspicious US features (namely blurred margins, spot microcalcifications, and hypoechogenicity) were significantly associated with the risk of cancer, whereas age <40 years, although associated with malignancy, did not reach statistical significance. Considering these 3 features, we calculated a risk score of harboring a cancer with a formula that had a high NPV. On clinical grounds, 735 patients had a risk score of −1.58 and then a predicted probability of cancer of 17.2%. More importantly, we questioned whether the above described risk score could identify patients with more unwieldy cancer disease. In general, the outcome of the whole series of patients included in this study was remarkably good. Indeed, 342 of 371 patients were free of disease after total thyroidectomy and ¹³¹I ablative treatment. Only 29 patients needed further treatment, and in 12 of 371 (3.2% of the patients with cancer and thus 0.79% of the total 1520 patients), the disease was persistent at the end of follow-up. Seven of these 12 patients had only a biochemical persistence of the disease and were likely carriers of clinically meaningless disease, and only 5 had a structural evidence of tumor, stressing the low aggressiveness of thyroid cancer in nodules with Thy 3 cytology. Only 4 of the 29 patients who needed treatment after thyroidectomy and ablative radioiodine administration and 2 of the 5 patients with structural persistence of disease at the follow-up were in the group with the lowest risk score of −1.58.

In summary, the overall prognosis of patients carrying a Thy 3 nodule is remarkably favorable. Most patients with thyroid cancer were in remission after thyroidectomy and ¹³¹I ablation. Few patients with thyroid cancer had persistent disease. A formula including cytology and US features predicts the risk of cancer and the risk of more unwieldy cancer disease in patients with indeterminate nodules. Hypothesizing of avoiding surgery in the 735 patients with the lowest risk score, only 4 of the 29 patients who needed treatment after thyroidectomy and ablative radioiodine administration and 2 of the 5 patients with structural persistence of disease at the follow-up would not have had an appropriate treatment. The evaluation of the clinical outcome of patients with Thy 3 nodule and a low risk score who do not undergo surgery would be the definitive proof that they do not need surgery. These data are not available. At the moment, therefore, we suggest taking into account the risk score in patients with particular conditions such as a high surgical risk and old age.

The inclusion of elastography in the diagnostic work-up would increase the predictive value of thyroid US.

In conclusion, we propose a clinical risk score that has a fairly good NPV in the assessment of the risk of cancer and, more importantly, a very high NPV in the assessment of the risk of carrying a more unwieldy cancer disease in patients with indeterminate thyroid nodules. These data should be taken into account in the strategy of treatment of patients with nodules with indeterminate cytology at FNA.

Acknowledgments

Disclosure Summary: The authors have nothing to disclose.

Abbreviations

 
• FNA

  fine-needle aspiration

 
• NPV

  negative predictive value

 
• Tg

  thyroglobulin

 
• TgAb

  Tg autoantibody

 
• US

  ultrasound

 
• USE

  US elastography.