Digital Breast Tomosynthesis Superior to Mammography for Surveillance of Breast Cancer Survivors
Key findings
- This retrospective study compared 9,019 digital mammography (DM) examinations performed in 4,085 women with 22,887 digital breast tomosynthesis (DBT) examinations performed in 7,154 women
- DBT led to a lower abnormal interpretation rate (5.8% vs. 6.2%; OR, 0.80; P=0.001) and higher specificity (95.0% vs. 94.7%; OR, 1.23; P=0.003) than DM
- Massachusetts General Hospital's switch to DBT did not affect the cancer detection rate, and there were no differences between DM and DBT in the proportion of invasive versus in situ cancers detected
- Interval cancer rates were also similar with DM and DBT, and one-third of interval cancers were detected with MRI
Subscribe to the latest updates from Radiology Advances in Motion
According to current guidelines, two-dimensional digital mammography (DM) is appropriate for routine surveillance of breast cancer survivors who have not undergone a bilateral mastectomy. However, surgery and radiation can lead to scarring and retractions, which can obscure or mimic cancers on DM images.
Digital breast tomosynthesis (DBT), also called three-dimensional mammography, minimizes the masking effect of overlying tissue. In the largest study on this topic, published in Radiology, Manisha Bahl, MD, MPH, director of the Breast Imaging Fellowship Program in the Department of Radiology at Massachusetts General Hospital, and Constance D. Lehman, MD, PhD, director of the Division of Breast Imaging, and colleagues found that DBT was associated with a lower abnormal interpretation rate and higher specificity, compared with DM, in breast cancer survivors.
Study Methods
The study compared examinations obtained in breast cancer survivors before and after implementation of DBT at Mass General. During the three-year period before implementation (March 2008 to February 2011), 9,019 DM examinations were performed on 4,085 women. Afterward (January 2013 to December 2017), 22,887 DBT examinations were performed on 7,154 women.
Mammograms obtained during the transition period to DBT (when some patients had their first DBT examinations) were excluded to avoid selection bias.
Performance Metrics
After adjustment for age, race, breast density, presence of a previous screening mammogram and interpreting radiologist:
- The cancer detection rate was higher in the DM group by 2.3 cancers per 1,000 examinations; post hoc analysis showed that the study was underpowered to detect whether that difference was statistically significant
- The abnormal interpretation rate was lower in the DBT group (5.8% vs. 6.2%; OR, 0.80; P=0.001)
- Specificity (the true-negative rate) was higher in the DBT group (95.0% vs. 94.7%; OR, 1.23; P=0.003)
- The DBT group had fewer false-positive results (5.0% vs. 5.2%; OR, 0.82; P=0.004)
Tumor Characteristics
The ratios of screening-detected invasive to in situ cancers were statistically similar (72% invasive cancers with DBT versus 74% with DM).
At Mass General, surveillance mammography for high-risk patients is generally alternated with breast MRI every six months. Of the 86 interval cancers in the two groups, 58% manifested with symptoms and 33% were detected with screening MRI. (Interval cancers are those found after a negative screening test and before the next screening.)
Takeaway Messages
- The switch to DBT led to fewer women being recalled for additional imaging, although the absolute reduction was only 0.4%
- DBT may not improve cancer detection if an institution attains a high cancer detection rate with DM alone
- Because it reduces false-positive findings, DBT could lessen the anxiety experienced by most breast cancer survivors when they present for mammography
- Supplemental screening with modalities such as MRI will continue to be useful for breast cancer survivors in the DBT era
view original journal article Subscription may be required
Learn more about the Department of Radiology
Refer a patient to the Department of Radiology