Interobserver variation in reporting CT arthrograms of the shoulder

Area of Science:

• Radiology
• Orthopedic Imaging
• Musculoskeletal Imaging

Background:

• Computed tomography (CT) arthrography is a highly accurate imaging technique for diagnosing shoulder instability and labral lesions.
• Accurate interpretation is hindered by common anatomical variants and the complexity of bone and soft tissue injuries.
• Standardized reporting protocols are needed to improve diagnostic consistency.

Purpose of the Study:

• To assess the interobserver agreement among radiologists in interpreting shoulder CT arthrograms.
• To identify specific shoulder pathologies with significant interobserver variation.
• To suggest improvements for more reliable and clinically accurate CT arthrogram reporting.

Main Methods:

• A consecutive sample of 50 shoulder CT arthrograms was reviewed by two consultant musculoskeletal radiologists.
• Interobserver agreement was quantified using the Kappa (K) statistic for various shoulder pathologies.
• Pathologies assessed included Hill-Sachs lesions, soft tissue and bony Bankart lesions, anterior capsular laxity, and glenohumeral osteoarthritis.

Main Results:

• Interobserver agreement varied significantly across pathologies, ranging from slight (0.20 for osteoarthritis) to substantial (0.61 for bony Bankart lesions).
• Fair to moderate agreement was observed for Hill-Sachs lesions (K=0.37), soft tissue Bankart lesions (K=0.32), and anterior capsular laxity (K=0.41).
• Complete agreement was achieved in only 18% of the 50 scans, indicating considerable interobserver variation.

Conclusions:

• Significant interobserver variation exists in the reporting of shoulder CT arthrograms.
• This variability poses challenges for accurate diagnosis and clinical correlation.
• Focused reporting strategies are necessary to enhance the reliability and accuracy of shoulder CT arthrogram interpretations.