Comparing Two Approaches to Thyroid Nodule Testing: A Randomized Trial of Selective Versus Non-Selective Fine Needle Aspiration

Table of Contents

• Key Points
• Background: Why This Research Matters
• How the Study Was Conducted
• Randomization and Treatment Groups
• Ultrasound Examination and EU-TIRADS Scoring
• Fine Needle Aspiration Cytology
• Clinical Workup and Surgery
• Statistical Methods
• Key Findings: Study Results
• Surgical Outcomes
• Repeat Examinations
• Discussion: What the Results Mean
• Clinical Implications for Patients
• Limitations of the Study
• Recommendations and Future Directions
• Frequently Asked Questions
• Source Information

Key Points

• In a trial of 195 patients, using EU-TIRADS to select nodules for FNA reduced procedures from 83% to 71% of patients.
• The selective approach found suspicious nodules in 26% versus 13% with non-selective FNA, but cancer rates were similar (8% versus 5%).
• Only about 10% of patients could safely avoid FNA based on EU-TIRADS alone, so most still needed the procedure.
• The trial was not large enough to definitively prove that selective FNA does not miss cancers; larger studies are needed.
• Discuss with your doctor whether EU-TIRADS-based selection is appropriate for your thyroid nodule and attend all follow-up appointments.

Background: Why This Research Matters

Thyroid nodules are very common in adults. The vast majority of these nodules are benign, meaning they are not cancer. However, doctors need a reliable way to identify the small number of nodules that might be cancerous.

For many decades, the standard method for evaluating thyroid nodules has been fine needle aspiration (FNA) cytology. This procedure uses a thin needle to remove cells from a thyroid nodule. A pathologist then examines these cells under a microscope to check for signs of cancer.

Ultrasound is the first-line imaging test for thyroid nodules. It is non-invasive, widely available, and cost-effective. High-resolution ultrasound can detect thyroid nodules at rates close to what has been found in autopsy studies. Over the past 25 years, the use of cervical ultrasound (ultrasound of the neck) in thyroid disease has gradually increased. This has led to more thyroid nodules being detected.

To assess the risk of thyroid malignancy (cancer), several ultrasound risk stratification systems (RSS) have been developed. The first version of TIRADS (Thyroid Imaging Report and Data System) was published in 2009 by Horvath and colleagues. However, this original version was difficult to use in everyday clinical practice because it was very complex.

More recent versions of TIRADS, including EU-TIRADS, have been developed to be easier to use in clinical practice. The main goal of these systems is to reduce unnecessary FNA procedures by identifying benign nodules that do not need to be sampled.

Both retrospective studies (looking back at past data) and prospective studies (following patients forward in time) have shown that ultrasound risk stratification systems can identify benign lesions. These systems can safely omit FNA. No single system has been proven superior to others, although several studies comparing different ultrasound risk stratification systems have been published.

Before this study, no randomized controlled trials (RCTs) had compared thyroid risk stratification systems to non-selective FNA. A randomized controlled trial is considered the strongest type of clinical study because it randomly assigns patients to different treatment groups.

In 2017, Sweden introduced a national standardized care pathway (SCP) for thyroid cancer. The purpose was to ensure uniform and prompt evaluation, including ultrasound and FNA, for patients with signs or symptoms that could indicate thyroid cancer. In Western Sweden, healthcare providers decided to introduce a thyroid ultrasound risk stratification system with selective FNA, performed as a "one-stop-shop" procedure. EU-TIRADS was then implemented in the Swedish National Guidelines for thyroid cancer in 2021.

This multi-centre regional study aimed to investigate whether EU-TIRADS can identify benign nodules and thereby correctly omit FNA. The primary outcome was the proportion of cytology results suspicious of thyroid cancer (classified as Bethesda IV–VI). In the selective group, low-risk nodules did not undergo cytology. Therefore, the researchers expected the frequency of cytologically suspicious nodules to be higher in this group. The secondary outcome was the rate of malignancy (cancer). The researchers hypothesized that there would be no difference in cancer rate between the two groups.

How the Study Was Conducted

This was a regional, multi-centre, unblinded, interventional prospective randomized trial. It compared selective and non-selective FNA of thyroid nodules in the thyroid cancer diagnostic programme in Western Sweden.

Western Sweden is one of six healthcare regions in Sweden, with a population of approximately 1.7 million people in 2023. During the study period, diagnostic workup according to the standardized care pathway was available only through the public healthcare system in Western Sweden.

Four hospitals participated in the study:

• Sahlgrenska University Hospital (SU), a university hospital
• Northern Älvsborg Hospital (NÄL), a regional hospital
• Southern Älvsborg Hospital (SÄS), a regional hospital
• Kungälv Hospital (KS), a local hospital

These hospitals cover a geographical area with approximately 1.5 million inhabitants.

Inclusion criteria: All patients in Western Sweden who were 18 years of age or older and referred to the radiology department for ultrasound-guided FNA were eligible for the study. Patients were referred because of signs or symptoms that could indicate thyroid cancer according to the standardized care pathway.

The signs and symptoms that qualified a patient for referral were:

1. A previously undetected palpable nodule (a lump that can be felt) in the thyroid
2. A growing palpable nodule in the thyroid
3. An incidental focal PET-positive uptake in the thyroid (a finding on a PET scan that was done for another reason)

Exclusion criteria: Patients were excluded if they had a previous history of thyroid cancer or previous thyroid surgery.

Power calculation: The researchers performed a power calculation to determine how many patients they needed to include. A power calculation estimates the number of participants required to detect a meaningful difference between groups. They used a historic control group of patients who underwent non-selective FNA without ultrasound risk stratification between 2013 and 2017.

In the historical data, the frequency of Bethesda IV–VI was 46% in operated patients. Preliminary data came from patients referred according to standardized care pathway criteria. These patients were examined with selective FNA according to EU-TIRADS between March 2018 and October 2019. The data showed a frequency of Bethesda IV–VI of 69%. This yielded an effect size of 23%.

The researchers estimated power at 0.8 and significance level at 0.05. This required a total of 149 observations. To avoid being underpowered and to compensate for dropouts, the researchers decided to include 200 patients in total.

For the secondary outcome of malignancy rate, a non-inferiority power calculation showed that a minimum of 600 patients would be required. The researchers considered this number impossible to achieve in a reasonable time frame.

Initially, Bethesda III was included in the primary endpoint when the trial was first registered. However, before the trial began, the second edition of the Bethesda classification was implemented in clinical routine. As a result, Bethesda III became much less influential in clinical management and was removed from the primary endpoint.

Patient inclusion was planned between February 2022 and September 2023. The study period was prolonged by 3 months because only 170 patients had been included by September 2023. In December 2023, 195 patients were randomized. This was five fewer than planned but more than the power calculation estimated. The researchers decided to close inclusion due to slow recruitment.

Randomization and Treatment Groups

Before examination, patients provided oral and written informed consent. Patients were randomized 1:1 to either an intervention (selective) group or a control (non-selective) group. Randomization used random permuted blocks with an online tool. Staff at the radiology department performed the randomization before the examination.

In both groups, all low-risk and intermediate-risk nodules (EU-TIRADS 4 or lower) that were 1 cm or larger were recorded in a case report form (CRF). Nodules larger than 0.5 cm in diameter classified as EU-TIRADS 5 were also recorded.

In the intervention group, the indication for FNA was identical to EU-TIRADS recommendations. In the control group, all nodules larger than 1 cm scoring EU-TIRADS 4 or lower underwent FNA in up to five nodules per patient. Nodules scoring EU-TIRADS 5 underwent FNA regardless of size if technically possible. If a patient had more than five nodules, the five most suspicious nodules underwent FNA.

Ultrasound Examination and EU-TIRADS Scoring

Patients were referred from primary care and hospital-based services. The primary examination of included patients took place between February 2022 and December 2023. Patients were examined at the radiology department at four participating public hospitals.

Ultrasound was performed at the four centres using different ultrasound scanners:

• Sahlgrenska University Hospital: Canon Aplio i800 ultrasound scanner with a 14L5 linear array transducer probe
• Northern Älvsborg Hospital: Canon Aplio i700 ultrasound scanner with a 14L5 linear array transducer probe
• Kungälv Hospital and Southern Älvsborg Hospital: Philips EPIQ 7 ultrasound system

Thyroid nodules were classified according to the EU-TIRADS system. Ultrasound images of relevant nodules with measurements in three dimensions were stored. The size of the thyroid gland was also recorded. Additionally, 10 MHz transverse scan sequences over both thyroid lobes were stored to enable secondary assessment if needed.

In addition to routine documentation of thyroid nodules and cervical lymph nodes, the radiologist filled in a case report form directly after the examination for study purposes.

Fine Needle Aspiration Cytology

All ultrasound examinations were carried out with "one-stop" FNA, meaning the FNA was performed during the same visit as the ultrasound.

In the control group, according to the study protocol, all solid or partially solid nodules (EU-TIRADS 3 or higher) that were larger than 1 cm underwent FNA. Nodules smaller than 1 cm scoring EU-TIRADS 5 underwent FNA if possible.

In the intervention group, nodules underwent FNA according to EU-TIRADS criteria.

Nodule puncture was performed with a 22G needle with vacuum applied using a syringe. All FNAs were ultrasound-guided and performed by the radiologist. Two board-certified cytologists evaluated the samples. The 2017 Bethesda system for reporting thyroid cytology was used consistently throughout the study.

Clinical Workup and Surgery

An endocrine surgeon at Sahlgrenska University Hospital or Northern Älvsborg Hospital determined the clinical management of patients undergoing FNA. The surgeon based decisions on results from ultrasound and FNA.

All surgical procedures and histopathologic examinations were performed at Sahlgrenska University Hospital and Northern Älvsborg Hospital.

Indications for surgery included:

• FNA suspicious for or showing thyroid cancer
• Benign goitre (an enlarged thyroid gland)
• Hyperthyroidism (an overactive thyroid gland)

Patients with Bethesda category IV or higher were offered surgery within 31 days. The type of surgery was either hemi-thyroidectomy (removal of half the thyroid) or total thyroidectomy (removal of the entire thyroid), according to Swedish national guidelines.

Central lymph node dissection (removal of lymph nodes near the thyroid) was performed in some patients. These patients had clinically suspicious lymph nodes, FNA-verified lymph node metastasis, or a primary tumour larger than 4 cm. Lateral lymph node dissection was performed when ultrasound suggested metastatic spread, usually confirmed by FNA.

Nodules with indeterminate FNA results (Bethesda I and III) and nodules yielding EU-TIRADS 5 with Bethesda II were referred to follow-up by ultrasound and FNA after 1–6 months. This follow-up occurred unless surgery was indicated due to compression symptoms.

Microcarcinomas (very small cancers) detected after surgery and not recorded in the case report form were not included in the statistical analysis. Borderline tumours were regarded as benign.

All patients with a postoperative cancer diagnosis underwent secondary review of FNA and histopathology at the department of pathology at Sahlgrenska University Hospital. These patients received further treatment recommendations based on discussion at a multidisciplinary tumour board.

Statistical Methods

The researchers presented descriptive statistics as total numbers and percentages. They used the chi-square test, Wilcoxon rank sum test, and Student's t-test to compare the two groups. They performed calculations using R, version 4.3.2.

The study was approved by the Swedish Ethical Review Authority (reference number 2021-01490).

Key Findings: Study Results

Altogether, 311 patients were offered inclusion in the study. One hundred and sixteen patients declined participation. Of the 195 patients who accepted to participate, 93 patients (48%) with 150 nodules (58.1%) were randomized to the control (non-selective) group. One hundred and two patients (52%) with 108 nodules (41.8%) were randomized to the intervention (selective) group. There were no dropouts among included patients.

Patient characteristics:

• Female: 69 patients (74.2%) in the non-selective group, 81 patients (79.4%) in the selective group, 150 patients (76.9%) total
• Male: 24 patients (25.8%) in the non-selective group, 21 patients (20.5%) in the selective group, 45 patients (23.1%) total
• Mean age: 58 years (standard deviation 15) in the non-selective group, 55 years (standard deviation 15) in the selective group, 56 years (standard deviation 15) overall

Performing units:

• Sahlgrenska University Hospital: 40 patients (43.0%) in the non-selective group, 45 patients (44.1%) in the selective group, 85 patients (43.6%) total
• Northern Älvsborg Hospital: 29 patients (31.3%) in the non-selective group, 30 patients (29.4%) in the selective group, 59 patients (30.3%) total
• Kungälv Hospital: 24 patients (25.8%) in the non-selective group, 25 patients (24.5%) in the selective group, 49 patients (25.1%) total
• Southern Älvsborg Hospital: 0 patients (0%) in the non-selective group, 2 patients (2.0%) in the selective group, 2 patients (1.0%) total

Indications for examination:

• Palpable nodule or goitre symptoms: 60 patients (64.5%) in the non-selective group, 74 patients (72.5%) in the selective group, 134 patients (68.7%) total
• PET-positive uptake: 4 patients (4.3%) in the non-selective group, 4 patients (3.9%) in the selective group, 8 patients (4.1%) total
• Incidental imaging findings (CT, MR, or ultrasound): 29 patients (31.2%) in the non-selective group, 24 patients (23.5%) in the selective group, 53 patients (27.2%) total

Referral source:

• Primary care: 51 patients (54.8%) in the non-selective group, 58 patients (56.9%) in the selective group, 109 patients (55.9%) total
• Hospital-based: 42 patients (45.2%) in the non-selective group, 44 patients (43.1%) in the selective group, 86 patients (44.1%) total

Nodule characteristics:

• Total nodules: 150 (58.1%) in the non-selective group, 108 (41.9%) in the selective group, 258 total
• Mean nodule size: 24 mm (standard deviation 14) in the non-selective group, 29 mm (standard deviation 13) in the selective group, 26 mm (standard deviation 14) overall

FNA use: The frequency of patients undergoing FNA was 83% (77 patients) in the control (non-selective) group compared to 71% (72 patients) in the intervention (selective) group. When considering only patients with solid or partially solid nodules (EU-TIRADS 3 or higher), FNA was omitted according to study protocol in 9% of patients (7 patients) in the intervention group. No patients in the control group had FNA omitted.

In the intervention group, a mean of 0.9 FNA procedures per patient was performed compared to 1.5 in the control group (P=0.002). This difference was statistically significant.

Primary outcome — Bethesda category IV–VI: EU-TIRADS risk stratification for FNA selected patients with neoplastic lesions (Bethesda IV–VI) to a higher degree compared to the non-selective group. The frequency was 26% in the selective group versus 13% in the non-selective group (P=0.039). This was the primary endpoint of the trial.

The number of nodules in the intervention group yielding Bethesda IV–VI at primary examination was 21 (20%) compared to 10 (7%) for the control group (P=0.002).

Highest Bethesda category per patient (only patients undergoing FNA included):

• No FNA: 16 patients in the non-selective group, 30 patients in the selective group, 46 patients total
• Bethesda I: 17 patients (22.1%) in the non-selective group, 16 patients (22.2%) in the selective group, 33 patients (22.1%) total
• Bethesda II: 43 patients (55.8%) in the non-selective group, 32 patients (44.4%) in the selective group, 75 patients (50.3%) total
• Bethesda III: 7 patients (9.1%) in the non-selective group, 5 patients (6.9%) in the selective group, 12 patients (8.1%) total
• Bethesda IV: 7 patients (8.9%) in the non-selective group, 15 patients (20.8%) in the selective group, 22 patients (14.8%) total
• Bethesda V: 1 patient (1.3%) in the non-selective group, 3 patients (4.2%) in the selective group, 4 patients (2.7%) total
• Bethesda VI: 2 patients (2.6%) in the non-selective group, 1 patient (1.4%) in the selective group, 3 patients (2.0%) total

Cancer prevalence: The rate of malignancy was 5.4% (5 patients) in the non-selective group and 6.9% (7 patients) in the selective group. The overall cancer prevalence was 6.2% (12 patients).

Surgical Outcomes

Of the 195 included patients, 46 patients (23.6%) underwent surgery. The indication for surgery was ultrasonography and/or FNA suspicious of or showing malignancy in 20 of 30 patients (67%) in the intervention group. In the control group, the indication was ultrasonography and/or FNA suspicious of or showing malignancy in 10 of 16 patients (63%). No patients were operated on due to toxic goitre.

For the secondary endpoint of the trial, histopathology showed thyroid cancer in 8 patients (8%) in the intervention group versus 5 patients (5%) in the control group.

In the intervention group, no patients were diagnosed with a microcarcinoma. In the control group, one patient had papillary thyroid cancer microcarcinoma (EU-TIRADS 4, Bethesda VI).

Four patients (three in the intervention group and one in the control group) had incidental papillary thyroid cancer microcarcinomas upon final histopathology. These were not recorded in the case report form and thus were not included in the statistical calculations.

Repeat Examinations

In total, 42 patients underwent repeat examination. This included 19 patients (19%) in the intervention group and 23 patients (25%) in the control group.

Of these patients, eight finally underwent surgery due to suspicion of malignancy at repeat FNA. Histopathology showed benign results in all cases except one incidental papillary microcarcinoma. This microcarcinoma was not the indication for surgery.

The mean time between the first and second examination was 15 months.

Discussion: What the Results Mean

This randomized trial compared EU-TIRADS-based selective FNA (intervention) versus non-selective FNA (control) in a thyroid cancer diagnostic system. The study had three main findings.

First finding: The proportion of patients with FNA suspicious for or showing thyroid cancer was significantly higher in the intervention group compared with the control group. This difference was even more pronounced when accounting for individual nodules. Furthermore, the study demonstrated a 39% reduction in FNA procedures, indicating higher selectivity when using EU-TIRADS.

Second finding: The proportion of patients with solid lesions where FNA was omitted according to EU-TIRADS was 9% in the intervention group.

Third finding: The rate of malignancy was similar in the intervention and control groups. This supports the secondary hypothesis that diagnostic sensitivity is retained when using EU-TIRADS.

The primary aim of this study was to evaluate the performance of EU-TIRADS in selecting neoplastic nodules for FNA investigation. The frequency of Bethesda category IV–VI lesions was significantly higher in the intervention group (9%) compared to the control group (5%) when comparing individual nodules. This is the first randomized controlled trial to show that EU-TIRADS improves the selection of nodules for FNA.

When analyzing the effect of EU-TIRADS at the patient level, the frequency of patients with cytologically suspicious or malignant findings in the intervention group was higher. However, this difference was less pronounced compared to the analysis of individual nodules.

The researchers speculated that the power calculation was based on a cohort with a higher degree of symptomatic palpable nodules. In the current cohort, most patients had palpable goitre with a mean nodule size of 26 mm. However, patients with less obvious symptoms were also referred for suspected thyroid cancer and included in the study.

Consequently, a higher proportion of thyroid ultrasound examinations showed no focal lesions or ultrasound signs of thyroiditis, resulting in EU-TIRADS 1. For these patients, an ultrasound risk stratification system does not contribute to patient management. This raises the question of which patients should be referred for cervical ultrasonography and which groups benefit most from ultrasound risk stratification systems.

The Swedish national guidelines for referring patients to cervical ultrasonography according to the standardized care pathway represent an attempt to reduce the number of unnecessary examinations and possibly unnecessary treatments.

The frequency of FNA performed per patient was significantly lower in the group randomized to selective FNA according to EU-TIRADS. This finding is not surprising because all or most ultrasound risk stratification systems, including EU-TIRADS, have been developed mainly to identify benign nodules so that FNA may be safely avoided.

The many different ultrasound risk stratification systems currently used have varying degrees of ability to "rule out" benign nodules. However, the decision to recommend FNA is based on similar features, including:

• Echogenicity (how the nodule reflects sound waves)
• Shape
• Margins
• Presence of calcifications
• Nodule size

Modifications from the first TIRADS system have focused on simplifying the stratification system to enable clinical implementation. This comes at the expense of lower specificity (the ability to correctly identify patients without disease). The present trial illustrates that the performance of EU-TIRADS is affected by the patients examined. The experience of the examiner also seems to affect outcomes.

Looking forward, AI-guided pattern recognition could potentially improve diagnostic performance. A recently published review concluded that artificial intelligence or deep learning may improve thyroid nodule evaluation.

This randomized controlled trial supports the use of EU-TIRADS to correctly select neoplastic nodules for FNA without missing thyroid cancer. However, only 10% of patients were freed from suspicion of malignancy based on EU-TIRADS alone without need for FNA. This indicates a need for further refinement of risk stratification systems for thyroid cancer diagnostics.

Clinical Implications for Patients

This study has several important implications for patients with thyroid nodules.

Fewer unnecessary procedures: Using EU-TIRADS to guide decisions about FNA reduced the number of patients undergoing FNA from 83% to 71%. This means fewer patients needed to have a needle biopsy of their thyroid nodule.

Better selection of suspicious nodules: Among patients who did undergo FNA, the selective approach identified a higher proportion of nodules that were suspicious for or showed cancer (26% versus 13%). This means the FNA procedures were more likely to find concerning results when they were performed.

No increase in missed cancers: The rate of thyroid cancer found was similar between the two groups (8% in the selective group versus 5% in the non-selective group). This suggests that the selective approach did not miss cancers that the non-selective approach would have found.

Limitations of current risk systems: Despite these benefits, only about 10% of patients could safely avoid FNA based on EU-TIRADS alone. This means that most patients still needed FNA to rule out cancer. The researchers concluded that current risk stratification systems need further improvement.

What this means for patients: If you have a thyroid nodule, your doctor may use an ultrasound risk scoring system like EU-TIRADS to decide whether you need an FNA. This approach can help avoid unnecessary procedures while still detecting cancers. However, you should discuss with your doctor whether this approach is appropriate for your specific situation.

Limitations of the Study

The researchers acknowledged several limitations of this study.

Sample size: The study included 195 patients, which was five fewer than the planned 200 patients. The researchers closed inclusion due to slow recruitment. For the secondary outcome of malignancy rate, a non-inferiority power calculation showed that a minimum of 600 patients would be required. The study was not large enough to definitively prove that the two approaches were equivalent in terms of cancer detection.

Unblinded design: This was an unblinded study, meaning the researchers and participants knew which group they were in. This could potentially introduce bias in how outcomes were assessed.

Patient population: The study population had a higher degree of symptomatic palpable nodules than expected. Many patients had palpable goitre with a mean nodule size of 26 mm. The results might differ in patients with less obvious symptoms or smaller nodules.

Examiner experience: The experience of the examiner seems to affect outcomes. The study did not analyze how examiner experience influenced the results.

Microcarcinomas: Microcarcinomas detected after surgery and not recorded in the case report form were not included in the statistical analysis. Four patients had incidental papillary thyroid cancer microcarcinomas that were not included in the calculations.

Generalizability: The study was conducted in Western Sweden, and results might not apply to other populations or healthcare systems.

Recommendations and Future Directions

Based on the findings of this study, the researchers made several recommendations.

For clinical practice:

1. EU-TIRADS can be used to select neoplastic nodules for FNA without missing thyroid cancer.
2. Using EU-TIRADS reduces the number of unnecessary FNA procedures.
3. However, only about 10% of patients can safely avoid FNA based on EU-TIRADS alone.
4. Further refinement of risk stratification systems is needed to improve their performance.

For future research:

1. Larger studies are needed to definitively prove that selective FNA does not miss cancers.
2. Research should investigate which patients benefit most from ultrasound risk stratification systems.
3. Studies should examine how examiner experience affects outcomes.
4. AI-guided pattern recognition and deep learning should be investigated as tools to improve thyroid nodule evaluation.
5. Further refinement of risk stratification systems for thyroid cancer diagnostics is needed.

For patients:

1. Discuss with your doctor whether EU-TIRADS-based selection is appropriate for your thyroid nodule.
2. Understand that most thyroid nodules are benign, but FNA may still be needed to rule out cancer.
3. Ask your doctor about the benefits and limitations of ultrasound risk stratification systems.
4. If you have a thyroid nodule that is being followed with repeat examinations, attend all scheduled appointments.

Frequently Asked Questions

What is fine needle aspiration (FNA) and why might I need it for a thyroid nodule?

FNA is a procedure that uses a thin needle to remove cells from a thyroid nodule. A pathologist examines the cells under a microscope to check for cancer. It is the standard method for evaluating thyroid nodules. Most nodules are benign, but FNA helps identify the small number that might be cancerous. Your doctor may recommend FNA based on ultrasound findings and your symptoms.

What is EU-TIRADS and how does it affect whether I need an FNA?

EU-TIRADS is an ultrasound risk scoring system that classifies thyroid nodules by their appearance. It helps doctors decide which nodules should undergo FNA. In a trial of 195 patients, using EU-TIRADS reduced FNA procedures from 83% to 71% of patients. However, only about 10% of patients could safely avoid FNA based on EU-TIRADS alone, so most still needed the procedure.

Does using EU-TIRADS to select nodules for FNA miss cancers?

In a randomized trial of 195 patients in Western Sweden, the rate of thyroid cancer found was similar between selective and non-selective groups (8% versus 5%). This suggests the selective approach did not miss cancers. However, the study was not large enough to definitively prove the two approaches were equivalent in cancer detection, so more research is needed.

What do the Bethesda categories mean on my FNA report?

The Bethesda system classifies FNA results from I to VI. Categories IV, V, and VI are suspicious for or show thyroid cancer. In the trial, the selective group had a higher proportion of these suspicious results (26% versus 13%). Categories I and III are indeterminate and often lead to repeat testing. Your doctor will explain your specific category and what it means for your care.

If I have a thyroid nodule, will I definitely need surgery?

No. In the trial, only 46 of 195 patients (23.6%) underwent surgery. Surgery was recommended for reasons such as FNA results suspicious for cancer, benign goitre causing symptoms, or hyperthyroidism. Most patients with thyroid nodules do not need surgery. Your doctor will discuss whether surgery is appropriate based on your ultrasound, FNA results, and symptoms.

What happens during follow-up if my FNA result is indeterminate?

If your FNA result is indeterminate (Bethesda I or III), or if you have an EU-TIRADS 5 nodule with a benign (Bethesda II) result, you may be referred for repeat ultrasound and FNA after 1 to 6 months. In the trial, 42 patients had repeat examinations, and the mean time between first and second examination was 15 months. Attend all scheduled appointments.

Are there any risks or limitations of using ultrasound risk systems like EU-TIRADS?

The main limitation is that only about 10% of patients could safely avoid FNA based on EU-TIRADS alone, so most still needed the procedure. The trial was unblinded and included 195 patients, fewer than the 600 needed to definitively prove equivalent cancer detection. Examiner experience may also affect outcomes. Discuss benefits and limitations with your doctor.

If my doctor recommends fine needle aspiration for a thyroid nodule based on an ultrasound risk score like EU-TIRADS, when should I get a second opinion?

A second opinion can help when the decision to perform or skip fine needle aspiration rests on an ultrasound risk score. In a randomized trial of 195 patients, EU-TIRADS-based selection reduced FNA procedures from 83% to 71% and found suspicious nodules in 26% versus 13%, with cancer rates similar between groups (8% versus 5%). However, only about 10% of patients could safely avoid FNA based on EU-TIRADS alone, and examiner experience affects outcomes. A second review of your ultrasound and pathology can clarify whether sampling is warranted. Diagnostic Detectives Network provides independent expert second opinions.

Source Information

Original article title: A randomized controlled trial comparing non-selective versus selective TIRADS-based cytology in thyroid cancer diagnostics.

Authors: Dahlberg J, Carlqvist J, Örtoft A, Hammarstedt L, Aula E, Hellström M, Elias E, Muth A.

Author affiliations:

• Department of Surgery, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden
• Department of Surgery, Sahlgrenska University Hospital, Gothenburg, Sweden
• Department of Radiology, Institute of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden
• Department of Radiology, Sahlgrenska University Hospital, Gothenburg, Sweden
• Department of Radiology, Northern Älvsborg Hospital, Trollhättan, Sweden
• Department of Radiology, Kungälv Hospital, Kungälv, Sweden
• Department of Radiology, Southern Älvsborg Hospital, Borås, Sweden

Publication: BJS (British Journal of Surgery), 2026, Volume 113, Issue 7, znag076

DOI: https://doi.org/10.1093/bjs/znag076

Published: 20 June 2026

Trial registration: ClinicalTrials.gov (NCT05583097)

Funding: This study was funded by the LUA/ALF agreement, grants from the Gothenburg Medical Society, research funds at the department of surgery at Sahlgrenska University Hospital, and the Assar Gabrielssons Foundation.

Ethical approval: The study was approved by the Swedish Ethical Review Authority (2021-01490).

This patient-friendly article is based on peer-reviewed research.