Related Experiment Video
Updated: Jun 28, 2026
A Swin Transformer-Based Model for Thyroid Nodule Detection in Ultrasound Images
Published on: April 21, 2023
Can quantitative diffusion-weighted MR imaging differentiate benign and malignant cold thyroid nodules? Initial
C Schueller-Weidekamm1, K Kaserer, G Schueller
1Department of Diagnostic Radiology, Medical University of Vienna, Vienna, Austria. claudia.schueller-weidekamm@meduniwien.ac.at
Background And Purpose:
The characterization of cold nodules of the thyroid gland is mandatory because approximately 20% of these nodules are of malignant origin. The purpose of this study was to evaluate the distinction of cold thyroid nodules by using quantitative diffusion-weighted MR imaging (DWI).
Materials And Methods:
In 25 patients with cold nodules on scintigraphy and suggestive findings at fine-needle aspiration, thyroid carcinoma was suggested. In these patients, cold nodules and the normal parenchyma of the contralateral thyroid lobe were prospectively investigated with quantitative DWI (echo-planar imaging sequence; maximum b-value, 800 s/mm(2)) before surgery. The differences in the mean apparent diffusion coefficient (ADC) values in benign and malignant nodules were tested by using a Mann-Whitney U test.
Results:
Histologically, there were 20 carcinomas with a minimum size of 8 mm and 5 adenomas. The mean ADC values (in 10(-3) mm(2)/s) differed significantly among carcinoma, adenoma, and normal parenchyma (P < .05). The ranges (95% confidence interval) of the ADC values for carcinoma (2.43-3.037), adenoma (1.626-2.233), and normal parenchyma (1.253-1.602) showed no overlap. When an ADC value of 2.25 or higher was used for predicting malignancy, the highest accuracy of 88%, with 85% sensitivity and 100% specificity, was obtained.
Conclusions:
Quantitative DWI seems to be a feasible tool with which to differentiate thyroid carcinomas from adenomas; however, further studies are required including larger numbers of patients to confirm our results.
