EVALUATION OF PREDICTIVE CAPABILITIES OF TIRADS, EU-TIRADS, TLA_RU SYSTEMS IN THE ULTRASOUND DIAGNOSIS OF THYROID NODULAR PATHOLOGY

Abstract

The aim of the study was to evaluate the prognostic capabilities of EU-TIRADS, TIRADS, TLA_RU systems in ultrasound diagnostics of thyroid nodular masses. A retrospective independent continuous blind study of ultrasound examination protocols was conducted in 665 patients, 241 patients had benign non-cancerous thyroid diseases, 86 patients had benign tumors (adenomas), 338 patients had thyroid cancer. During the preoperative examination, all patients underwent a multiparametric ultrasound examination of the neck organs according to the standard procedure, with the registration of identified thyroid nodes, with examining the prognostic capabilities of stratification systems – TIRADS, EU-TIRADS, TLA_RU. It was found that the TIRADS system has a sensitivity of 91.04%, specificity – 91.41%. Focusing on the scale of AUC values reflecting the quality of the diagnostic test, it can be stated that TIRADS is a test with excellent quality (AUC =0.972). Basing on the results of the analyzing the data obtained by the EU-TIRADS system, it was revealed that it is a high-quality test (AUC=0.826), but its predictive capabilities are worse than those of TIRADS. The original TLA_RU model has 87.5% sensitivity and 95.7% specificity. During the ROC analysis, it was found that the AUC is equal to 0.954±0.00894, which suggests that the TLA_RU model is an excellent quality test in the differential diagnosis of thyroid nodular masses. Multivariate statistical comparative analysis of thyroid imaging assessment systems (TIRADS, EU-TIRADS and TLA_RU) from the standpoint of evidence-based medicine has shown that thyroid cancer risk stratification systems based on the assessment of multiparametric ultrasound signs have great diagnostic capabilities.