Imaging Characteristics of Thyroid Nodules on Ultrasound by AI-TIRADS 2019 Classification, Compare with Histopathological at E Hospital

Abstract

To describe the ultrasound imaging characteristics of thyroid nodules according to the AI-TIRADS 2019 classification compared with pathological anatomy. Research methods: A retrospective cross-sectional study of medical record data of patients with nodules undergoing thyroid ultrasound from October 2024 to March 2025 at E Hospital. Results: The study included 45 patients, with 68 thyroid nodules evaluated by fine-needle aspiration and postoperative histopathology at E Hospital, indicated that: The average age: 52.1 ± 13.5, most commonly from 40-60 years: 48.9%. Mainly female, female-to-male ratio: 6.5/1. Benign nodules were characterized by: wider than tall (95.24%), mixed cystic/solid (52.38%), smooth margins (85.72%) and absence of echogenic foci (71.43%). Malignant nodules were characterized by: taller than wide (53.19%), solid composition (85.11%), hypoechoic (91.49%), lobulated or irregular (51.06%), punctate echogenic foci (25.53%). The diagnostic value of AI-TIRADS 4 and 5 for malignant nodules showed: sensitivity 100%, specificity 71.43%, positive predictive value 88.68%, negative predictive value 100%, and accuracy 91.18%. The diagnostic value of fine-needle aspiration for malignant nodules showed: sensitivity 93.62%, specificity 57.14%, positive predictive value 83.02%, negative predictive value 80%, and accuracy 82.35%.