Interobserver variability in thyroid ultrasound.

Area of Science:

• Endocrinology
• Radiology
• Medical Imaging

Background:

• Ultrasound is the primary imaging modality for thyroid nodules.
• Operator interpretation leads to significant inter-observer variability.
• Standardized classification systems aim to reduce variability.

Purpose of the Study:

• To assess inter-physician agreement on thyroid nodule characteristics.
• To evaluate consensus on risk stratification using multiple classification systems.
• To determine agreement on the need for fine needle aspiration (FNA).

Main Methods:

• Four endocrinologists evaluated 100 thyroid nodule ultrasound images.
• Assessment included nodule composition, echogenicity, margins, and calcifications.
• Nodules were classified using ATA, EU-TIRADS, K-TIRADS, and ACR-TIRADS systems.
• Krippendorff's alpha was used to measure interobserver agreement.

Main Results:

• Interobserver agreement varied for ultrasound features, with composition showing the highest (0.80) and microcalcifications the lowest (0.50).
• Concordance for classification systems ranged from 0.63 (EU-TIRADS) to 0.70 (ATA).
• Agreement on the indication for fine needle aspiration (FNA) was generally acceptable, ranging from 0.73 to 0.86.

Conclusions:

• Interobserver agreement is acceptable for identifying nodules requiring cytologic study (FNA).
• Limited concordance exists in the risk stratification of thyroid nodules.
• Variability persists in assessing specific ultrasonographic characteristics, impacting precise risk assessment.