Investigations in thyroid swelling

Reading File
Finding Sources
Finding Sources
Searching PubMed

"thyroid nodule"[MeSH Terms] AND "diagnosis"

Reading File
Reading File
Reading File
Searching Images

thyroid nodule ultrasound malignant microcalcification

This composite diagnostic image illustrates the multimodal ultrasound evaluation of Papillary Thyroid Carcinoma. Panel (a) is a two-dimensional B-mode and Color Doppler ultrasound showing a hypo-echoic, irregular thyroid nodule containing internal microcalcifications, appearing as punctate hyperechoic foci. Panel (b) presents Contrast-Enhanced Ultrasound (CEUS) findings, demonstrating heterogeneous low enhancement within the lesion compared to the surrounding thyroid parenchyma, suggesting irregular vascularization and complex tissue composition. Panel (c) shows Virtual Touch Tissue Quantification using Acoustic Radiation Force Impulse (ARFI) imaging to measure Shear Wave Velocity (SWV). The panel displays a region of interest (ROI) centered on the nodule to assess tissue stiffness. These imaging modalities are combined to differentiate malignant from benign thyroid nodules by evaluating echogenicity, microcalcification presence, vascular perfusion patterns, and mechanical tissue elasticity (stiffness).

This composite diagnostic image illustrates the multimodal ultrasound evaluation of Papillary Thyroid Carcinoma. Panel (a) is a two-dimensional B-mode and Color Doppler ultrasound showing a hypo-echoic, irregular thyroid nodule containing internal microcalcifications, appearing as punctate hyperechoic foci. Panel (b) presents Contrast-Enhanced Ultrasound (CEUS) findings, demonstrating heterogeneous low enhancement within the lesion compared to the surrounding thyroid parenchyma, suggesting irregular vascularization and complex tissue composition. Panel (c) shows Virtual Touch Tissue Quantification using Acoustic Radiation Force Impulse (ARFI) imaging to measure Shear Wave Velocity (SWV). The panel displays a region of interest (ROI) centered on the nodule to assess tissue stiffness. These imaging modalities are combined to differentiate malignant from benign thyroid nodules by evaluating echogenicity, microcalcification presence, vascular perfusion patterns, and mechanical tissue elasticity (stiffness).

This composite diagnostic image displays ultrasound scans of a thyroid nodule analyzed by two deep learning models, ResNet50 (top) and DenseNet201 (bottom), for TI-RADS classification. The panel includes original grayscale B-mode images, binary nodule masks, and Grad-CAM heatmaps showing regions of interest for benign versus malignant predictions. The nodule is characterized by a mixed solid-cystic composition with hyperechoic regions and largely smooth margins, leading to a TI-RADS 2 (benign) classification. Detailed heatmap sub-panels evaluate specific ACR TI-RADS categories: Composition (Cystic, Spongiform, Mixed Cystic, Solid), Echogenicity (Anechoic, Hypoechoic, Hyperechoic), Margin (Smooth, Ill-defined, Irregular, Extension), and Echogenic Foci (Macrocalcification, Microcalcification). Each sub-panel provides a quantitative confidence score for the feature identification. The Grad-CAM overlays illustrate that the models focus on the internal solid components and the interface of the nodule borders to differentiate benign and malignant features, demonstrating the utility of explainable AI in radiological decision-making for thyroid oncology.

This composite diagnostic image displays ultrasound scans of a thyroid nodule analyzed by two deep learning models, ResNet50 (top) and DenseNet201 (bottom), for TI-RADS classification. The panel includes original grayscale B-mode images, binary nodule masks, and Grad-CAM heatmaps showing regions of interest for benign versus malignant predictions. The nodule is characterized by a mixed solid-cystic composition with hyperechoic regions and largely smooth margins, leading to a TI-RADS 2 (benign) classification. Detailed heatmap sub-panels evaluate specific ACR TI-RADS categories: Composition (Cystic, Spongiform, Mixed Cystic, Solid), Echogenicity (Anechoic, Hypoechoic, Hyperechoic), Margin (Smooth, Ill-defined, Irregular, Extension), and Echogenic Foci (Macrocalcification, Microcalcification). Each sub-panel provides a quantitative confidence score for the feature identification. The Grad-CAM overlays illustrate that the models focus on the internal solid components and the interface of the nodule borders to differentiate benign and malignant features, demonstrating the utility of explainable AI in radiological decision-making for thyroid oncology.

Searching the Web

Bethesda System thyroid cytopathology 2023 categories risk of malignancy ATA 2025 thyroid nodule guideline

Investigations in a thyroid swelling

The aim is to answer three questions in order. Is the gland functioning normally? What is the structure, and is any nodule suspicious? Is there cytological or histological evidence of cancer? Bailey and Love lists TSH, thyroid autoantibodies and FNAC of a discrete swelling as the essential tests (Bailey and Love's Short Practice of Surgery, 28th ed., Summary box 55.1).

1. Biochemical (thyroid function)

  • Serum TSH is the first-line test. If it is abnormal, add free T4 and T3.
    • Low TSH suggests thyrotoxicosis, including an autonomous ("hot") nodule.
    • High TSH suggests hypothyroidism, such as Hashimoto thyroiditis.
  • Free T4 and free T3 confirm and grade dysfunction. Total T3 matters in T3 thyrotoxicosis, where T4 is normal (Schwartz's Principles of Surgery, 11th ed., p. 1661). Total T4 is affected by binding proteins, for example in pregnancy, with oestrogen use or in nephrotic syndrome, so free T4 is more reliable.
  • Thyroid autoantibodies (anti-TPO and anti-thyroglobulin) point to autoimmune thyroiditis. About 80% of patients with Hashimoto thyroiditis have raised levels. They do not determine thyroid function (Schwartz, p. 1661).
  • TSH receptor antibodies (TRAb/TSI) are used for Graves disease.
  • Serum thyroglobulin is not a diagnostic test for a swelling. It is a tumour marker for follow-up of differentiated cancer after total thyroidectomy and radioiodine ablation. Anti-Tg antibodies interfere with the assay, so measure them at the same time (Schwartz).
  • Serum calcitonin (CEA is an alternative) is optional. Use it if medullary carcinoma is suspected, for example with a family history, MEN 2, or an FNAC result suggesting MTC.
  • Corrected serum calcium is optional. It is useful if parathyroid disease is possible and before surgery.
  • Phaeochromocytoma screening (plasma or urinary metanephrines) is needed before surgery if MTC or MEN 2 is suspected.

2. Imaging

Ultrasound is the workhorse investigation for the surgeon (Bailey and Love).
  • It describes the number, size, shape, margins, vascularity and calcification of nodules.
  • It assesses the regional lymph nodes, especially in the lateral neck.
  • It guides FNA, which is more accurate than free-hand sampling.
  • It involves no ionising radiation and is cheap and non-invasive.
  • It cannot see well behind the sternum.
Features suggesting malignancy (Scott-Brown's Otorhinolaryngology, Box 60.6, in decreasing order of specificity):
  • microcalcification
  • irregular border
  • increased intranodular flow on Doppler
  • hypoechogenicity
  • absent halo
  • a taller-than-wide shape (from ATA/TI-RADS rather than this box)
Features suggesting a benign nodule:
  • simple cyst with no solid component
  • comet-tail artefact (colloid)
  • honeycomb or spongiform appearance
Ultrasound of papillary thyroid carcinoma showing a hypoechoic, irregular nodule with microcalcifications
Ultrasound findings are graded with a risk system (ATA pattern or ACR TI-RADS). The grade and the nodule size together decide whether FNA is needed.
Radionuclide scan (Tc-99m pertechnetate or I-123):
  • It is indicated mainly when TSH is suppressed, to see whether a nodule is hot (autonomous, rarely malignant) or cold.
  • Cold nodules carry a higher malignancy risk, but most cold nodules are benign, so the scan is not a good test for cancer.
  • Bailey and Love also list it when a discrete swelling and toxicity coexist.
  • It can also show retrosternal goitre and ectopic thyroid tissue.
Chest X-ray and thoracic inlet views show tracheal deviation or compression and retrosternal extension.
CT or MRI are not routine. They are used for:
  • retrosternal goitre, to define its extent
  • invasive or fixed tumours
  • known cancer, to stage nodes and look for mediastinal and lung disease
  • some reoperations
CT is the modality of choice for retrosternal extension. If iodinated contrast is used, delay radioiodine therapy.
PET-CT has a limited role. It is used in selected cases, such as radioiodine-negative, thyroglobulin-positive recurrence or aggressive disease.

3. Cytology and histology

FNAC (ultrasound-guided) is the key test for a discrete swelling or a dominant nodule. Cytology is reported using the Bethesda System (TBSRTC, 3rd ed., 2023). The risk of malignancy (ROM) figures below are approximate, taken from the literature and web sources rather than the library textbooks.
CategoryTypical ROMUsual action
I Non-diagnosticVariableRepeat FNA under ultrasound guidance
II BenignLow (about 3%)Clinical and ultrasound follow-up
III AUS (atypia of undetermined significance)IntermediateRepeat FNA, molecular testing or lobectomy
IV Follicular neoplasmHigherMolecular testing or diagnostic lobectomy
V Suspicious for malignancyAbout 74%Lobectomy or total thyroidectomy
VI MalignantAbout 97%Surgery
Other points on FNAC:
  • Follicular carcinoma cannot be diagnosed on FNAC, because it needs histology to show capsular or vascular invasion. Any follicular lesion therefore needs diagnostic lobectomy.
  • Cytology can diagnose papillary, medullary and anaplastic carcinoma and lymphoma.
  • Core biopsy is useful for suspected lymphoma or anaplastic carcinoma.
  • Molecular testing on indeterminate cytology (BRAF, RAS, RET/PTC and others) helps with risk stratification.
  • Diagnostic hemithyroidectomy gives the final histology when cytology is indeterminate.
Laryngoscopy (direct or indirect) assesses vocal cord mobility. Do it pre-operatively if there is hoarseness, a malignant FNAC, a reoperation, or a retrosternal goitre. In practice it is done for every patient before thyroid surgery, and again afterwards if there is a voice change.

4. Practical approach to a thyroid swelling

  1. Take the history and examine the patient, including the neck nodes. Look for risk factors: neck irradiation, family history, rapid growth, hoarseness, dysphagia, and a hard fixed nodule.
  2. Measure TSH.
    • Low TSH: do a radionuclide scan. A hot nodule is treated as hyperthyroidism and usually does not need FNA.
    • Normal or high TSH: do ultrasound, then FNA according to the risk grade and size.
  3. Add imaging as needed: CXR or CT for a retrosternal goitre or compressive symptoms, and calcitonin if MTC is suspected.
  4. Do laryngoscopy before surgery.

Recent evidence check

A PubMed search of the last two years found three systematic reviews and meta-analyses. They were on clinical predictors of malignancy in thyroid nodules (PMID 42718662), temporal trends in nodule size on ultrasound (PMID 39446361), and incidental thyroid nodules on CT (PMID 39283819, with data on prevalence, follow-up and malignancy risk). I did not read these in full. None looks likely to overturn the approach above, but they may refine risk stratification and the follow-up of incidental nodules. The 2025 ATA guideline for differentiated thyroid cancer has also been published (Thyroid 2025;35:841). I did not review it in detail, so check current local protocols for the exact size thresholds that trigger FNA.
This is a shared conversation. Sign in to Orris to start your own chat.