Correlation between ultrasound TI-RADS classification system and Bethesda cytology result in assessment of thyroid nodule among a sample of Iraqi patient in wasit province

Background: Thyroid nodule is a common presentation. The estimated prevalence of thyroid nodules is 5-10% by clinical examination and up to 70% on ultrasonographic (USG) evaluation. Most are benign without any symptoms or cosmetic concerns. Only around 7-15% are found to be malignant.

Objective: This study was conducted to assess the association between Thyroid Imaging, Reporting, Data System (TI-RADS) and the corresponding category of Bethesda System for Reporting Thyroid Cytology (TBSRTC) in evaluation of thyroid nodules and assessing risk of malignancy.

Methods: This retrospective study evaluated a total of 172 patients with thyroid nodules presenting to out-patient laboratory in AL-Kut in a period of three years, by categorizing them into different categories using Thyroid Imaging Reporting and Data System (TIRADS) by USG and The Bethesda System for Reporting Thyroid Cytopathology (TBSRTC) by fine needle aspiration (FNA). The correlation between TIRADS and TBSRTC was evaluated using SPSS Statistics.

Result: By FNA, 93% were benign nodules (category II-III), whereas 7.4% were malignant (category IV-V). By ultrasound 85% were benign (TI-RADS 1-3). The risk of malignancy for TI-RADS1, 2, 3, 4, 5 was 0, 0, 2.9, 38, and 80% respectively. No statistically significant difference between males and females. Overall agreement between the cases by USG and FNA using the TIRADS and TBSRTC respectively was significant (p<0.0001), and the area under the ROC curve was 0.913.

Conclusion: Our study observed a substantial agreement between the diagnosis made by TIRADS on ultrasound and TBSRTC on FNA, so using TI-RADS as the initial step in stratification of all thyroid nodules and as the only step in TI-RADS 1 and TIRADS2 nodules, so that only suspicious lesions undergo FNA, can reduce the need for invasive more expensive procedures without risking cancer missing.