Accuracy of the European Thyroid Imaging Reporting and Data System (EU-TIRADS) in the valuation of thyroid nodule malignancy in reference to the post-surgery histological results.

PURPOSE: To assess the clinical usefulness of the European Thyroid Imaging and Reporting Data System (EU-TIRADS) in the valuation of thyroid nodules malignancy in reference to post-surgery histological results.
MATERIAL AND METHODS: Pre-operative ultrasound was performed in consecutive patients admitted for thyroid surgery between June 2017 and January 2018. Thyroid nodules were classified according to EU-TIRADS to five groups: 1-5. At least one fine-needle aspiration biopsy (FNAB)/patient (dominant or suspected nodule) was performed in an outpatient clinic. The final diagnosis was based on the histological result. The percentage of cancers in each EU-TIRADS group was evaluated. Finally, sensitivity, specificity, accuracy, as well as positive and negative predictive values for malignancy were assessed.
RESULTS: Fifty-two patients with a total of 140 thyroid nodules (median: 3 nodules/thyroid [minimum-maximum: 1-6]) were enrolled in the study. Thyroid cancer was diagnosed in 0% (0/6) in EU-TIRADS 2; 0% (0/92) in EU-TIRADS 3; 5.9% (2/34) in EU-TIRADS 4, and 75% (6/8) in EU-TIRADS 5. In nodules assessed as EU-TIRADS ≥ 4 sensitivity, specificity, positive and negative predictive values for malignancy were, respectively: 75% (CI 95%: 40.7-93.5), 94.1% (CI 95%: 86.0-98.5), 75% (CI 95%: 40.7-93.5), and 94.1% (CI 95%: 86.0-98.5).
CONCLUSIONS: EU-TIRADS is a valuable and simple tool for assessment of the risk of malignancy of thyroid nodules and demonstrates a high ultrasound correlation with histological post-surgery results. FNAB should be performed in all nodules assessed as EU-TIRADS ≥ 4, due to higher risk of malignancy.

Introduction

The widespread use of imaging techniques has generated increasing detection rates of thyroid nodules. The prevalence of thyroid nodules is still high and ranges from 50% to 60% in healthy people [1]. Ultrasound (US) is currently the best detection tool of choice available for the initial work-up of thyroid nodules. The main aim of thyroid US examination is to distinguish benign nodules from those suspicious for malignant features and requiring further investigation. Approximately 90% of thyroid lesions are benign, and 95% are asymptomatic during recognition and further follow-up [2]. Fine-needle aspiration biopsy (FNAB) is used to differentiate between nodules that require surgery and those that can be followed-up. Approximately 10-20% of FNAB are non-diagnostic and need to be repeated [3]. The limitations of FNAB include indeterminate or follicular lesions. The prevalence of unnecessary thyroid surgery is still a therapeutic problem [4]. For all these reasons there was a need to establish some basic criteria to select nodules for FNAB to minimise costs and maximise benefits.

US uses a wide range of features that present different sensitivity and specificity in prediction of thyroid cancer. However, none of them could be used alone in detecting or rejecting malignancy efficiently. The main disadvantage of US examination is relatively low specificity and substantial inter-operator variability [5,6]. To minimise the impact of these limitations, attempts are being made to develop and refine a standardised system.

The American Association of Clinical Endocrinologists and American Thyroid Association have created a practical clinical statement, for better communication between ultrasound practitioners and clinicians, called Thyroid Imaging, Reporting, and Data System (TIRADS) classification. It was first introduced by Horvath et al. [7], originating from the Breast Imaging Reporting and Data System [8]. Since its introduction, several versions of TIRADS have been endorsed by international medical societies to estimate the final assessment and management of thyroid lesions. All of them classify nodules based on suspicious ultrasound features related to composition, echogenicity, shape, margin, and echogenic foci.

In 2017 the European Thyroid Association developed a new and simple European Thyroid Imaging and Reporting Data System (EU-TIRADS), which classifies the risk of malignancy of thyroid nodules in adults to the following categories: benign, low-, intermediate-, and high-risk (e.g. EU-TIRADS 1 is a normal exam, EU-TIRADS 5 means high risk of malignancy) [9]. The aim of this study was to assess the clinical usefulness of EU-TIRADS in the valuation of thyroid nodule malignancy in reference to post-surgery histological results.

Material and methods

Patients

This prospective study was conducted at an academic referral centre between 06.2017 and 01.2018. Patients were admitted to the Department of General and Endocrine Surgery for thyroid or parathyroid surgery. The decision about surgical treatment was made before admission to the hospital, based on clinical, US, laboratory, and FNAB evaluation. Patients were consecutively recruited to the study if they fulfilled the following criteria: 1) preoperative US reassessment of the neck performed by two certified ultrasonographers; 2) preoperative FNAB of dominant or suspicious lesion; 3) surgical thyroid resection with histological evaluation.

All the procedures were performed in accordance with the Helsinki Declaration of 1975, as revised in 2000, and with the ethical standards of the local Research Ethics Committee.

Ultrasound examination

All patients were reassessed with a preoperative US of the neck for nodal staging by two certified ultrasonographers (AS and MŻ) (certificate of Polish Ultrasound Society) with two and 15 years of experience, respectively, in thyroid US. US was performed in the Department of Endocrinology a few days before the thyroid surgery. All the US scans were performed using a Hitachi Avius Medical ultrasound technique system equipped with a 7.5-12 MHz high-frequency linear array transducer. Thyroid US procedures were performed on the basis of the Ultrasound Examination Standards of the Polish Ultrasound Society (2014) [10]. The neck was scanned in sagittal, transverse, and oblique sections to optimally visualise both lobes of thyroid, isthmus, and carotid arteries, as well as internal jugular veins. All images were examined on real-time two-dimensional B-mode grey-scale and Doppler imaging. All obtained sonograms were saved in a picture archiving and communication system. The US examination study did not change the management of surgical interventions. The interobserver agreement was measured using a weighted kappa statistic with 95% confidence intervals (CI 95%). Levels of agreement included values: ≤ 0 as no agreement, 0.01-0.20 as slight, 0.21-0.40 as fair, 0.41-0.60 as moderate, 0.61-0.80 as substantial, and 0.81-1.00 as almost perfect agreement. In this study the interobserver agreement was almost perfect.

EU-TIRADS

According to the EU-TIRADS, we assessed the presence of five equally-weighted nodule features (irregular/non-oval shape, irregular margins, microcalcifications, marked hypoechogenicity, solid component), and assigned the lesion to one of five categories based on the number of suspicious features. To define and specify each US pattern, the following variables were considered: composition (solid, predominantly solid, predominantly cystic, simple cyst, spongiform), echogenicity (hyperechoic, isoechoic, hypoechoic, markedly hypoechoic), margins (circumscribed, non-circumscribed), and calcifications (absent, microscopic, macroscopic – the latter including eggshell calcifications). Other parameters were also recorded: 1) nodule diameters (transverse, anteroposterior, and longitudinal); and 2) nodule shape (oval [wider than tall], non-oval [taller than wide], irregular).

Cervical lymph nodes were evaluated for their size, loss of the central, echogenic hilum, the presence of irregular and indistinct margin, microcalcifications, and necrotic changes.

Normal thyroid gland was classified as EU-TIRADS 1, a benign nodule as EU-TIRADS 2 (Figure 1A,B), a nodule with low suspicion for malignancy as EU-TIRADS 3 (Figure 1C,D), a nodule with intermediate suspicion for malignancy as EU-TIRADS 4 (Figure 1E), and a nodule highly suggestive of malignancy as EU-TIRADS 5 (Figu-re 1F,G).

Figure 1

Representative images of EU-TIRADS categories. A) EU-TIRADS 2: a well-defined oval cystic nodule; B) EU-TIRADS 2: a multiple spongiform cystic nodule; C) EU-TIRADS 3: a large well-defined normoechoic nodule; D) EU-TIRADS 3: a hyperechoic nodule. E) EU-TIRADS 4: a well-defined mild hypoechoic nodule; F, G) EU-TIRADS 5: an ill-defined oval hypoechoic nodule surrounded by incomplete halo sign with a microcalcification inside (arrow)

Fine-needle aspiration biopsy

FNAB of dominant or suspicious lesions in US (at least in one thyroid lesion) was performed before the admission to the hospital. The majority of FNABs were completed in the outpatient clinic before admission to the hospital. The Bethesda System for reporting thyroid cytopathology was used in the classification of cytology reports.

Surgery

Thyroid surgeries were performed in the Department of General and Endocrine Surgery. Total thyroidectomy was performed in 40 patients (eight with suspicion of carcinoma in FNAB, 32 with multinodular goitre), lobectomy in nine (17.3%) patients, and lobectomy plus isthmectomy in three (5.8%) patients. Total thyroidectomy with neck exploration was done in seven (13.5%) patients, of these lymph nodule dissection was performed in three patients.

Histological evaluation

Histological evaluation was performed by one pathologist (MM). All of the nodules analysed in US were separately assessed in histological examination.

Statistical analysis

Statistical analysis was performed using the statistical program STATISTICA 10 (StatSoft Inc.). Qualitative characteristics were described along with their percentage distributions. For quantitative variables, averages and standard deviation (SD) or median (range) were used. The consensus decisions were compared with surgical histology reports (reference standard). Specificity, sensitivity, positive predictive value (PPV), and negative predictive value (NPV) (with 95% confidence interval [CI 95%]) were calculated to evaluate the reliability of EU-TIRADS classification in differentiation between benign and malignant features.

Results

Fifty-two patients (44 female) with mean age of 55 ± 14 years were enrolled in the study. Patient characteristics including age, sex, indications for surgery, clinical symptoms, thyrotropin, and free thyroid hormone levels are presented in Table 1.

Table 1

Demographic and clinical profile of patients

Factor	All patients (n = 52)	Patients with benign nodules (n = 44)	Patients with malignant nodules (n = 8)
Age (years)
Mean	55 ± 14	56 ± 12	35 ± 16
< 30	3 (5.8%)	1 (2.3%)	2 (25%)
30-39	9 (17.3%)	7 (15.9%)	2 (25%)
40-49	9 (17.3%)	9 (20.5%)	–
50-59	10 (19.2%)	9 (20.5%)	1 (12.5%)
60-69	15 (28.8%)	14 (31.8%)	1 (12.5%)
≥ 70	6 (11.5%)	4 (9.1%)	2 (25%)
Sex
Female	44 (84.6%)	39 (88.6%)	5 (62.5%)
Male	8 (15.4%)	5 (11.4%)	3 (37.5%)
Clinical symptoms
Goitre compression	25 (48.1%)	24 (54.5%)	1 (12.5%)
Laryngeal nerve palsy	2 (3.8%)	–	2 (25%)
Rapidly enlarging mass	4 (7.7%)	2 (4.5%)	2 (25%)
Hyperthyroidism	5 (9.6%)	4 (9.1%)	1 (12.5%)
Cervical lymphadenopathy	2 (3.8%)	–	2 (25%)
Thyroid function
TSH (0.27-4.2 mIU/l)	1.33 ± 0.84	1.31 ± 0.85	1.44 ± 0.78
FT4 (12-22 pmol/l)	16.30 ± 2.45	16.15 ± 2.43	17.70 ± 2.21
FT3 (3.1-6.8 pmol/l)	4.84 ± 0.76	4.84 ± 0.77	5.25 ± 0.68
Indications for surgery
Suspicion of malignancy	12 (23.1%)	6 (13.6%)	6 (75%)
Goitre size and signs of compression	20 (38.5%)	19 (43.2%)	1 (12.5%)
Toxic multinodular goitre	5 (9.6%)	4 (9.1%)	1 (12.5%)*
Cosmetic problems	2 (3.8%)	2 (4.5%)	–
Other
Multinodular goitre and primary hyperparathyroidsm	13 (25%)	13 (29.5%)	–

FT3 – free triiodothyronine, FT4 – free thyroxine, TSH – thyrotropin

Incidental, histologically detected follicular microcarcinoma

Ultrasound evaluation

The total amount of thyroid nodules found in US was 140 (median: 3 nodules/patient [minimum-maximum: 1-6]). The details of US examination are shown in Table 2.

Table 2

Ultrasound features of benign and malignant nodules

Ultrasound features	All nodules (n = 140, 100%)	Benign nodules (n = 132, 94.3%)	Malignant nodules (n = 8, 5.7%)
Composition
Solid	41	34 (25.8%)	7 (87.5%)
Predominantly solid	62	61 (46.2%)	1 (12.5%)
Predominantly cystic	30	30 (22.7%)	–
Simple cyst	3	3 (2.3%)	–
Spongiform	4	4 (3.1%)	–
Echogenicity
Hypo-/markedly hypo-	73	65 (49.2%)	8 (100%)
Hyper-/iso-	67	67 (50.8%)	–
Margins
Circumscribed	134	132 (100%)	2 (25%)
Non-circumscribed	6	–	6 (75%)
Calcifications
Microcalcifications	5	2 (1.5%)	3 (37.5%)
Macrocalcifications	17	16 (12.1%)	1 (12.5%)
None	118	114 (86.4%)	4 (50%)
Shape
Oval (wider than tall)	134	131 (99.2%)	3 (37.5%)
Non-oval (taller than wide	6	1 (0.8%)	5 (62.5%)
Irregular	8	–	8 (100%)
Lymphatic nodes
Suspected for malignancy	2	–	2 (25%)

n – number of nodules, US – ultrasound

The overall distribution in EU-TIRADS categories was as follows: six EU-TIRADS 2 (4.3%), 92 EU-TIRADS 3 (65.7%), 34 EU-TIRADS 4 (24.3%), and eight EU-TIRADS 5 (5.7%) (Table 3).

Table 3

The results of EU-TIRADS classification

EU-TIRADS n (%)	Nodule size (mm), median (range)	Preoperative FNAB n (%)	Histopathological findings n (%)
EU-TIRADS 26 (4.3%)	19.5 (6.0-41.0)	II – 3 (4.8%)	Colloid nodule – 3 (50%)Simple cyst – 3 (50%)
EU-TIRADS 392 (65.7%)	18.0 (5.0-47.5)	II – 28 (44.4%)III – 7 (11.1%)IV – 1 (1.6%)	Colloid nodule – 66 (71.7%)Hyperplastic nodule – 25 (27.2%)Adenoma – 2 (2.2%)
EU-TIRADS 434 (24.3%)	11.6 (5.0-29.0)	II – 9 (14.3%)III – 4 (6.3%)IV – 3 (4.8%)V – 1 (1.6%)VI – 1 (1.6%)	Colloid nodule – 20 (58.8%)Hyperplastic nodule – 12 (35.3%)Adenoma – 2 (5.9%)Papillary cancer – 1 (2.9%)Anaplastic cancer – 1 (2.9%)
EU-TIRADS 58 (24.3%)	17.8 (14.2-34)	IV – 2 (3.2%)V – 3 (4.8%)VI – 1 (1.6%)	Hyperplastic nodule – 2 (25%)*Lymphoma – 1 (12.5%)Papillary cancer – 2 (25%)Follicular cancer – 2 (25%)Medullary cancer – 1 (12.5%)

FNAB not performed

EU-TIRADS – European Thyroid Imaging Reporting and Data System; FNAB – fine-needle aspiration biopsy; n – number of nodules

FNAB of 63 nodules was performed (details in Table 3). Among the examined nodules, 37 (58.7%) were assessed as benign, four (6.3%) were suspected for malignancy, and two (3.2%) were reported as malignant. The rest of them (20 [31.7%]) were categorised as III or IV according to the Bethesda system. In the final histological examination, thyroid cancer was not recognised in nodules with category III in FNAB evaluation, in two among six nodules with category IV (anaplastic and follicular cancer), and in all nodules with category V (medullary and papillary cancer) and with category VI (lymphoma, anaplastic cancer).

Histological examination revealed 132 (94.3%) benign nodules (adenomas, hyperplastic, cystic, and colloid nodules) and eight (5.7%) cancers (details in Table 4).

Table 4

Histological evaluation (n = 140)

Histological evaluation	Nodules n (%)	Ultrasonography
Colloid nodule	89 (63.6%)	Solid – 19 (21.3%)Predominantly solid – 38 (42.7%)Predominantly cystic – 21 (23.6%)Simple cyst – 3 (3.4%)Spongiform – 8 (8.9%)Hypo-/markedly hypoechoic – 14 (15.7%)Hyper-/isoechoic – 75 (84.3%)
Hyperplastic nodule	39 (27.9%)	Solid – 12 (31.6%)Predominantly solid – 19 (48.7%)Predominantly cystic – 8 (21.1%)Hypo-/markedly hypoechoic – 6 (15.8%)Hyper-/isoechoic – 32 (8.2%)
Adenoma	4 (2.9%)	Solid – 4 (100%)Hypo-/markedly hypoechoic – 4 (100%)
Follicular cancer	2 (1.4%)	Solid – 1 (50%)Predominantly solid – 1 (50%)Hypo-/markedly hypoechoic – 2 (100%)Non-circumscribed – 1 (50%)
Papillary cancer	3 (2.1%)	Solid – 3 (100%)Hypo-/markedly hypoechoic – 3 (100%)Non-circumscribed – 3 (100%)
Medullary cancer	1 (0.7%)	Solid – 1 (100%)Hypo-/markedly hypoechoic – 1 (100%)
Anaplastic cancer	1 (0.7%)	Solid – 1 (100%)Hypo-/markedly hypoechoic – 1 (100%)Non-circumscribed – 1 (100%)
Lymphoma	1 (0.7%)	Solid – 1 (100%)Hypo-/markedly hypoechoic – 1 (100%)Non-circumscribed – 1 (100%)

n – number of nodules

Comparison of ultrasound and histological evaluation

The US evaluation of selected patterns of thyroid nodules with comparison between histologically recognised malignant and benign nodules is summarised in Table 2. Solid composition, hypoechogenicity, microcalcifications, “taller than wide”, irregular shape, and non-circumscribed margins were more frequent ultrasound features in malignant lesions than in benign nodules. Similarly, predominantly cystic component, simple cyst, spongiform, hyper-/isoechogenic, oval, and with circumscribed margins nodules were associated more often with the histological diagnosis of benign lesions. There was no significant difference between the size of benign (16 ± 17.5 mm) and malignant nodules (17.5 ± 6.8 mm).

Comparison of EU-TIRADS and histological evaluation

The percentage of malignancy for each EU-TIRADS category was as follows: 0% (0/6) in EU-TIRADS 2; 0% (0/92) in EU-TIRADS 3; 5.9% (2/34) in EU-TIRADS 4; and 75% (6/8) in EU-TIRADS 5. In nodules assessed as EU-TIRADS ≥ 4, predictive values for malignancy were assessed: sensitivity of 75% (CI 95%: 40.7-93.5); specificity of 94.1% (CI 95%: 86.0-98.5); PPV of 75% (CI 95%: 40.7-93.5); and NPV of 94.1% (CI 95%: 86.0-98.5).

Discussion

For the last nine years several attempts have been made to create a standardised systematic categorisation of thyroid nodules according to their risk of malignancy, in order to optimise focal lesion management through US follow-up, cytological analysis, and surgical treatment with pathological verification.

In 2009, Horvath et al. and then Park et al. [7,11], first presented the Thyroid Imaging Reporting and Data System (TIRADS). The system, based on 10 US patterns proposed in a prospective study by Horvath et al. did not cover the entire spectrum of lesions that may occur in thyroid nodules, especially multinodular goitre. The classification described in a retrospective study by Park et al., based on a formula for calculation of TIRADS, seemed to be too complicated for common use by US practitioners. In 2011 Kwak et al. [12] published another approach for TIRADS classification, a more accurate tool in daily practice for the evaluation of thyroid cancer in multinodular goitre, particularly to select cases that require biopsy [13]. Russ et al. [14] presented in 2013, in a large prospective study based on the results of 4550 thyroid nodules, a simple classification that included only five suspicious sonographic patterns, together with optional elastography. In 2017, the American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS) was published [15]. Unlike other guidelines, the ACR TI-RADS does not recommend FNAB of nodules with benign US characteristics regardless of size. The novel EU-TIRADS classification created recently by the European Thyroid Association assesses the presence of five equally-weighted lesion patterns and assigns the nodule to one of five categories, based on the number of suspicious features. This system is both practical and accurate.

To our knowledge, our study is the first prospective trial that assesses EU-TIRADS classification in the correlation with histological findings received from surgical specimens.

Among nodules evaluated as EU-TIRADS 2 or 3 we did not find any cancer. In nodules assessed as EU-TIRADS 4 and 5 we diagnosed thyroid carcinoma in two (6%) and six (75%) nodules, respectively. These results are in accordance with EU-TIRADS statement: FNAB is not necessary in EU-TIRADS 2, but should be performed in EU-TIRADS 4 and 5. In EU-TIRADS 3 FNAB should be performed only for nodules > 20 mm [9].

With a cut-off point at EU-TIRADS ≥ 4, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for recognition of nodules malignancy were: 75%, 94.1%, 75%, and 94.1%, respectively. In the prospective validation of TIRADS classification performed by Horvath et al. [16] in 502 resected thyroid nodules with a cut-off point at TIRADS ≥ 4 to perform FNAB, the specificity was lower (74.4%), but the sensitivity, PPV, and NPV were higher (99.6%, 82.1%, and 99.4%, respectively). According to the results described by Russ et al. [14] for TIRADS grey-scale score in combination with cytological results sensitivity, specificity, and negative predictive value were as follows: 95.7%, 61%, and 99.7%, respectively.

Additionally, we found that solid composition, hypoechogenicity, microcalcification, “taller than wide”, irregular shape, and non-circumscribed margins were more frequent ultrasound features in malignant lesions. Nodules with predominantly cystic component, simple cyst, spongiform, hyper-/isoechogenic, oval, and with circumscribed margins were associated with the histological diagnosis of benign lesions. This is in accordance with the results of previous studies [17-20].

The limitations of this study include: (i) relative small sample size; (ii) small number of thyroid cancers; and (iii) the lack of US examination with the usage of elastography. It would be interesting to compare the results with those presented by Russ et al. [14], Migda et al. [21], and Dobruch-Sobczak et al. [22]. According to Gietka et al. [23], Wolinski et al. [24], and Zhao et al. [25], real- time elastography, shear wave elastography, and three-dimensional shear wave elastography are useful in selecting nodules to FNAB. However, other authors reported clearly that grey-scale with elastography has poorer results for discriminating benign from malignant thyroid lesions in comparison to single grey-scale assessment [26-28].

Conclusions

To conclude, EU-TIRADS is a valuable and simple tool in assessing the malignancy of thyroid nodules, which demonstrates a high clinical correlation with histological results. Because of its high sensitivity it can be used in selecting nodules with a high risk of cancer. EU-TIRADS improves qualification for surgical treatment and leads to the avoidance of unnecessary procedures such as FNAB of benign thyroid nodules.

Conflict of interest

The authors report no conflict of interest.