Observer variation in the interpretation of lower limb venograms

Area of Science:

• Radiology
• Vascular Imaging
• Diagnostic Accuracy

Background:

• Venous thrombosis diagnosis relies on imaging techniques.
• Venography has been a reference standard, but inter-observer variability can affect its reliability.
• Standardized diagnostic criteria are crucial for consistent interpretation.

Purpose of the Study:

• To assess inter-observer agreement among radiologists in diagnosing lower limb venous thrombosis using venography.
• To evaluate the impact of refined diagnostic criteria on diagnostic consistency.
• To determine the extent of observer variation in venographic interpretation.

Main Methods:

• Two experienced radiologists independently reviewed 40 lower limb venograms twice.
• A pilot study and rediscussion of criteria preceded the main review.
• The kappa statistic was used to analyze agreement, correcting for chance.
• Eleven major veins were assessed for the presence of thrombus (absent, doubtful, presumed, definite).

Main Results:

• Overall agreement levels were higher than in many previous radiologic studies, likely due to refined criteria.
• Despite improvements, observers disagreed on thrombus presence in approximately 10% of examinations.
• The kappa statistic indicated substantial agreement, but not perfect concordance.

Conclusions:

• Refined diagnostic criteria improved inter-observer agreement in lower limb venography for venous thrombosis.
• Observer variation, though reduced, still exists and must be considered.
• Venography's role as a reference standard requires careful consideration of this residual variability.