Mammogram interpretations vary between community radiologists, experts
The level of agreement between community-based radiologists and an expert panel of radiologists when it comes to interpreting mammograms is high for cancer cases, but less so for noncancers such as asymmetric lesions, calcifications, architectural distortions, and asymmetric densities, according to a study in the November issue of the Journal of the American College of Radiology.
Since the effectiveness of screening mammography depends on accurate image interpretations, and because the performance of radiologists in community practice for screening mammography has been shown to vary widely, Tracy Onega, Ph.D., of Dartmouth Medical School and her colleagues wanted to assess the agreement between mammographic interpretations by community radiologists and a gold standard--in this case the consensus interpretations of a panel of experts.
More than 100 community-based radiologists were recruited to assess one of four randomly assigned test sets of 109 mammograms with comparison studies for recall or no recall, and for giving the most significant findings and locations.
The researchers found that recall agreement was 100 percent for masses and examinations with obvious findings in cancer and noncancer cases. For cancer cases recall agreement was lower for subtle (50 percent) or asymmetric (60 percent) lesions, and for subtle noncancer findings and benign calcifications (33 percent). Recall agreement also was low for architectural distortions (43 percent) and asymmetric densities (40 percent), while asymmetric densities accounted for the most unnecessary recalls.
"By understanding the extent to which case difficulty and finding type affects interpretive agreement, we can develop targeted training modules and educational interventions that yield the greatest improvement in radiologist interpretive performance," the authors wrote. "Our analysis of mammographic interpretation agreement between radiologists and an expert panel suggests that mammographic training should focus on identification and correct interpretation of asymmetric densities and architectural distortions."