Keywords: Diagnosis/Prediction, AI/ML Software, thyroid nodule; quantification; segmentation
Motivation: To explore effectiveness of AI in automated detection and characterization of thyroid nodules using T2-weighted neck MRIs
Goal(s): To assess AI-driven segmentation and characterization accuracy of thyroid nodules on multi-vendor 1.5T T2-weighted neck MRIs
Approach: nnUNet was applied to a dataset of 278 annotated cases for 3D-thyroid nodule segmentation, followed by post-processing to measure volume, size, laterality, and location. Model performance was evaluated on 750 test-cases (500 without nodules, 250 with nodules).
Results: The model achieved an 86.34 Dice for segmentation, with sensitivity, and specificity of 83.2%, and 86.4%, respectively. Accuracy was higher for nodules over 1cm (87.37%) compared to smaller ones (84%)
Impact: Having the ability to not only automatically detect thyroid nodules but automatically to characterize them provides valuable insights as well as saving valuable time to radiologists in dealing with this condition.
How to access this content:
For one year after publication, abstracts and videos are only open to registrants of this annual meeting. Registrants should use their existing login information. Non-registrant access can be purchased via the ISMRM E-Library.
After one year, current ISMRM & ISMRT members get free access to both the abstracts and videos. Non-members and non-registrants must purchase access via the ISMRM E-Library.
After two years, the meeting proceedings (abstracts) are opened to the public and require no login information. Videos remain behind password for access by members, registrants and E-Library customers.
Keywords