Related Experiment Video
Updated: Dec 14, 2025
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
Approaching Indeterminate Thyroid Nodules in the Absence of Molecular Markers: "The BETH-TR Score"
Sushma Mehta1, Subramanian Kannan2
1Department of Head and Neck Surgical Oncology, Narayana Hrudhalaya Hospitals, Bengaluru, Karnataka, India.
Context:
Given the lack of easy access to molecular markers for indeterminate thyroid nodules (Bethesda (BETH) category III, IV), the clinician can either decide to get a second opinion from an expert high-volume thyroid cytopathologist, redo the FNAC after a period of 3-6 months, or send the patient for a diagnostic hemithyroidectomy. Reviewing the sonographic risk features is also one way of triaging these nodules. The ACR-TIRADS (TR) is an objective method of sonographic risk assessment and is superior to other forms of sonographic classification.
Aim:
We propose combining the scoring of the TR category and BETH category (both expressed as a numerical value and summated) and look at the score which could potentially guide the clinician in deciding whom to send for surgery.
Settings And Design:
Observational prospective collection of consecutive patient data from the thyroid FNAC clinic.
Statistical Analysis Used:
The BETH categories were represented numerically and summated with the TR category. The categorical outcome variables of benign and malignant nodules and the summated score was analyzed using the Kruskal-Wallis test.
Results:
We analyzed 450 FNAC data, out of which 403 were thyroid nodule aspirates. Out of these nodules, 96 of them underwent surgery and 64% of these nodules were malignant on final histopathology (malignant = 62 and benign = 34). The mean size of the benign nodules was 3.6 ± 2.2 cm compared to 2.8 ± 1.8 cm of the malignant nodules. After excluding those with BETH 1 (n = 4), the mean BETH-TR score for benign nodules was 6 ± 1.4 and malignant nodules 9.4 ± 2.1 (P < 0.0001). The BETH-TR score progressively increased from 7.3 ± 0.92 in follicular thyroid cancers (FTC) to 8.6 ± 1.4 in follicular variant papillary thyroid cancer (FVPTC) to 10 ± 1.3 in classic papillary thyroid cancers (PTC). Among the indeterminate nodules (BETH III and IV; n = 40), the BETH-TR score of benign nodules was 6.75 ± 1 and malignant nodules was 7.5 ± 0.72 (P value = 0.01). A BETH-TR score ≥7 gave a sensitivity of 92% specificity of 74% and correctly identified malignant nodules in 86% of cases (likelihood ratio 3.5; ROC area: 0.8841; CI 0.79-0.94).
Conclusion:
A combined sonocytological BETH-TR score is one way to triage the management of indeterminate thyroid nodules. A BETH-TR score ≥7 gave a sensitivity of 92% specificity of 74% and correctly identified malignant nodules in 86% of cases.
