Mammography Technology Doesn’t Replace Doctors’ Judgment
According to a study published in the American Journal ofRoentgenology, use of computed-assisted detection (CAD) technology for the interpretation of mammograms cannot replace the judgment of a radiologist.
Computer-aided detection (CAD) technology involves the use of a computer to evaluate a digital mammographic image. The technology marks areas of the image that appear abnormal and that should be reviewed further by the radiologist. CAD doesn’t replace a radiologist’s review of a mammogram, but it can supplement it.
Whether CAD improves the accuracy of mammographic interpretations is still uncertain. Two measures of the accuracy of a screening test are sensitivity and specificity. Sensitivity refers to the ability of a test to detect cancer in individuals with cancer. Specificity refers to the ability of a test to correctly classify cancer-free individuals. Both of these measures are important. If sensitivity is low, many cancers will be missed by the test. If specificity is low, many cancer-free individuals will undergo unnecessary diagnostic tests such as biopsies.
To explore how use of CAD affects the sensitivity and specificity of mammographic screening, researchers asked 19 radiologists to review mammograms with and without the use of CAD. Some of the mammograms were from women who were cancer-free and some were from women with breast cancer.
- Overall, use of CAD did not increase the sensitivity of mammography. A similar number of breast cancers were detected with and without CAD.
- The effect of CAD on cancer detection varied by whether or not a visible mammographic abnormality was marked by CAD. When CAD marked a visible abnormality as abnormal, use of CAD increased cancer detection compared to non-use of CAD. When CAD failed to mark a visible abnormality as abnormal, however, use of CAD decreased cancer detection compared to non-use of CAD. This suggests that radiologists sometimes defer to the technology, even though radiologists sometimes see abnormalities that are missed by CAD.
- Use of CAD increased the specificity of mammography. Of the women without breast cancer, 75% were correctly classified as cancer-free when CAD was used, and 72% were correctly classified as cancer-free when CAD was not used.
The researchers conclude that CAD increased the correct classification of cancer-free women but did not increase the detection of cancers. The study indicates that radiologists cannot rely entirely on CAD results, but must still use their own judgment when interpreting mammograms.
Reference: Taplin SH, Rutter CM, Lehman CD. Testing the Effect of Computer-assisted Detection on Interpretive Performance in Screening Mammography. American Journal of Roentgenology. 2006;187:1475-1482.
Copyright © 2018 CancerConnect. All Rights Reserved.