Related Experiment Videos
Variability in radiologists' interpretations of mammograms
J G Elmore1, C K Wells, C H Lee
1Department of Internal Medicine, Yale University, New Haven, Conn.
The New England Journal of Medicine
|December 1, 1994
Summary
Radiologist interpretation variability impacts mammography effectiveness. Studies show moderate agreement, with significant differences in management recommendations, highlighting a need for improved accuracy in breast cancer screening.
Area of Science:
- Radiology
- Oncology
- Medical Imaging
Background:
- Mammography is valuable for breast cancer screening.
- Radiologists' interpretations significantly influence mammography efficacy.
- Variability in interpretation is not well understood.
Purpose of the Study:
- To assess the extent of inter-reader variability in mammogram interpretation.
- To analyze disagreement in diagnostic recommendations among radiologists.
Main Methods:
- 150 mammograms (27 with cancer, 123 without) were interpreted by 10 radiologists.
- Stratified random sampling was used for mammogram selection.
- Disagreement was analyzed pairwise and for the group.
Main Results:
- Moderate diagnostic consistency observed (median weighted kappa 0.47).
- Significant disagreement in management recommendations occurred in 25% of group comparisons.
- Disagreement on lesion location (breast laterality) occurred in 9% of group comparisons.
Conclusions:
- Radiologists exhibit substantial variability in mammogram interpretation and management recommendations.
- Reducing interpretation variability may enhance mammography's effectiveness in early breast cancer detection.
- Further efforts are needed to improve accuracy and consistency in mammographic interpretation.