Related Experiment Video
Updated: Dec 22, 2025

A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
[Accuracy of two thyroid imaging, reporting and data systems for differential diagnosis of benign and malignant
Shuiqing Lai1, Yuancheng Chen1, Zhijiang Chen1
1Department of Endocrinology, Guangdong Provincial People's Hospital, Guangdong Academy of Medical Sciences, Guangzhou 510080, China.
Objective:
To compare the accuracy of two widely used thyroid imaging, reporting and data systems (TI-RADS), namely ACR TI-RADS and Kwak TI-RADS, in the differential diagnosis of benign and malignant thyroid nodules.
Methods:
We reviewed the data of 350 thyroid nodules with definite diagnoses by surgical histopathology (n=144, 41.14%) or fine needle aspiration (FNA) cytopathology (n=206, 58.86%). The nodules were graded using ACR TI-RADS and Kwak TI-RADS based on the ultrasound images, and the diagnostic accuracy of these two systems was evaluated by the area under the receiveroperating characteristic curve (AUC).
Results:
The AUCs of ACR TI-RADS and Kwak TI-RADS were both 0.879. For a differential diagnosis of the thyroid nodules, ACR TI-RADS had a diagnostic sensitivity, specificity, positive predictive value, negative predictive value, positive likelihood ratio, negative likelihood ratio, diagnostic odds ratio, Youden's index and accuracy of 77.3%, 89.1%, 83.0%, 85.1%, 7.101, 0.255, 27.848, 0.664 and 0.843, respectively, with an optimal threshold of TR5, as compared with 84.8%, 84.0%, 78.3%, 89.0%, 5.283, 0.181, 29.265, 0.688 and 0.843, respectively, of Kwak TI-RADS, which had an optimal threshold of 4c.
Conclusions:
Both ACR TI-RADS and Kwak TI-RADS have good performance for differential diagnosis of thyroid nodules, but ACR TI-RADS has a higher specificity and a lower sensitivity compared with Kwak TI-RADS.
Related Concept Videos
Radiological Investigation I: X-ray and CT
Imaging Studies II: Positron Emission Tomography and Scintigraphy
Fundamental Principles of PET

