Background: Indeterminate thyroid nodules (Bethesda III) are challenging to characterise without diagnostic surgery. Auxiliary strategies including molecular analysis, machine learning models and ultrasound grading with TI-RADS can help to triage accordingly, but further refinement is needed to prevent unnecessary surgeries and increase positive predictive values.
Design: Retrospective review of 88 patients with Bethesda III nodules who had diagnostic surgery with final pathological diagnosis.
Measurements: Each nodule was retrospectively scored through TI-RADS. Two deep learning models were tested, one previously developed and trained on another dataset, mainly containing determinate cases and then validated on our dataset while the other one trained and tested on our dataset (indeterminate cases).
Results: The mean TI-RADS score was 3 for benign and 4 for malignant nodules (p=0.0022). Radiological high risk (TI-RADS 4, 5) and low risk (TI-RADS 2,3) categories were established. The PPV for the high radiological risk category in those with >10mm nodules was 85% (CI 70-93%). The NPV for low radiological risk in patients >60years (mean age was 100% (CI 83-100%). The AUC value of our novel classifier was 0.71 and differed significantly from the chance-level (p=0.0009).
Conclusions: Novel radiomic and radiologic strategies can be employed to assist with pre-operative diagnosis of indeterminate thyroid nodules.