Related Experiment Video
Updated: Jun 25, 2026
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
An ultrasonogram reporting system for thyroid nodules stratifying cancer risk for clinical management
Eleonora Horvath1, Sergio Majlis, Ricardo Rossi
1Thyroid Board, Clinica Alemana de Santiago, Av. Vitacura 5951 Santiago, Chile.
Context:
There is a high prevalence of thyroid nodules on ultrasonographic (US) examination. However, most of them are benign. US criteria may help to decide cost-effective management.
Objective:
Our objective was to develop a standardized US characterization and reporting data system of thyroid lesions for clinical management: the Thyroid Imaging Reporting and Data System (TIRADS).
Design:
This was a prospective study using the TIRADS, which is based on the concepts of the Breast Imaging Reporting Data System of the American College of Radiology.
Materials:
A correlation of the US findings and fine needle aspiration biopsy (FNAB) results in 1959 lesions biopsied under US guidance and studied histologically during an 8-yr period was divided into three stages. In the first stage, 10 US patterns were defined. In the second stage, four TIRADS groups were defined according to risk. The percentages of malignancy defined in the Breast Imaging Reporting and Data System were followed: TIRADS 2 (0% malignancy), TIRADS 3 (<5% malignancy), TIRADS 4 (5-80% malignancy), and TIRADS 5 (>80% malignancy).
Results:
The TIRADS classification was evaluated at the third stage of the study in a sample of 1097 nodules (benign: 703; follicular lesions: 238; and carcinoma: 156). Sensitivity, specificity, positive predictive value, negative predictive value, and accuracy were 88, 49, 49, 88, and 94%, respectively. The ratio of benign to malignant or follicular FNAB results currently is 1.8.
Conclusions:
The TIRADS has allowed us to improve patient management and cost-effectiveness, avoiding unnecessary FNAB. In addition, we have established standard codes to be used both for radiologists and endocrinologists.
