2023 European Thyroid Association Clinical Practice Guidelines for thyroid nodule management

1A thyroid nodule is a discrete lesion within the thyroid gland that is ultrasonographically distinct from the surrounding thyroid parenchyma. With the widespread use of sensitive imaging techniques, which include neck visualization, a conspicuous number of thyroid nodules emerge and demand attention. Up to 60% of adults in the general population harbor one or more thyroid nodules. The likelihood of malignancy is an overriding concern, but the actual prevalence of cancer in unselected thyroid nodule populations generally ranges from 1 to 5%, with variation related to selection criteria and the population under evaluation, for example, whether papillary micro-carcinomas are included or not. Thus, most lesions are benign, asymptomatic, and do not warrant treatment. In the case of cancer diagnosis, most are small, intrathyroidal and indolent neoplasms (up to 53.6%, as shown in one contemporary large-scale study, in an unselected population) that can safely be managed conservatively. Accepting that not all reasons for surgery are disclosed in the available literature, in European countries a proportion of such lesions are superfluously referred for surgery, leading to an unfavorable risk and cost–benefit ratio. A more conservative approach results in reduced cost as well as a lower risk of complications. Implementation of active surveillance and minimally invasive techniques (MITs) has been limited and measured.

2Overall, these observations cause concern for unwarranted expense and excess morbidity associated with thyroid nodule over-diagnosis and -treatment. This is exemplified by the abundantly demonstrated overdiagnosis of benign and malignant thyroid lesions without this leading to a significant lowering of thyroid cancer mortality. There is a pronounced need for more cost-effective, risk-adapted approaches to the management of this highly prevalent condition, taking the wishes of the patient into consideration. Based on the aforementioned, ultrasound (US) screening of asymptomatic adults is discouraged. Thus, the present guidelines aim at providing a clinical practice guide for the initial workup and the subsequent management of adult individuals harboring thyroid nodules. In the present manuscript, nodular thyroid disease includes both the solitary nodule, whether functioning or non-functioning, and multinodular goiter, independent of the patient having clinical goiter or not. Importantly, these guidelines are not intended to cover the management of thyroid malignancy. Notably, all recommendations should take into consideration the clinical setting, medical expertise, available technology, and patient preference. As a consequence, nationally adapted guidelines may well, in certain areas, deviate from this European Thyroid Association (ETA) guideline. Over the past years, the ETA has produced guidelines on the topics of US risk stratifications of thyroid nodules and lymph nodes, molecular cytology diagnostics, as well as MIT for benign and malignant nodules. The present document incorporates and updates aspects of these guidelines, where appropriate. For more in-depth information the reader is referred to the guidelines in question.

3Recommendations for thyroid nodule management are summarized in Table 1 and Figs. 1 and 2.

4The Executive Committee of the ETA, upon consultation with its Guideline Board, commissioned the development of this guideline to a multidisciplinary team led by two chairpersons (C D, L H). The task force consisted of four endocrinologists (C D, L H, R P, E P), one internist/clinical cytologist (T S), one endocrine surgeon (A C), a radiologist (G R), one pathologist (F S), and a biologist (P S). All panelists had to be members of the ETA and be experts in nodular thyroid disease. In hindsight, the document could have benefitted from including even more specialists, for example, a nuclear medicine physician. However, we have attained this by incorporating the comments from nuclear medicine experts within the ETA.

5The literature considered was retrieved based on a systematic search on the MEDLINE database, through the PubMed search engine. The manuscript and the specific recommendations were developed by reconciling the best available research evidence with the knowledge and clinical experience of the panelists. The grading of recommendations, assessment, development, and evaluation (GRADE) framework was used for grading the quality of evidence and making clinical practice recommendations. The quality of a body of evidence was rated as high (ØØØØ), moderate (ØØØO), low (ØØOO), or very low (ØOOO), while the strength of a recommendation was categorized as either strong (recommended for all or almost all patients; indicated by 1) or weak (different choices may be appropriate for some patients and settings; indicated by 2). The statements that panelists considered worthy of being presented, but inappropriate for rating the evidence despite a considerable and compelling amount of indirect evidence, are described as ‘ungraded good practice statement’. Consensus on all the recommendations was reached following a modified Delphi process, which involved two rounds of voting using an online survey tool (Google Forms platform). Panelists rated each recommendation on a 5-point scale: strongly agree, agree, neutral, disagree, or strongly disagree. Consensus was defined as >80% of voters (that is a minimum of eight out of nine voters) strongly agreeing, agreeing or being neutral with a statement. The total number and percent of individuals who agreed with each recommendation and the round in which agreement was achieved are reported in Table 1. The final draft of the manuscript was sent to the Guideline Board for comments and thereafter posted on the ETA website for four weeks for critical evaluation by the ETA members. All proposed changes and comments were taken into consideration within the guideline task force, and the resulting changes were incorporated into the final document submitted to the European Thyroid Journal, after approval by the Guideline Board.

6The initial evaluation of any patient suspected of nodular thyroid disease includes the combination of personal and family history, physical examination, evaluation of thyroid function, and US of the neck. This should take into consideration pre-treatment symptomatology, ideally including a thyroid specific patient-reported outcome (PRO), which has been cross-culturally validated and for which instrument minimally important change in the quality of life has been determined, allowing the use for treatment effect evaluation.

7As a minimum, laboratory assessment needs to include thyroid-stimulating hormone (TSH) measurement. If TSH is decreased, we recommend determining free thyroxine (FT4). If the latter is normal, free tri-iodothyronine should be measured. Based on the clinical context, TSH receptor antibody determination may be considered to define the etiology of hyperthyroidism. If TSH is elevated, FT4 and anti-thyroid peroxidase (TPO) antibodies should be measured to aid in the classification of the etiology of thyroid dysfunction. In case of clinical or US suspicion of chronic lymphocytic thyroiditis and negative anti-TPO antibodies, measurement of anti-thyroglobulin (Tg) antibodies may be considered (Fig. 1).

8Using calcitonin for medullary thyroid carcinoma (MTC) screening in unselected thyroid nodule populations provides an early diagnosis and thereby potentially improves prognosis. However, the value of such screening is still under debate, based on the fact that at most 1 in 200 screened individuals will have MTC and that some might also have been diagnosed without routine screening. In the following scenarios, calcitonin evaluation is appropriate: patients with thyroid nodules scheduled for surgery or MIT; thyroid nodules with indeterminate cytology or suspicious US findings. Calcitonin determination should be performed in case of personal or family history of medullary thyroid cancer or multiple endocrine neoplasia type 2. Cut-off points to separate non-medullary reasons for increased calcitonin from medullary thyroid carcinoma have been established (>30 pg/mL in females and >34 pg/mL in males, even if several variables may affect this threshold).

9Serum Tg and Tg antibody determination has no role in the initial evaluation of nodular thyroid disease.

10US is more sensitive and specific than palpation for the evaluation of nodular thyroid disease.

11Diagnostic thyroid and neck US should be performed in all patients clinically suspected of having nodular thyroid disease, or if a nodule is incidentally detected using another imaging modality (e.g. US scan of carotid arteries, computed tomography (CT) of the neck).

12The thyroid bed and anterior neck from hyoid bone to sternal notch and below, if accessible.

13From levels II to V in the lateral neck and level VI in the central neck for lymph nodes (Fig. 3).

14A high-frequency linear probe (e.g. up to 14 MHz) is adequate for exploring these regions. To visualize the inferior pole of an intrathoracic thyroid a convex probe (frequency range: 2.5–5.0 MHz) may be useful.

15Assess the presence, location, size, and features of nodule(s) and thyroid lobes. Details of what is expected in the US report are provided in Table 2. For a description of thyroid nodule features, the use of an appropriate lexicon is warranted (Appendix, see section on supplementary materials given at the end of this article).

16Visualize any intrathoracic extension and its relationship with the cervical structures.

17Stratify the risk with the use of the European Thyroid Imaging and Reporting Data System (EU-TIRADS) (Table 3).

18Investigate for the presence of suspicious lymph nodes.

19Select nodules to be addressed to fine-needle aspiration (FNA) biopsy.

20Monitor growth (an increase ≥20% in at least two nodule diameters with a minimum increase of 2 mm, or nodule volume increase >50% at the time of re-evaluation) of thyroid nodule(s) (Fig. 2).

21Monitor US feature changes that may modify risk stratification.

22Monitor lateral neck lymph nodes.

23Re-evaluate in case of the appearance of local pressure symptoms and/or voice changes.

24The usefulness of the Doppler vascular pattern for defining the risk of malignancy (ROM) of thyroid nodules is controversial. Even if more sensitive techniques are now available (e.g. super-resolution, microvascular imaging), their role remains unclarified. However, besides providing complementary information, Doppler imaging may be indicated in order to differentiate between cases where vascularization is diminished or absent (e.g. thick colloid, cystic or necrotic nodules) from solid nodules. So, it may indirectly be useful for risk stratification and for guiding FNA and minimally invasive procedures in mixed thyroid lesions.

25The role of elastography remains unsettled. The studies concerning the usefulness of these techniques have not resolved a consensus on their use or reporting. Importantly, while the classical variant of papillary thyroid carcinoma (PTC) has demonstrated high stiffness, other variants of PTC and follicular thyroid carcinoma (FTC) may show a normal stiffness. Thus, the contribution of this method to the standard US imaging does not justify its routine use and inclusion in the risk-stratification systems (RSSs).

26A few meta-analyses report that contrast-enhanced ultrasound (CEUS) has a rather high positive and negative predictive value for the assessment of the ROM. However, authors have reported inconsistency across studies and publication bias. Thus, even if these methods may add some information to gray-scale US examination, their use is limited because contrast agents are expensive, invasive, and not universally licensed for this purpose. However, CEUS provides a clear depiction of the ablated areas after thermal ablation (TA) of thyroid nodules and offers an advantage for guiding the need for repeat treatments.

27TIRADS scores have been proposed for improving inter-observer reproducibility in the description of US features and quantifying malignancy risk of thyroid nodules. Additional goals are to unify the reports, to facilitate their understanding and to standardize the management recommendations. Currently, there are several stratification systems, either ‘pattern-based’ – when all the features of the nodule are taken into account to analyse its risk – or ‘point-based,’ when points are assigned to each US sign. No system has consistently demonstrated superiority over another, and the inter-observer reproducibility of the scores is similar. Each system provides recommendations for FNA, based on the score and size of the nodules. As all TIRADS reduce the number of unnecessary FNAs, all relevant specialty societies endorse using TIRADS. In Europe, the preferred system is the EU-TIRADS (Table 2).

28Even if TIRADS scores provide quantitative malignancy risk stratification, users should be aware of the following:

29TIRADS have been designed and mainly tested for PTCs, although they are proposed to estimate the malignancy risk of any thyroid neoplasm. The sensitivity for the detection of the classical variant is excellent but decreases substantially for the follicular variant and even more for FTCs. Accuracy in identifying medullary thyroid carcinomas is debated.

30Misdiagnoses may occur, especially in cystic nodules and in sub-acute as well as chronic thyroiditis.

31Composition of the nodule is not included in the cardinal features of the EU-TIRADS. However, the users should consider that the ROM is higher in completely solid than mainly cystic ones.

32Ultrasound features suggestive of extra-thyroidal extension (i.e. capsular bulging, disruption, or abutment by the thyroid nodule) are not included in the cardinal features of the EU-TIRADS. However, they should be described in the report as they are associated with a higher ROM and should prompt FNA irrespectively of EU-TIRADS score.

33When using EU-TIRADS and in case of difficulties with ascertaining the presence of features of high suspicion, we suggest classifying these nodules as EU-TIRADS 4.

34TIRADS scores do not include lymph node evaluation. Cervical lymph nodes should be described according to the 2013 ETA guidelines terminology as normal, indeterminate, or suspicious, and located using the six cervical levels nomenclature (Fig. 3). Features of high suspicion are the presence of cystic areas, microcalcifications, thyroid tissue-like appearance, and anarchic vascularity in the absence of a visible hilum.

35Having obtained clinical, biochemical and US evaluation, and in a dialogue with the patient, a decision regarding the indication for FNA can be made. How to perform the FNA is beyond the scope of this guideline and we refer to a recent video. As a rule, FNA should be performed under US guidance. Each US RSS has its own cut-offs for guiding indications for FNA cytology, and there is a continuous debate on the optimum threshold. These recommendations have been shown to reduce the number of superfluous FNAs. For instance, compared prospectively in 477 patients, the performances of five internationally endorsed sonographic classification systems (those of the American Thyroid Association (ATA), American Association of Clinical Endocrinologists (AACE)-American College of Endocrinology (ACE)-Associazione Medici Endocrinologi (AME), American College of Radiology (ACR), ETA, and Korean Society of Thyroid Radiology), application of the FNA criteria would have reduced the number of biopsies performed by 17.1–53.4% (17.1% using K-TIRADS, 30.7% using EU-TIRADS, 34.9% applying AACE/ACE/AME, 43.8% for ATA, and 53.4% with ACR TIRADS). The percentage of missed carcinomas in nodules >1 cm was low comprising between 2.2% for ACR TIRADS and 4.1% for ATA TIRADS. Thus, all RSSs seem effective to reduce the number of unnecessary FNAs, at the expense of temporarily postponing the diagnosis of a minute proportion of carcinomas. The diagnosis of such carcinomas will be postponed until they eventually grow and are diagnosed after reaching the cut-off threshold defined for FNA, according to their US risk category. There is no evidence of this strategy implying a significant loss of quality of life or an increase in morbidity and mortality.

36The indications for FNA, based on EU-TIRADS, and the factors that may influence this choice are described in Tables 3 and 4, respectively. Severe coagulation disorders represent a contraindication to FNA, while the use of anticoagulant therapy does not, as long as INR is below 3. Antiaggregant therapy is not an absolute contraindication to FNA.

37Unless highly suspicious for malignancy, hyperfunctioning thyroid nodules should not be biopsied. Scintigraphy should be performed in case of subnormal serum TSH, with or without elevated free thyroid hormones. In certain situations, scintigraphy may be warranted also when TSH is normal (e.g. in current or formerly iodine deficient regions and in case of a multinodular goiter). The reasons mainly being to decide eligibility for FNA and/or radioactive iodine (RAI) treatment.

38FNA should be repeated in case of:

39a first non-diagnostic sample;

40a Bethesda class III cytology;

41discrepancy between US risk score (i.e. high risk) and cytological findings (i.e. benign cytology).

42Core-needle biopsy (CNB), performed with a large bore spring-activated device, may provide a micro-histological sample from thyroid lesions. There are no clear advantages of using CNB, a more invasive and expensive procedure compared to FNA, based on cost and risk–benefit analysis.

43CNB may be considered in the following situations:

44repeat inadequate FNA as an alternative to diagnostic surgery;

45repeat Bethesda class III cytology;

46when histological assessment can improve pre-operative diagnosis (e.g. suspicion of poorly differentiated or undifferentiated thyroid cancer, thyroid lymphoma, thyroid metastases).

47The only available pathology reporting system has been proposed by the Korean thyroid CNB study group.

48In patients suspected of lymph node metastases, diagnostic confirmation should be obtained by US-guided FNA, before offering therapy. In the case of differentiated thyroid cancer, Tg washout determination should be added, and in the case of medullary thyroid cancer, calcitonin measurement. FNA-Tg washout levels in metastatic thyroid lymph nodes are usually much higher than in the circulation and allow a definitive diagnosis. In some cases of small nodules with normal serum calcitonin but cytological suspicion of medullary thyroid cancer, or in case of elevated serum calcitonin, calcitonin washout assessment of the nodule can be useful. In the rare cases of intranodular parathyroid adenomas, parathyroid hormone determination in FNA washout may confirm the clinical suspicion.

49Thyroid cytopathology should be reported using a widely accepted and endorsed classification system. Currently, the most widely used system is ‘The 2017 Bethesda System for Reporting Thyroid Cytopathology’ (TBSRTC), enjoying wide acceptance internationally. Advantages include well-defined ROM rates, management algorithms linked to each diagnostic category, and integration of molecular tests in the reporting. TBSRTC should be integrated into the local setting, with local calculation of ROM based on demography, including local thyroid cancer incidence. There are clearly formulated requirements for cytopathology quality control. The details are beyond the scope of this document. Table 5 shows the distribution of benign and malignant entities across the different Bethesda categories.

50Identification of the patient

51Imaging findings and, if available, TIRADS score

52Adequacy of the sample

53Microscopic description of the material including cellular and colloid components

54Ancillary testing (if performed)

55Reporting category and subclassification (specific diagnosis)

56The local ROM of the diagnostic category

57The morphology remains the cornerstone to distinguish benign from malignant nodules. HBME-1, Galectin-3, and CK19 immunostaining are more frequently positive in thyroid carcinomas, but their expressions are variable, and no single stain has sufficient sensitivity and specificity to be recommended for routine practice. There are recent studies applying antibodies to detect genetic alterations such as BRAF V600E (VE1), Pan-Trk, and ALK among others. However, they are not recommended for routine use. In some special situations, stains for calcitonin, Tg, and TTF1 can be useful, for example, in FNA of suspicious metastatic cervical lymph nodes.

58Using an integrated approach combining careful clinical, US, and cytology malignancy risk assessment with the local outcome and test performance data, molecular testing may improve diagnostic outcomes for thyroid nodules by identifying patients with indeterminate cytology as most likely benign. Based on local integrated diagnostic pathway outcome data, this strategy may obviate diagnostic surgery or identify patients with a high likelihood of malignancy, allowing surgical treatment.

59Currently, available molecular FNA tests are based on examining for somatic mutations, evaluation of gene expression, and microRNA-based classifiers. For details of the ThyroSeq, the Afirma Genomic Sequencing Classifier (GSC) and the ThyGeNEXT/ThyraMIR, see Table 6. Up to 13.4% avoided diagnostic surgeries have been reported for the ThyroSeq and the Afirma GSC. ThyroidPrint is a gene expression classifier based on the interrogation of only ten mRNAs. ThyroSPEC™ is a MALDI-TOF mass spectrometry-based mutation detection panel that detects the most prevalent 117 point mutations and 23 gene fusions in thyroid cancer.

60Although a fully comprehensive genomic profile of thyroid FNAs is offered by centralized laboratories in North America, recent efforts in other countries have concentrated on developing local laboratory-developed thyroid molecular tests in research settings of a few centers in Europe and East Asia and a publicly funded test in Canada with promising results.

61Notably, benign lesions frequently exhibit RAS mutations and have been reported positive for PAX8/PPARG rearrangements. The pronounced variation in reports of the prevalence of these and other biomarkers in benign lesions suggests the need for data with sufficient sample size and histologic evaluation according to WHO guidelines, and caution when including these markers especially in binary instead of complementary diagnostic decisions to avoid overtreatment of patients with benign thyroid tumors. Considerable inter-observer variation for differentiating benign follicular cell-derived tumors and minimally invasive FTCs hamper translational studies assessing the value of molecular markers that aim to improve presurgical diagnosis because the histological reference for these studies can be ambiguous or discrepant.

62Currently, molecular FNA testing of indeterminate FNA outside of the USA is limited to mainly research use of local laboratory-developed tests and a publically funded test in Alberta/Canada. Molecular tests marketed in the USA are currently not used as reimbursed tests outside of the USA. Possible reasons for this are health care system differences, issues pertaining to lack of independent validation studies, lack of long-term outcome studies for ‘benign’ molecular tests, and the high cost that currently limits their use outside the USA.

63Whether the comprehensive molecular profile of thyroid nodules can provide prognostic information and guide the extent of surgery is still debated. However, high-risk molecular profiles (e.g. coexistence of either BRAF p.V600E or RAS mutations with late-hit mutations like those in TERT promoter, PIK3CA or TP53 genes) have been strongly associated with the presence of distant metastases in DTC patients, thus increasing the odds of an indeterminate thyroid nodule with high-risk mutations being aggressive cancer.

64Thyroid scintigraphy allows the assessment of regional thyroid function and the detection of functioning nodules. For routine use, most often 99mTc is used, based on a combination of low cost, wide availability, and low radiation burden. However, in a minority, this approach may lead to the misclassification of hypofunctioning as hyperfunctioning nodules. Thyroid scintigraphy should be performed when serum TSH is suppressed or at the lower normal limits. Of note, in areas of current or previous iodine deficiency, hyperfunctioning nodules may also be seen in individuals with normal TSH, meriting the use of thyroid scintigraphy.

65While we do not suggest [99mTc]Tc-MIBI imaging for routine use, it may be of value in case of indeterminate cytology, based on its relatively high negative predictive value for malignancy. Similarly, in patients with indeterminate cytology, 18F-fluoro-2-deoxy-D-glucose positron emission tomography ([18F]FDG-PET)/CT, although still debated, has shown promising results for excluding malignancy. Routine use of this resource, however, is limited in clinical practice by cost and limited accessibility.

66Thyroid scintigraphy provides useful information in:

67solitary hyperfunctioning nodules, to avoid FNA biopsy, as hyperfunctioning nodules are rarely malignant;

68multinodular goiter, to differentiate hypofunctioning nodules suitable for FNA from hyperfunctioning lesions that do not need cytologic evaluation;

69to determine the eligibility for radioiodine therapy.

70While there is no indication for initial evaluation of nodular thyroid disease using cross-sectional (i.e. contrast-enhanced CT and magnetic resonance imaging (MRI)) or functional (e.g. [18F]FDG-PET/CT) studies, incidentally detected thyroid lesions do have a ROM of 5–13% when using CT and MRI and of about 35% of high activity lesions when using [18F]FDG-PET/CT. These types of nodules should be investigated according to the diagnostic workup proposed in this guideline.

71Neck and upper mediastinal CT scan should be performed in case of US or clinical suspicion of substernal extension. If using contrast media, the risk of thyrotoxicosis should be considered. The aim is to assess the extension into the upper mediastinum, the location and dimensions of the trachea, and the anatomical relation of the goiter to the oesophagus.

72Most benign thyroid nodules are incidentally diagnosed and asymptomatic. In the absence of elevated TSH, the use of thyroid hormone in order to decrease TSH should be discouraged in order to limit the increased morbidity and mortality seen with such therapy, but mainly because of its lack of efficacy in adequately decreasing size in symptomatic nodules. Iodine as well as selenium deficiency is associated with increased goiter prevalence. However, neither iodine nor selenium supplementation is recommended in iodine and selenium replete populations. During 1 year, significant growth, defined as an increase ≥20% in at least two nodule diameters with a minimum increase of 2 mm, or nodule volume increase >50%, occur only in a minority of cases (over a period of 5 years it is about 16%). Thus, asymptomatic nodules should be followed based on the sonographic pattern and cytological assessment (Table 1, Fig. 2).

73An in-depth account is given in. Approximately 5–10% of solitary/dominant thyroid nodules are functioning on thyroid scintigraphy, with suppression of the peri-nodular thyroid tissue. Such nodules are, with extremely rare exceptions, benign, should not be biopsied, and are eligible for RAI treatment. Most patients are euthyroid or subclinically hyperthyroid at the time of diagnosis. Unless severely hyperthyroid or with cardiac comorbidity, individuals rarely need pre-treatment with anti-thyroid drugs. RAI is most often given as a fixed activity (e.g. 185–370 MBq), most often achieves euthyroidism, may cause hypothyroidism, and reduces nodule size by 30–50% in 12 months. Life-long follow-up is recommended.

74The diagnostic workup of non-hyperfunctioning multinodular goiters, including FNA, should accord with the previously described algorithm (Figs. 1 and 2). When symptomatic and benign, thyroid nodules may, as an alternative to surgery, be eligible for RAI, especially in case of patients at surgical risk. Most patients are offered fixed activity RAI, dependent on local regulations. Hypothyroidism is rare (10–20% after 10 years), but life-long follow-up is recommended. Thyroid volume is typically reduced by 40% within 12 months and alleviates symptoms in most. In case of low RAI uptake and/or a large goiter, prestimulation with rhTSH has been demonstrated to augment thyroid volume reduction by 35% and increases the smallest cross-sectional area of the trachea, improves pulmonary function, and reduces pressure symptoms.

75MITs are out-patient procedures, performed under US guidance, for non-surgical management of thyroid lesions that cause local pressure symptoms or esthetic concerns. MITs include ethanol ablation (EA), based on the direct injection of ethanol into a cystic cavity, and TA techniques, which use various energy sources: laser, radiofrequency, microwaves, or high-intensity focused ultrasound. MITs result in a relevant and long-lasting decrease of nodule volume (57–77% at 5 years) that is paralleled by improvement of local symptoms and disease-related quality of life. These treatments do not require general anesthesia, rarely cause complications, and even more rarely thyroid dysfunction. EA is preferred as an effective, safe, and inexpensive treatment for cystic (or predominantly cystic) symptomatic thyroid nodules while TA procedures, due to their geometric and predictable volume of tissue destruction, are the first-line treatment for solid thyroid lesions. TA treatments should aim at a balance between nearly complete nodule destruction, to prevent late regrowth, and sparing of at-risk areas, to minimize the risk of complications. TA is an alternative to RAI and surgery in small hyperfunctioning nodules (<10 mL), while it performs poorer in larger nodules. In such cases, TA may be considered for patients who decline or are not candidates for RAI therapy or surgery. Currently, a major limitation of TA procedures use is their limited availability and lack of long-term data.

76For MIT procedure indications, refer to recommendations (Table 1). For a more detailed account of the use of TA in benign nodules or even very low-risk thyroid carcinomas, we refer to the recent pertinent guidelines.

77This section is focused on providing the indications for the use of surgery in the management of thyroid nodules, rather than defining the extent of surgery and exploring the currently available surgical techniques and their advantages and disadvantages.

78Currently, surgery is one of the possible management options in patients with nodular thyroid disease due to the development of sensitive diagnostic tools (US, FNA) enabling active surveillance and use of MIT. As detailed previously, any treatment decision, not only restricted to surgery, needs to take into consideration patient preference and available alternatives in a dialogue with the patient.

79According to the management algorithm of the present guideline (Fig. 2), surgery may be appropriate in the following scenarios:

80Symptomatic nodular thyroid disease, as an alternative option to MIT and RAI therapy.

81Nodules that have been classified as benign at cytology and/or low risk at US (i.e. EU-TIRADS 2 or 3) and become symptomatic over time.

82Nodules with indeterminate cytology (Bethesda class III and IV) that are not suitable for active surveillance (i.e. large size, high suspicion of malignancy on US, symptomatology).

83Nodules with a Bethesda class V and VI cytology.

84For nodules of uncertain malignant potential (Bethesda class III–V cytology), surgery allows for a definitive diagnosis. Molecular test results (if available) should be considered prior to operation. For diseases limited to one lobe, lobectomy/hemithyroidectomy is recommended. If such a disease is diagnosed in a nodular goiter, near-total thyroidectomy should be considered. In benign lesions (Bethesda class II cytology), even if asymptomatic, surgical treatment may be considered for nodules ≥4 cm (due to the ROM and increased probability of a false negative FNA), symptomatic nodules (airway or esophagus compression), in case of cosmetic concern, and retro-clavicular and mediastinal extension. Surgery is one of the available therapeutic options, besides RAI and MIT, for hyperfunctioning nodules and toxic multinodular goiter.

85Figures 1 and 2 and Table 1 summarize the diagnostic workup, the recommendations facing a newly diagnosed thyroid nodule and the following management strategies.

86The present document offers clinicians a guide to the rational management of nodular thyroid disease. Although very few nodules are malignant, the possibility of cancer represents the overriding concern when thyroid gland lesions are brought to our attention. Neck US is the central hub in the diagnostic algorithm aimed at stratifying the ROM, allowing guided FNA, and promoting the individualized management currently advocated. For the vast majority of patients, both tools provide a reliable foundation for defining the initial management strategy, whether follow-up, surgery or non-surgical treatment with MIT. However, in a number of cases, the estimates of the ROM will remain uncertain and the subsequent choices unsettled. In such cases, the clinicians’ experience and individualized choices based on patient-related factors and preferences represent the way forward. New tools are emerging in clinical practice, such as molecular testing, which are not yet generally available in Europe. While the analytic and clinical validity have been extensively studied, the ability of the molecular analysis to improve patient outcomes sufficiently to justify its incorporation into real-world clinical practice has yet to be proven. Such studies are warranted for years to come. The current focus is on developing an internationally endorsed lexicon of US findings allowing a common RSS, refining and implementing the rapidly evolving technology of non-surgical techniques, and involving disease-specific PRO tools, for example, the ThyPRO, in the management of our patients.

87Agnieszka Czarniecka, Gilles Russ, Fernando Schmitt, Paula Soares, and Tamas Solymosi have no conflicts of interest to disclose. Cosimo Durante has reported advisory board honoraria from EISAI, Eli Lilly, and Roche. Laszlo Hegedüs has reported consultancy fees from Berlin-Chemie, Horizon, IBSA, Lundbeck, Merck-Serono, and Novo Nordisk Foundation. Enrico Papini has reported advisory board honoraria from IBSA and Terumo and consultancy fees from Novo Nordisk. Ralf Paschke has reported ThyroSPEC™ license fees, received grant from Bayer, and advisory board honoraria from Bayer, Eisai, and Ipsen.

88This work did not receive any specific grant from any funding agency in the public, commercial, or not-for-profit sector.