Related Experiment Video
Updated: Mar 3, 2026
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
Cytology-Ultrasonography Risk-Stratification Scoring System Based on Fine-Needle Aspiration Cytology and the
Min Ji Hong1, Dong Gyu Na2, Jung Hwan Baek3
11 Department of Radiology, Gachon University Gil Medical Center , Incheon, Korea.
Background:
The malignancy risk of a cytology diagnosis may depend on the ultrasonography (US) patterns of thyroid nodules, and management should be determined by the combined malignancy risk of fine-needle aspiration (FNA) cytology and US patterns. This study was performed to develop a clinically applicable cytology-ultrasonography (CU) scoring system for malignancy risk stratification based on FNA cytology and US patterns, according to the Korean-Thyroid Imaging Reporting and Data System (K-TIRADS).
Methods:
This retrospective Institutional Review Board-approved study included 1651 thyroid nodules (≥1 cm) with final diagnoses. The malignancy risk was assessed of the combined results of FNA cytology and the K-TIRADS for the development of the CU system. The interaction between FNA cytology and US pattern (K-TIRADS) in the malignancy risk of nodules was investigated by using a binominal test.
Results:
The malignancy risk of nodules could be stratified into four CU scores (very low risk, <3%; low risk, ≥3%, <30%; high risk, ≥30%, <90%; very high risk, ≥90%). In nodules with non-diagnostic, benign, and atypia of undetermined significance/follicular lesion of undetermined significance cytology results, low-suspicion US pattern (K-TIRADS 3) significantly decreased the malignancy risk of nodules (p = 0.003, 0.013, and 0.027, respectively), and a high-suspicion US pattern (K-TIRADS 5) significantly increased the malignancy risk of nodules (p ≤ 0.001). A Bethesda 1 or 4 cytology result did not significantly change the malignancy risk of any K-TIRADS (p ≥ 0.518 and p ≥ 0.137, respectively). A Bethesda 2 cytology result decreased and a Bethesda 5 or 6 cytology result increased the malignancy risk of K-TIRADS 3, 4, and 5 (p ≤ 0.001). A Bethesda 3 cytology result increased the malignancy risk of K-TIRADS 3 and 4 (p < 0.001 and p = 0.024, respectively).
Conclusion:
The malignancy risk of thyroid nodules can be stratified by the CU risk-stratification system, based on FNA cytology and the K-TIRADS. The proposed CU scoring system may be helpful in the management of thyroid nodules after FNA.
