Interpretation varies among radiologists

For patients who have signs or symptoms that could signal breast cancer, diagnostic mammograms (which involve more x-ray views than are used for screening) are critical to helping radiologists make accurate diagnoses. However, researchers discovered recently that diagnostic accuracy depends strongly on which radiologist reads the mammograms.

As reported in the Dec. 19 issue of the Journal of the National Cancer Institute, Diana L. Miglioretti from Group Health Center for Health Studies in Seattle and her colleagues evaluated whether radiologists’ experience and practice characteristics were associated with the interpretive skill in diagnostic mammography.

The researchers examined the performance of 123 radiologists from 72 facilities across the US, who collectively interpreted 35,895 diagnostic mammograms between 1996 and 2003. The radiologists completed a survey that included questions about their demographic characteristics, their mammography experience and their clinical practice the prior year. The researchers compared the survey responses with the radiologists’ interpretive ability.

Sensitivity — the percentage of positive findings in women who eventually were diagnosed with breast cancer — ranged from 27 to 100 percent among the radiologists, with a median of 79 percent, whereas the false-positive rate ranged from 0 to 16 percent, with a median of 4.3 percent.

The researchers found that radiologists working at academic medical centers had higher sensitivity (88 percent) and more false-positives (7.8 percent) compared with other radiologists. Moreover, those who spent 20 percent or more of their time with breast imaging also had higher sensitivity (80 percent) and higher false-positive rates (4.6 percent) compared with those who did not.

The researchers believe that their results indicate that diagnostic mammography readings vary significantly among radiologists and that future research should focus on identifying ways to improve accuracy and to reduce variability.