Endocrine Abstracts (2015) 38 P437 | DOI: 10.1530/endoabs.38.P437

Reliability of thyroid ultrasounds in assessment of thyroid nodules

Salini Sumangala, Gurmit Gill, Paul Wilson, Zafar Hashim, Muthukumarasamy Balasubramaniam & Lakshminarayanan Varadhan


University Hospitals of North Midlands NHS Trust, Stoke-on-Trent, UK.


Aim: The BTA guidelines published recently suggested an uniform U1–U5 classification of thyroid nodules on ultrasound (US) and organising management and follow-up based on these grades. The aim of our retrospective analysis was to assess the reliability and usefulness of US in assessment of thyroid nodules, in comparison with post-operative histological diagnosis.

Methods: All patients who underwent thyroid surgery for thyroid nodules were assessed, with a review of pre-operative US over the last 2 years.The post-operative histology results were verified and confirmed by histopathologist. US images were reviewed and graded by radiologists, after being blinded for histology and previous radiological reports.

Results: Five hundred and ninty one cases were analysed. Based on the availability of the actual US images to report on grading of nodules, 88 patients were included in the study. 27/88 had malignancy proven on post-op histology. 45 patients were graded U2 and 43 were graded U3 or above.

Of the U2 nodules (n=45): based on histology: 36 were benign on histology (FNA was Thy1 or Thy2 in all these nodules); further four were follicular adenoma; five were malignant (two were incidental micropapillary carcinomas and three were overt malignant nodules) – FNA was Thy1 or Thy2 in all these nodules.

Of the U3-U5 nodules (n=43): based on histology: 18 were benign (ten were U3 and the rest U4); 10/18 had FNA being Thy1 or Thy2 as well; three were follicular adenomas; 22 were malignant nodules (five had Thy1 or Thy2 on FNA). The sensitivity and specificity of U3–U5 to diagnose malignancy was 82% and 66%; the same for U3–U5+ FNA being Thy3 or above was 93 and 80%.

Conclusions: i) A small but significant proportion (11%) of patients with U2 may still have malignancy and hence clinical correlation is required, ii) 42% of patients with U3–U5 could actually have a benign nodule, iii) inter-rater variability could be high among US reports and adequate expertise should be mandatory, and iv) concurrent use of FNA and US could improve sensitivity and specificity further.

Article tools

My recent searches

No recent searches.