Thyroid Nodule Characterization: Which Thyroid Imaging Reporting and Data System (TIRADS) Is More Accurate? A...

Area of Science:

• Radiology
• Oncology
• Medical Imaging

Background:

• Thyroid nodules require accurate risk stratification for appropriate management.
• Standardized classification systems like TIRADS aim to improve diagnostic consistency.
• Computer-aided diagnosis (CAD) tools are emerging to assist in nodule characterization.

Purpose of the Study:

• To compare the diagnostic performance of K-TIRADS, EU-TIRADS, and ACR TIRADS across different observer experience levels.
• To evaluate the diagnostic performance of CAD software (S-Detect) against TIRADS systems.
• To assess these systems against the gold standard of cytology.

Main Methods:

• 323 thyroid nodules were evaluated by three observers with varying experience.
• Nodules were characterized using ACR TIRADS, EU-TIRADS, and K-TIRADS.
• Computer-aided design software (S-Detect) was applied to the nodules.
• Results were compared with cytology examination.

Main Results:

• K-TIRADS demonstrated the highest specificity (89%) and positive predictive value (50%) among the TIRADS systems.
• ACR-TIRADS (86% specificity, 43% PPV) and EU-TIRADS (79% specificity, 33% PPV) also showed high sensitivity (100%).
• S-Detect combined with EU-TIRADS showed high agreement (>95%) with cytology, outperforming standalone TIRADS in positive predictive power.

Conclusions:

• K-TIRADS offers slightly superior positive predictive power for correctly diagnosing malignant cases.
• S-Detect, particularly when combined with EU-TIRADS, enhances diagnostic accuracy, especially for positive cases.
• CAD software like S-Detect serves as a valuable support tool for radiologists, aiding less experienced observers and improving reporting efficiency.