Comparison of ACR-TIRADS to the ATA Guidelines for Thyroid Nodules: A Neck to Neck Comparison

Introduction: The goal of this study was to compare the performance characteristics of the American College of Radiology Thyroid Imaging Reporting and Data System (ACR-TIRADS) and the American Thyroid Association (ATA) systems in identifying malignant thyroid nodules. Methods: In a retrospective chart review, ultrasound images of all thyroid nodules biopsied in 2014- 2015 at a Canadian academic centre were reviewed by two radiologists. The ultrasound characteristics of thyroid nodules were compared with cytologic or pathologic results to determine the positive predictive value (PPV), negative predictive value (NPV), sensitivity and specificity for TIRADS and ATA in predicting cancer risk. Clinical course of nodules not requiring follow up or intervention according to ACR-TIRADS was described. Vascularity was added to ACR-TIRADS to determine whether sensitivity of TIRADS improves. Results: A total of 417 thyroid nodules were reviewed, 82% were benign (Bethesda II). The sensitivity, specificity, PPV, and NPV were 97%, 11%, 9%, 98%, and 70%, 29%, 18%, and 81% for ATA and TIRADS, respectively. Of the 10 nodules that did not need ultrasound follow up based on TIRADS criteria, 2 were malignant, the rest were FLUS. If vascularity was added to TIRADS (TIRADS-Vasc), the number of malignant cases missed could have been reduced by 43% (from 7 to 4 cases). Conclusions: TIRADS is more specific but less sensitive than ATA, and misses a small number of malignant nodules. Clinicians need to use their judgement to decide which nodules require biopsy as some malignant cases will be missed using TIRADS alone.