Inter-observer variability in the interpretation of mammograms.

Area of Science:

• Radiology
• Medical Imaging
• Oncology

Background:

• Mammography is a key tool for breast cancer screening.
• Inter-observer agreement in mammogram interpretation is crucial for consistent diagnostic accuracy.
• Variability in interpretation can impact patient management and outcomes.

Purpose of the Study:

• To evaluate the inter-observer agreement among radiologists in interpreting mammograms.
• To assess the reliability of mammographic interpretation for parenchymal patterns, diagnoses, and further testing recommendations.

Main Methods:

• Eight radiologists with varying experience levels interpreted 180 mammograms each from 45 women.
• The study included women with confirmed breast cancer, benign conditions, and normal findings.
• Weighted kappa statistics were used to quantify agreement for parenchymal patterns, five diagnostic categories, and recommendations for further tests.

Main Results:

• Weighted kappa values for parenchymal patterns ranged from 0.27 to 0.82 (median 0.60).
• Agreement for diagnosis in five categories ranged from 0.33 to 0.67 (median 0.48).
• Agreement for indications for further diagnostic tests ranged from 0.22 to 0.57 (median 0.38).

Conclusions:

• Inter-observer agreement in mammogram interpretation is comparable to previous studies.
• The findings suggest moderate to substantial agreement for key aspects of mammographic interpretation.
• Ensuring consistent interpretation is vital for effective breast cancer screening programs.