The Influence of Knowledge of Mammography Findings on the Accuracy of Breast Ultrasound in Symptomatic Women

Abstract

Breast ultrasound is generally interpreted with knowledge of the mammographic examination. This study examined the influence of knowledge of mammography findings on the accuracy of ultrasound in women with breast symptoms. Subjects were sampled from all women 25-55 years of age consecutively attending a breast clinic. This included all 240 women shown to have breast cancer and 240 age-matched women shown not to have cancer. Ultrasound films were prospectively reviewed and reported by two radiologists independent of each other and in a blinded manner. A two-phase design was used. In the first phase, the radiologists provided an opinion on the ultrasound films. In the second phase, the ultrasound films were reread with consideration of the corresponding mammographic examination. The accuracy of reading the ultrasound with and without knowledge of the findings on mammography was compared using sensitivity and specificity, and receiver operating characteristics (ROC) curves. Reporting the ultrasound with knowledge of mammography (compared to without mammography) improved sensitivity and reduced specificity for both radiologists. For one reader, sensitivity increased from 77.5% to 86.7% (p = 0.0002) and specificity decreased from 89.7% to 85.4% (p = 0.04). For the other reader, sensitivity increased from 81.3% to 87.5% (p = 0.0023) and specificity decreased from 87.1% to 85.0% (p = 0.27). ROC curves for both radiologists showed that reporting ultrasound with knowledge of mammography resulted in small (about 3%), but significant improvement in the area under the ROC curve. Our study indicates that knowledge of the findings of mammography improves the interpretation of breast ultrasound in symptomatic women.

Measuring the accuracy of diagnostic imaging in symptomatic breast patients: team and individual performance

“…In a review of 22 studies Flobbe et al [7] found that ultrasound performed better than mammography in 4 of 10 studies in which prior knowledge of mammography results were available and in 8 of 11 studies in which ultrasound was interpreted independently. Their results indicate a higher relative diagnostic performance when no prior knowledge of mammography was available; however, this has not been corroborated by other investigators [8]. As well as calculating individual test performance we have shown that the addition of mammography increased sensitivity by 18.3% over clinical examination and that the addition of ultrasound further raises sensitivity by 9.8%.…”

Does Providing Previous Results Change the Accuracy of Cervical Cytology?

“…In studies of mammography and ultrasound for the diagnosis of breast cancer, providing clinical history improved accuracy. 2,3 However, the effect was small, with an increase in the area under the ROC curve of only 2% or 3%. In a systematic review, 9 studies assessed the effect on ROC curves of using actual history.…”

“…2 Similarly, breast ultrasonography was found to be slightly more accurate when reported with access to mammograms. 3 The accuracy of screening mammography was found to be unaltered when radiologists were supplied with clinical history. 4 However, clinical history had adverse effects on recommendations for further tests to identify cancer.…”

New methods give better estimates of changes in diagnostic accuracy when prior information is provided

“…All of the 240 women who had cancer, and 233 of the women who did not have cancer, underwent both tests. The study was conducted as a component of the Sydney Breast Imaging Accuracy Study, and full details of the subjects' symptoms and histological classification have been published elsewhere [6,7].…”