Related Experiment Video
Updated: May 21, 2026
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
Nomogram for selecting thyroid nodules for ultrasound-guided fine-needle aspiration biopsy based on a quantification
Iain J Nixon1, Ian Ganly, Lucy E Hann
1Department of Head and Neck Surgery, Memorial Sloan-Kettering Cancer Center, New York, New York, USA.
Background:
Our aim through this study was to develop a statistical tool to quantify risk of malignancy in thyroid nodules based on clinical, biochemical, and ultrasound features, which could be used to select which nodules require ultrasound-guided fine-needle aspiration.
Methods:
Clinical records, biochemical profiles, pathology reports, and ultrasound images were reviewed. Multivariate logistic regression was used to rank variables in their ability to predict malignancy.
Results:
In all, 190 nodules were reviewed. The final diagnoses were papillary carcinoma in 105 patients (66%), other carcinoma in 8 patients (5%), and benign thyroid pathology in 45 patients (29%). After exclusions, 182 nodules remained for analysis on a per nodule basis. The 8 variables with highest predictive value were: age; thyroid-stimulating hormone; and ultrasound size, shape, echo texture, calcification, margin, and vascularity. The nomogram had a concordance index of 75%.
Conclusion:
We produced a nomogram able to accurately predict the need to perform ultrasound-guided fine-needle aspiration on a thyroid nodule based on biochemical, clinical, and ultrasound features.
