Related Experiment Video
Updated: Jul 9, 2025
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
A comprehensive comparative assessment of eight risk stratification systems for thyroid nodules in the elderly
Xiao Ma1, Jing Yu1, Yuanjing Huang1
1Department of Ultrasound, The First Affiliated Hospital of Zhengzhou University, Zhengzhou, China.
Objective:
This study aims to investigate the diagnostic value of eight risk stratification systems (RSSs) for thyroid nodules in the elderly and explore the reasons in comparison with a younger group.
Methods:
Cases of thyroid nodules that underwent ultrasound examination with thyroidectomy or fine-needle aspiration (FNA) at our hospital between August 2013 and March 2023 were collected. The patients were categorized into two groups: an elderly group (aged ≥60) and a younger group (aged <60). Eight RSSs were applied to evaluate these nodules respectively.
Results:
The malignant rate in the elderly group was significantly lower than that in the younger group (28.2% vs. 49.6%, P=0.000). There were statistically significant differences in nodule diameter, multiplicity, composition, echogenicity, orientation, margin, and echogenic foci between the elderly and younger groups (P<0.05). Among the eight RSSs evaluated in elderly adults, the artificial intelligence-based Thyroid Imaging Reporting and Data System (AI TIRADS) demonstrated the highest overall diagnostic efficacy, but with relatively high unnecessary FNA rate (UFR) and missed cancer rate (MCR) of 55.0% and 51.3%, respectively. By modifying the size thresholds, the new AI TI-RADS achieved the lowest UFR and MCR while maintaining nearly the lowest FNA rate (FNAR) among all the RSSs (P=0.172, 0.162, compared to the ACR and original AI, respectively, but P<0.05 compared to the other six RSSs).
Conclusion:
Among the eight RSS systems, AI demonstrated higher diagnostic efficacy in the elderly population. However, the size thresholds for FNA needed to be adjusted.
