RELIABILITY OF THE 2017 ACR TI-RADS CLASSIFICATION SYSTEM IN DETECTING MALIGNANT THYROID NODULES

Abstract

INTRODUCTION
In 2017, the American College of Radiology (ACR) proposed a new point-based reporting system, the ACR Thyroid imaging reporting and data system (ACR TIRADS) consisting of 5 categories ranging from TR1 to TR5, stratifying thyroid malignancy risk based on ultrasound features and a predetermined size cut off for fine needle aspiration cytology (FNAC) or follow up. Since the introduction of ACR TIRADS, this system has been validated in other countries. We aim to evaluate the reliability and practicality of the ACR TIRADS scoring system in our centre.

METHODOLOGY
A cross-sectional observational study of 592 adult patients (716 thyroid nodules) from 2019-2021 in Hospital Queen Elizabeth and Hospital Queen Elizabeth II, Kota Kinabalu who had thyroid ultrasound and available FNAC and/or histopathological examination (HPE) results. The ACR TIRADS system was applied to categorize thyroid nodules. The performance and diagnostic accuracy of the ACR TIRADS scoring system, the risk of malignancy in each TR category and the percentage of unnecessary FNAC rates were determined.

RESULTS
The ACR TI-RADS performance showed a specificity, sensitivity, positive predictive value (PPV), negative predictive value (NPV) of 94.6%, 41.8%, 26.4%, 97.2%. The diagnostic accuracy was 51.5%. The ROC curve analysis showed AUC 0.794 (95% CI: 0.753-0.834). The risk of malignancy was 0% for TR 1 and 2, 4.1% for TR3, 17.2 % for TR4, and 52.9% for TR5. The unnecessary FNAC rate was 32.5%.

CONCLUSION
The ACR TI-RADS system applied in our centre is synonymous with other validated studies and is a reliable system to differentiate malignant from benign thyroid nodules. The high unnecessary FNAC rates inform that FNAC should be deferred in patients with TR 1 and 2 nodules.