Inter- and Intra-Observer Variability and the Effect of Experience in Cine-MRI for Adhesion Detection.

Area of Science:

• Medical Imaging
• Radiology
• Surgical Outcomes

Background:

• Post-abdominal surgery pain affects many patients, often due to adhesions.
• Cine-magnetic resonance imaging (MRI) is a novel technique for adhesion detection.
• Limited data exists on cine-MRI's diagnostic accuracy and observer variability.

Purpose of the Study:

• To evaluate inter- and intra-observer variability in cine-MRI adhesion detection.
• To assess the diagnostic accuracy of cine-MRI for adhesions.
• To determine the impact of observer experience on cine-MRI interpretation.

Main Methods:

• Retrospective analysis of 61 sagittal cine-MRI slices.
• 15 observers with varying experience annotated suspected adhesions.
• Inter-observer variability assessed using Fleiss' kappa; intra-observer using Cohen's kappa; diagnostic accuracy via ROC analysis.

Main Results:

• Inter-observer agreement ranged from poor to fair (Fleiss' kappa 0.04–0.34).
• Higher general and cine-MRI experience significantly improved observer agreement (p < 0.001).
• Group AUC scores were 0.66–0.72, with individual scores up to 0.78; intra-observer agreement was moderate (Cohen's kappa 0.37–0.53).

Conclusions:

• Cine-MRI can diagnose adhesions, validated against a radiologist consensus panel.
• Observer experience is crucial for improving cine-MRI interpretation agreement.
• Further research, including guidelines or AI, is needed for consistent interpretation.