Related Experiment Video
Updated: Nov 17, 2025
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
Association of the Implementation of a Standardized Thyroid Ultrasonography Reporting Program With Documentation of
Amr F Hamour1, Weining Yang1, John J W Lee1
1Department of Otolaryngology-Head and Neck Surgery, University of Toronto, Toronto, Ontario, Canada.
Importance:
Although most thyroid nodules are benign, the potential for malignant neoplasms is associated with unnecessary workup in the form of imaging, fine-needle aspiration, and diagnostic surgery. The American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS) is commonly used to assess the malignant neoplasm risk potential of thyroid nodules imaged by ultrasonography. However, standardized reporting of ACR TI-RADS descriptors is inconsistent.
Objective:
To increase the documentation rate of ACR TI-RADS thyroid nodule characteristics to 80% in 18 months.
Design, Setting, And Participants:
This prospective interrupted time series quality improvement study was conducted from December 1, 2018, to March 31, 2020, at a tertiary outpatient head and neck clinic among 229 patients who had at least 1 documented thyroid nodule identified on bedside clinic ultrasonography. Data analysis was performed throughout the entire study period because this was a quality improvement study with iterative small cycle changes; final analysis of the data was performed in April 2020.
Main Outcomes And Measures:
The primary outcome was the documentation rates of 6 ACR TI-RADS ultrasonographic descriptors. Secondary outcomes included nodule fine-needle aspiration biopsy rate and physician-reported clinic flow efficiency.
Results:
A total of 229 patients had at least 1 documented thyroid nodule and were included in the analysis. Size was the most frequently documented nodule characteristic (72 of 74 [97.3%]) at baseline, followed by echogenic foci (31 of 74 [41.9%]), composition (23 of 74 [31.1%]), echogenicity (17 of 74 [23.0%]), margin (6 of 74 [8.1%]), and shape (1 of 74 [1.4%]). After 3 Plan, Do, Study, Act (PDSA) cycles, the final intervention consisted of a standardized ultrasonography reporting form and educational initiatives for surgical trainees. After the third PDSA cycle (n = 36), reporting of nodule size, echogenic foci, and composition increased to 100%. Similarly, reporting of echogenicity (34 of 36 [94.4%]), shape (28 of 36 [77.8%]), and margin (25 of 36 [69.4%]) all increased. This represented a cumulative 90.3% documentation rate (195 of 216), a 56.5% increase from baseline (95% CI, 50.0%-61.9%). The standardized reporting form was used in 83.3% of eligible thyroid ultrasonography cases (30 of 36) after PDSA cycle 3, demonstrating good fidelity of implementation. There were no unintended consequences associated with clinic workflow, as a balancing measure, reported by staff surgeons.
Conclusions And Relevance:
This study suggests that implementation of an ACR TI-RADS-based reporting form in conjunction with educational initiatives improved documentation of ultrasonographic thyroid nodule characteristics, potentially allowing for improved bedside risk stratification and communication among clinicians.
