THE USE OF THE THYROID IMAGING REPORTING AND DATA SYSTEM (TIRADS) IN THE DIAGNOSIS OF THYROID NODULES

Abstract

Thyroid nodules represent one of the most common diagnoses in medicine, having been found to be present in as many as two-thirds of adults upon the application of high-resolution ultrasound scanning. Though the vast majority turn out to be benign, the lack of ability to distinguish the relatively small number of clinically meaningful malignancies from those of low significance in terms of treatment has led to large numbers of patients undergoing fine needle aspiration (FNA) biopsies, diagnostic operations, and an over-diagnoses of thyroid cancer pandemic across the world. This is the challenge which the TIRADS system was designed to overcome. This narrative review aims to explore the history, architecture, efficacy, and application of TIRADS schemes, with special attention given to the most widely used classification scheme, the American College of Radiology TIRADS (ACR TI-RADS) and alternative classifications, namely EU-TIRADS, K-TIRADS, and C-TIRADS. In this review, using the results of recent meta-analyses and validations across multiple institutions, the performance of these systems is evaluated in relation to their sensitivity, specificity, discriminative ability, and potential for avoiding unnecessary biopsies. Common problems such as inter-rater variability, uncertain diagnosis of follicular lesions, and management of indeterminate cytology are reviewed in detail, as well as the increasing use of machine learning in improving the accuracy and efficiency of feature extraction and risk stratification. There is significant evidence that TIRADS, and the ACR TI-RADS system in particular, has greatly increased the specificity of nodules assessment and avoided low-yield interventions, and that convergence between competing classification systems and long-term outcome assessments are essential needs.