THE ROLE OF ULTRASOUND IN EARLY DETECTION OF THYROID PATHOLOGY: MODERN CRITERIA AND CLASSIFICATIONS (TIRADS 2024)

Abstract

Ultrasound (US) examination for the early detection of thyroid pathology is widely used today as the main, primarily - risk stratifying and screening tool for further diagnosis of patients - mainly cytology (CFT) or/and surgical management. This article analyzes the influence of modern criteria and classification systems - ACR-TI-RADS, EU-TIRADS, K-TIRADS, ATA recommendations, and TIRADS models based on artificial intelligence that have emerged in recent years on clinical practice. The compliance of the ultrasound lexicon and classification with the international consensus, the relationship of ultrasound features (eccogenicity, shape, marginal, microcalcifications, central and peripheral circulation) with oncological risk, and the differences in biopsy size limits used in different TIRADS  systems  are  considered.  It  was  noted  that  the  ACR-TI-RADS  and  European recommendations provided a clear and repeatable approach to patient selection, but the presence of differences in the measurement and monitoring schedule for biopsy between different systems is important  in  clinical  decision-making.  Also  discussed  are  recent  studies  on  the  possibility  of integrating US data using MRI and artificial intelligence (for example, combining MRI morphological features with ACR-TI-RADS or AI-TIRADS) increasing diagnostic effectiveness and reducing the number of unnecessary biopsies. In this work, the advantages and limitations of modern TIRADS criteria based on US, their comparison with histopathological results, and practical recommendations for their application to clinical protocols are presented analytically. The results show that early detection of thyroid diseases and efficient resource allocation can be achieved by legalizing and standardizing TI-RADS approaches, but adaptation to the country's practice requires modification depending on local epidemiology, existing diagnostic capabilities, and the patient's risk profile. In the future,  the  prospects  for  further  improvement  of  risk  stratification  through  the  integration  of international lexicon and artificial intelligence models are considered promising.