Risk stratification of indeterminate thyroid nodules using ultrasound and machine learning algorithms
Clinical Endocrinology Oct 19, 2021
Gild ML, Chan M, Gajera J, et al. - Findings support the use of novel radiomic and radiologic strategies to assist with preoperative diagnosis of indeterminate thyroid nodules.
Indeterminate thyroid nodules (Bethesda III) are difficult to characterize without diagnostic surgery.
This retrospective review included 88 patients with Bethesda III nodules who underwent diagnostic surgery with final pathological diagnosis.
Via TI-RADS (Thyroid Imaging, Reporting and Data System), retrospective scoring of each nodule was performed, and two deep learning models were tested.
For benign and for malignant nodules, the mean TI-RADS score was 3 and 4, respectively, and radiological high risk (TI-RADS 4,5) and low risk (TI-RADS 2,3) categories were defined.
PPV for the high radiological risk class in those with >10 mm nodules was 85%, and NPV for low radiological risk in those >60 years (mean age) was 100%.
Area under the curve value of the novel classifier was identified to be 0.75 and differed significantly from the chance-level.
-
Exclusive Write-ups & Webinars by KOLs
-
Daily Quiz by specialty
-
Paid Market Research Surveys
-
Case discussions, News & Journals' summaries