Related Experiment Video
Updated: Jan 10, 2026
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
Multicenter Evaluation of Thyroid Nodule Size: Implications for FNA Diagnostic Accuracy and Malignancy Risk
Reem J Al Argan1, Dania M Alkhafaji1, Abdulmohsen H Al Elq1
1Department of Internal Medicine, College of Medicine, Imam Abdulrahman Bin Faisal University, King Fahad Hospital of the University, Dammam, Eastern Province, Saudi Arabia.
Background:
Nodular thyroid disease is increasingly prevalent, largely due to advances in imaging techniques. Key risk factors include female sex and older age. Thyroid nodules are clinically significant because they can cause thyroid dysfunction, lead to compressive symptoms if large, which can adversely affect patients' quality of life and, most importantly, pose a risk of malignancy. The malignancy rate in thyroid nodules ranges from 5% to 15%.
Objective:
This study aimed to investigate the association between thyroid nodules with clinical characteristics, imaging features, and FNAC accuracy. It also explored the relationship between nodule size, malignancy and outcome predictors.
Methods:
We retrospectively reviewed 733 thyroid nodules from 679 patients who underwent thyroid surgery between 2016 and 2022 at four healthcare institutions in Saudi Arabia's Eastern Province. Nodule sizes were categorised using ultrasound and histopathology. Malignancy risk, FNAC accuracy, and false-negative rates were compared. Statistical comparisons were made between nodules <4.0 cm and ≥4.0 cm, with multivariate analysis to identify independent predictors.
Results:
Nodules ≥4.0 cm were independently associated with benign pathology (OR 0.59, p = 0.004) but showed increased rates of extrathyroidal extension, follicular thyroid carcinoma, and rare malignancies. Multivariate analysis revealed higher odds of intermediate (OR 2.39, p = 0.003) and high (OR 5.33, p < 0.001) recurrence risk. Suspicious ultrasound features and Bethesda V-VI cytology were predictive of malignancy. FNAC performance in nodules ≥4.0 cm showed higher false-negative rates (40.0% vs. 25.9%), lower sensitivity (60.0% vs. 74.1%), and reduced accuracy (66.7% vs. 75.4%).
Conclusion:
Nodule size ≥4.0 cm is benign as well as associated with distinct diagnostic profiles and greater recurrence risk. Importantly, larger nodules are aggressive and demonstrate a substantially higher FNAC false-negative rate. Therefore, a careful evaluation by ultrasound risk stratification and consideration of diagnostic lobectomy are advisable, even with benign cytology, particularly when suspicious ultrasound features are present.
