Diagnostic performance evaluation of different TI-RADS using ultrasound computer-aided diagnosis of thyroid nodules: An experience with adjusted settings

Background Thyroid cancer diagnosis has evolved to include computer-aided (CAD) approaches overcome the limitations of human ultrasound feature assessment. This study aimed evaluate diagnostic performance a CAD system in thyroid nodule differentiation using varied settings. Methods Ultrasound images 205 nodules from 198 patients were analysed this retrospective study. AmCAD-UT software was used at default settings and 3 adjusted diagnose nodules. Six risk-stratification systems classify nodules: The American Association (ATA), College Radiology Imaging, Reporting, Data System (ACR-TIRADS), British (BTA), European Union (EU-TIRADS), Kwak (2011) Korean Society (KSThR). determined relative histopathology and/or cytology each nodule. Results At setting, EU-TIRADS yielded highest sensitivity, 82.6% lowest specificity, 42.1% while ATA-TIRADS 66.4%. had AUROC (0.74) which comparable that ACR, ATA, KSThR TIRADS (0.72, 0.73, 0.70 respectively). hyperechoic foci setting 3.5 with other median values; ATA best-balanced specificity good (70.4%; 67.3% 0.71 Conclusion achieved best all for maximizing sensitivity EU-TIRADS. Adjusting by only reducing echogenic may be most helpful improving minimal change sensitivity.