Thyroid nodule: TI-RADS and who needs biopsy

A 48-year-old man has a carotid ultrasound to assess cardiovascular risk. The artery exam is normal, but the radiologist notes an extra finding: "solid isoechoic nodule of 14 mm in the right lobe of the thyroid". The patient goes to the clinician with the report in hand and a direct question: "Do I need a biopsy?"

The thyroid nodule is one of the most common findings in medicine. It appears on neck exams done for any reason, and the vast majority are benign. At the same time, the decision whether to puncture or not defines whether the patient enters a sequence of cytology, repeat testing, surgery and follow-up that can last years.

The TI-RADS systems were created to bring order to this decision. They translate the nodule's features on ultrasound into a risk category and link each category to a minimum size for puncture. This text shows how to read the report, what studies show about the real risk and how to avoid both the missed cancer and the unnecessary puncture.

Key points

• Thyroid nodules are very common, and cancer is the minority among them.
• Ultrasound is the central exam, and TI-RADS translates its features into risk.
• The indication for puncture combines the risk category with the nodule size.
• Isolated features, such as microcalcifications, have value, but none decides on its own.
• Using TI-RADS reduces unnecessary punctures without abandoning the search for clinically relevant cancer.

A finding almost everyone can have

The frequency of thyroid nodules depends on how you look for them. The review by the Korean Society of Thyroid Radiology cites a prevalence of 2 to 6 per 100 people on palpation, 19 to 68 per 100 on ultrasound and 8 to 65 per 100 in autopsies (Shin et al., 2016). The 2023 guideline of the European Thyroid Association states that up to 60 per 100 adults in the general population have one or more nodules (Durante et al., 2023).

The same European guideline notes that the frequency of cancer in unselected populations of nodules usually falls between 1 and 5 per 100, varying according to selection criteria and whether papillary microcarcinomas are included (Durante et al., 2023). The math is simple: many nodules, few cancers, and most cancers found are small and indolent.

The weight of imaging exams

A Canadian study notes that thyroid nodules may be present in up to 76 per 100 adults, with an estimated prevalence of 68 per 100 on ultrasound, 25 per 100 on CT and 18 per 100 on MRI (Abou Shaar et al., 2022). Every exam of the neck, chest or cervical spine is a chance to find a nodule that would never cause symptoms.

What the cancer numbers hide

Malignancy numbers vary widely between sources, and understanding why is part of the reasoning. The Korean review cites a malignancy rate of about 5 to 15 per 100 among nodules detected by palpation or ultrasound (Shin et al., 2016). These data, however, come largely from specialized centers, which receive patients who have already been selected.

A primary and secondary care cohort

A German cohort followed 17,592 patients with a nodule larger than 1 cm, assessed at a primary and secondary care center, with follow-up of up to 23 years (Grussendorf et al., 2022). The authors note that the risk of malignancy is usually estimated between 7 and 15 per 100 in the specialized center data used by guidelines.

The result was very different. Malignancy was confirmed in 189 patients, or 1.1 per 100, excluding papillary microcarcinomas found by chance (Grussendorf et al., 2022). Of the 189 diagnoses, 155 were diagnosed in the first year, 25 between the second and fifth year, 9 between the sixth and tenth year and none among the 1165 patients followed for more than 10 years.

This study illustrates well the base rate in diagnosis. The same nodule, seen in a general clinic or a referral center, carries different probabilities of cancer, because the population that reaches each place is different.

What ultrasound looks for

Ultrasound describes the nodule through a set of features: composition, echogenicity, shape, margins and echogenic foci. Some of them are strongly associated with malignancy.

The weight of each feature

A meta-analysis gathered 52 observational studies with 12,786 nodules and showed that all the features assessed were associated with malignancy in unselected nodules (Remonti et al., 2015). Microcalcifications, irregular margins and a taller-than-wide shape had high specificities, of 87.8, 83.1 and 96.6 per 100, with positive likelihood ratios of 3.26, 2.99 and 8.07, respectively.

The authors concluded that isolated features are not enough to diagnose thyroid cancer and that combining them is the way forward (Remonti et al., 2015). This is the principle behind all TI-RADS systems.

Calcifications deserve attention

The 2021 Korean guideline pays special attention to fully calcified nodules. According to the text, the risk of malignancy of these nodules, when they are 1 cm or larger, is between 18.4 and 23.3 per 100 (Ha et al., 2021). Calcification is not a synonym for benignity, and a report describing a "calcified nodule" does not end the conversation.

Interpreting these features requires the same care as interpreting signs, symptoms and history: each finding changes the probability, but the diagnosis comes from the whole picture.

When the nodule shows up on another exam

Many nodules do not come from thyroid ultrasound but from a chest CT, a spine MRI or a PET done for another reason. In these cases, the reasoning starts with a simple question: is the exam that found the nodule good for characterizing it?

Most of the time, the answer is no. The European guideline recommends limiting the use of CT and MRI in the workup of thyroid nodules to the assessment of local extension or retrosternal growth of a goiter (Durante et al., 2023). A nodule seen on CT should be characterized by ultrasound, and it is ultrasound that defines the risk category and the need for aspiration.

Ultrasound as the entry point

This recommendation has a practical consequence. Ordering an MRI or a new CT to "see better" a thyroid nodule rarely adds useful information and can delay the right exam. The Korean review describes ultrasound as the basis for detection, risk stratification and guidance of aspiration (Shin et al., 2016).

What to ask the patient

Some information from the history changes the probability before any imaging. Radiation exposure to the neck in childhood, family history of thyroid cancer, rapid growth of the nodule, new hoarseness and difficulty swallowing are data the report does not bring. In the Colombian study, agreement between ultrasound and cytology tended to be higher in patients with accelerated nodule growth, fixation to neighboring structures, vocal cord paralysis and a history of head and neck radiotherapy (Vargas-Uricoechea et al., 2017).

These findings do not enter the TI-RADS calculation, but they enter the reasoning. A low-risk nodule on ultrasound in a patient with new hoarseness and a palpable lymph node deserves another conversation. The system organizes the decision, but it does not replace clinical decision-making at the bedside.

What TI-RADS is

TI-RADS stands for Thyroid Imaging Reporting and Data System. The name does not refer to a single system but to a family of classifications created by different societies, all with the same idea: standardize the vocabulary of the report and associate each nodule pattern with a risk of cancer.

The European guideline recalls that these scores were proposed to improve reproducibility between examiners, quantify the risk of malignancy, unify reports and standardize management recommendations (Durante et al., 2023). The systems can be based on patterns, when the nodule is assessed as a whole, or on points, when each feature adds a score.

The main systems

The French TI-RADS is a five-level system with standardized vocabulary and reporting (Russ et al., 2016). In it, category 5 indicates a nodule that is very likely malignant, with a risk above 95 in 100, and category 4C indicates high risk, between 50 and 95 in 100. The European system, EU-TIRADS, the American ACR TI-RADS and the Korean K-TIRADS follow a similar logic, with differences in criteria and cutoffs.

The 2021 Korean guideline revised its biopsy criteria with the explicit aim of reducing unnecessary biopsies of benign nodules without losing sensitivity for cancer in small nodules, from 1 to 2 cm (Ha et al., 2021).

EU-TIRADS in practice

The 2023 European guideline summarizes the categories and estimated risks of malignancy in a simple table (Durante et al., 2023).

EU-TIRADS category	Ultrasound features	Estimated risk of malignancy	Biopsy indicated
EU-TIRADS 2, benign	Pure cyst or entirely spongiform nodule	0 in 100	No, unless treatment is planned
EU-TIRADS 3, low risk	Isoechoic or hyperechoic, with no sign of high suspicion	2 to 4 in 100	If larger than 20 mm
EU-TIRADS 4, intermediate risk	Mildly hypoechoic, with no sign of high suspicion	6 to 17 in 100	If larger than 15 mm
EU-TIRADS 5, high risk	Irregular shape or margins, microcalcifications or marked hypoechogenicity	26 to 87 in 100	If larger than 10 mm

Table source: Durante et al., 2023.

The same guideline recommends biopsy regardless of category when there are pathological lymph nodes in the neck (Durante et al., 2023). Minimum size is never the only criterion.

Who needs a biopsy

Fine needle aspiration biopsy is the test that provides the cytological diagnosis. It is safe, but it is not harmless: it produces indeterminate results, repeats and, sometimes, surgeries that were not necessary.

Risk and size go together

TI-RADS systems combine risk with size. In the European guideline, the cutoffs are more than 10 mm for EU-TIRADS 5, more than 15 mm for EU-TIRADS 4 and more than 20 mm for EU-TIRADS 3 (Durante et al., 2023). The logic is that low-risk nodules only justify biopsy when they are large, while suspicious nodules deserve biopsy earlier.

The impact of adopting the system

A Canadian community hospital compared biopsies performed before and after adopting ACR TI-RADS (Abou Shaar et al., 2022). Of the 57 nodules biopsied in the previous year, 30, or 52.6 in 100, would have been only followed if ACR TI-RADS had already been in use. The authors concluded that the system significantly reduces the number of unnecessary biopsies.

This finding connects with the problem of overdiagnosis. Each biopsy avoided in a low-risk nodule is one less chance of finding a microcarcinoma that would never harm the patient.

The nodule that will not be biopsied

The European guideline also guides the follow-up of asymptomatic nodules that are not sent for biopsy (Durante et al., 2023). In EU-TIRADS 2, nodules of 5 to 10 mm do not need additional evaluation, and those larger than 10 mm can be reassessed in 3 to 5 years. In EU-TIRADS 3 smaller than 20 mm, nodules of 5 to 10 mm also need no additional evaluation, and those of 10 to 20 mm can be reassessed in 3 to 5 years.

The Korean guideline proposes closer follow-up for the higher-risk categories that were not biopsied, with ultrasound every 6 months for 1 to 2 years in K-TIRADS 5 nodules and, after that, annually if there is no growth (Ha et al., 2021).

When TSH changes the path

Before biopsy, thyroid function enters the reasoning. The European guideline recommends scintigraphy when TSH is below normal, to identify functioning nodules and avoid biopsies (Durante et al., 2023). A hyperfunctioning nodule is rarely malignant, and biopsying it often produces confusing cytology.

TI-RADS and cytology speak to each other

TI-RADS does not replace cytology, but the two tests complement each other. Studies that compare ultrasound with the Bethesda cytological classification show how far this agreement goes.

Agreement between ultrasound and cytology

A Colombian study of 180 patients compared TI-RADS categories with Bethesda categories and found observed agreement of 87.2 per 100, with a weighted kappa of 0.69 (Vargas-Uricoechea and colleagues, 2017). Agreement was higher between the benign categories of the two systems.

When cytology is indeterminate

An indeterminate cytology result is one of the hardest moments of the investigation. A study of 140 operated nodules with indeterminate cytology showed that 74, or 52.9 per 100, were benign on histology (Barbosa and colleagues, 2019). In this group, malignancy increased with the ultrasound suspicion category, and the combination of Bethesda III with the lower-risk categories had a sensitivity of 95.3 per 100.

The authors concluded that ACR TI-RADS and the American Thyroid Association patterns help manage nodules with indeterminate cytology (Barbosa and colleagues, 2019). The European guideline adds that molecular tests may be considered in this situation, when available (Durante and colleagues, 2023).

The limits of the system

TI-RADS is a tool, and every tool has limits. Knowing them prevents blind trust in the number on the report.

Variation between examiners

An Italian study asked two raters to classify 1055 nodule images from 265 patients using five different systems (Grani and colleagues, 2018). Agreement between them in the first series, measured by Krippendorff's alpha, ranged from 0.47 to 0.61, depending on the system. Agreement on the indication for aspiration was higher, with kappa between 0.61 and 0.82.

The same study showed that agreement improved in the second series, after a joint reading to align the definitions (Grani and colleagues, 2018). Training matters. Reports from different examiners can diverge in the details, even when they agree on management.

The risk of overdiagnosis at scale

The rise in thyroid cancer worldwide is a topic of debate. A review of epidemiology points out that much of the increase appears to be apparent, linked to the detection of small cancers in the preclinical phase, although a real increase is also possible (Pellegriti and colleagues, 2013).

A Swiss population study showed that, between 1998 and 2012, the standardized incidence of thyroid cancer rose from 5.9 to 11.7 per 100 000 women and from 2.7 to 3.9 per 100 000 men (Jegerlehner and colleagues, 2017). The increase was restricted to the papillary subtype, the most indolent, mortality fell slightly and the thyroidectomy rate increased three to four times. It is the picture of too much diagnosis and too much treatment.

Building the reasoning step by step

The decision about a thyroid nodule can be organized into a short sequence, useful in the office and in case discussions.

1. Read the report and identify the TI-RADS category, the system used and the nodule size.
2. Check TSH: if it is low, consider scintigraphy before thinking about aspiration.
3. Look for signs that change the risk: suspicious lymph nodes, rapid growth, hoarseness, a history of neck radiation.
4. Apply the size cutoff for the category to decide on aspiration.
5. For the nodule that will not be aspirated, define follow-up according to the category.
6. Explain the reasoning to the patient, including the choice not to aspirate.

Back to the patient in the case

The 48-year-old patient has a 14 mm solid isoechoic nodule. With no other suspicious features described, this pattern is close to EU-TIRADS 3, whose estimated risk is 2 to 4 per 100 and whose cutoff for aspiration is more than 20 mm (Durante and colleagues, 2023). By the European guideline, the nodule would not need aspiration now, and follow-up could be done with a new assessment in 3 to 5 years.

Before closing the reasoning, it is worth ordering TSH, palpating the neck, asking about radiation and family history and checking whether the report describes the lymph nodes. If none of this changes the picture, the answer to the patient's question is that biopsy is not indicated at this time, and explaining why helps deal with diagnostic uncertainty.

Where the reasoning usually fails

The most frequent mistake is to perform a biopsy based on size without looking at the category, or out of anxiety without looking at the risk. The other mistake is the opposite: reassuring the patient without checking whether the report described the features that matter. Recording the category, the size and the plan in the medical record helps to organize the clinical reasoning and provides continuity of follow-up.

How NEXO helps train this reasoning

In NEXO, the case of the day is fictional, with a real clinical presentation and up to six progressive clues, and you propose the diagnosis by searching for the code in ICD-10.

The limits of use of the game are described on the NEXO economy page. If you research the teaching of clinical reasoning, the page for researchers explains how to get in touch. Professionals who write or review cases find information under curation and under creators, and institutions can learn about the proposal on the partnerships page. Training pretest probability and searching for the right code in ICD-10 complete this training.

Frequently asked questions

Does every thyroid nodule need a biopsy?

No. The indication depends on the ultrasound risk category and on size, and many low-risk nodules are only biopsied once they exceed a minimum size, such as the more than 20 mm of EU-TIRADS 3 (Durante et al., 2023). Purely cystic or spongiform nodules generally do not need a biopsy.

Which TI-RADS system is the best?

None has prevailed over the others. The systems have similar reproducibility between examiners, and agreement on the indication for biopsy is greater than agreement on isolated features (Grani et al., 2018). The most important thing is that the report states which system was used.

Is a large nodule more dangerous?

Size alone does not define the risk. The ultrasound features weigh more, and that is why the biopsy cutoff changes according to the category (Durante et al., 2023). Large nodules can cause compressive symptoms, which is another discussion.

What does indeterminate cytology mean?

It is the result in which cytology cannot classify the nodule as benign or malignant. In a study of operated indeterminate nodules, more than half were benign, and the ultrasound category helped separate those at lower risk (Barbosa et al., 2019).

Does finding more nodules mean saving more lives?

Not necessarily. The Swiss study showed an increase in the incidence of papillary cancer and in thyroidectomies, with slightly lower mortality in the period, a pattern consistent with overdiagnosis (Jegerlehner et al., 2017). Looking for cancer where the risk is low can cause more harm than benefit.

Take-home messages

• Thyroid nodules are very common, and cancer is the minority among them.
• The risk of cancer depends on the population served and is lower in general outpatient clinics than in referral centers.
• TI-RADS combines ultrasound features into risk categories linked to size cutoffs for biopsy.
• Low TSH calls for scintigraphy before considering a biopsy.
• Adopting TI-RADS reduces unnecessary biopsies.
• Explaining to the patient why not to perform a biopsy is part of the reasoning.

References

1. Abou Shaar B, Meteb M, Awad El-Karim G, et al. Reducing the Number of Unnecessary Thyroid Nodule Biopsies With the American College of Radiology (ACR) Thyroid Imaging Reporting and Data System (TI-RADS). Cureus. 2022. PMID 35425684.
2. Barbosa TLM, Junior COM, Graf H, et al. ACR TI-RADS and ATA US scores are helpful for the management of thyroid nodules with indeterminate cytology. BMC Endocr Disord. 2019. PMID 31664992.
3. Durante C, Hegedüs L, Czarniecka A, et al. 2023 European Thyroid Association Clinical Practice Guidelines for thyroid nodule management. Eur Thyroid J. 2023. PMID 37358008.
4. Grani G, Lamartina L, Cantisani V, et al. Interobserver agreement of various thyroid imaging reporting and data systems. Endocr Connect. 2018. PMID 29196301.
5. Grussendorf M, Ruschenburg I, Brabant G. Malignancy rates in thyroid nodules: a long-term cohort study of 17,592 patients. Eur Thyroid J. 2022. PMID 35635802.
6. Ha EJ, Chung SR, Na DG, et al. 2021 Korean Thyroid Imaging Reporting and Data System and Imaging-Based Management of Thyroid Nodules: Korean Society of Thyroid Radiology Consensus Statement and Recommendations. Korean J Radiol. 2021. PMID 34719893.
7. Jegerlehner S, Bulliard JL, Aujesky D, et al. Overdiagnosis and overtreatment of thyroid cancer: A population-based temporal trend study. PLoS One. 2017. PMID 28614405.
8. Pellegriti G, Frasca F, Regalbuto C, et al. Worldwide increasing incidence of thyroid cancer: update on epidemiology and risk factors. J Cancer Epidemiol. 2013. PMID 23737785.
9. Remonti LR, Kramer CK, Leitão CB, et al. Thyroid ultrasound features and risk of carcinoma: a systematic review and meta-analysis of observational studies. Thyroid. 2015. PMID 25747526.
10. Russ G. Risk stratification of thyroid nodules on ultrasonography with the French TI-RADS: description and reflections. Ultrasonography. 2016. PMID 26324117.
11. Shin JH, Baek JH, Chung J, et al. Ultrasonography Diagnosis and Imaging-Based Management of Thyroid Nodules: Revised Korean Society of Thyroid Radiology Consensus Statement and Recommendations. Korean J Radiol. 2016. PMID 27134526.
12. Vargas-Uricoechea H, Meza-Cabrera I, Herrera-Chaparro J. Concordance between the TIRADS ultrasound criteria and the BETHESDA cytology criteria on the nontoxic thyroid nodule. Thyroid Res. 2017. PMID 28184253.