This article is more than 5 years old. Information may no longer be current.
Tomosynthesis plus mammography increases breast cancer detection by 90%
Click Here to Manage Email Alerts
The addition of tomosynthesis to digital mammography improved the detection of breast cancer by 90%, according to researchers in Italy.
“Tomosynthesis and digital mammography is much more sensitive than digital mammography,” Pierpaolo Pattacini, MD, radiologist and director of the radiology department at AUSL Reggio Emilia in Italy, said in a press release. “The vast majority of the advantage is due to tomosynthesis alone.”
The researchers performed a randomized controlled trial in which women aged 45 to 70 years were randomly assigned to receive either digital mammography alone (n = 9,783; mean age, 56.3 years) or with tomosynthesis, also called 3-D mammography (n = 9,777; mean age, 56.2 years). Pattacini and colleagues evaluated sensitivity, recall rate and positive predictive value for each arm, and determined the ratios of each rate between arms.
The recall rate was 3.5% in both arms.
In the combined group, detection was approximately 90% greater than that of the control group (8.6 per 1,000 vs. 4.5 per 1,000; relative detection, 1.89; 95% CI, 1.31-2.72).
The positive predictive value in the combined arm was 24.1%, compared with 13% in the control group. The combined arm also had a lower false-positive rate (27 per 1,000 vs. 30 per 1,000).
Tomosynthesis alone detected 72 of the 80 cancers detected in the combined arm.
The greater detection rate in the combined arm was stronger for ductal carcinoma in situ than for invasive cancers (relative detection, 2.8; 95% CI, 1.01-7.65).
For invasive cancers, the increase in detection appeared similar both for tumors smaller than 10 mm and those between 10 mm and 20 mm.
Researchers observed no gain in the detection of large cancers, however.
They noted that it would be necessary to weigh the benefits of the combined screening method against possible harms.
“If these small cancers would never become life-threatening, then we are increasing overdiagnosis and not impacting mortality,” Paolo Giorgi Rossi, PhD, director of the epidemiology unit at Reggio Emilia and a co-author of the study, said in the press release. “Thus, we need to have a measure of the impact of this higher detection rate on the incidence of advanced cancers and interval cancers in the following years.”
The combined method also required a longer reading time compared with mammography.
“For publicly funded screening programs, this increased reading time would be a big issue, destroying their sustainability,” Rossi said. “Implementing tomosynthesis in public screening programs would require rethinking protocols and reading technologies to reduce or eliminate the extra costs.” – by Andy Polhamus
Disclosures: Pattacini reports travel and lodging expenses and fees for presentation of the study design and preliminary results at several congresses and meetings from GE Healthcare for the submitted work, as well as speaker fees and travel grants from GE Healthcare outside of the submitted work. Please see the study for all other authors’ relevant financial disclosures.
Perspective
Back to Top
Breast cancer screening decreases mortality. The problems of overdiagnosis, false positives and false negatives remain as limitations to screening. The Reggio Emila trial evaluates "state-of-the-art" imaging with digital mammograms alone vs. digital mammograms and digital breast tomosynthesis (DBT) in a prospective randomized trial with more than 19,000 women. The results report the first imaging only — reflecting the randomization — with plans for digital mammograms subsequently. The primary endpoint of the trial is interval cancers, and the cumulative incidence of advanced cancers after 4.5 years is not reported.
These detailed results focus on the initial recall rate, detection rate, positive predictive value and number of false negatives. The total numbers of breast cancers and ductal carcinomas in situ was statistically significantly higher with digital mammograms plus DBT vs. digital mammograms only (83 vs. 44). The overall recall and false-positive rates were not statistically significantly different. That the recall rates were the same is surprising given the addition of DBT to digital mammogram usually decreases recall rates. The positive predictive value rates were just about double for digital mammograms plus DBT vs. digital mammograms (24% vs. 13%).
Several things about these results are worth highlighting. First, a 13% positive predictive value for digital mammograms is low in the Reggio Emilia trial, whereas it was 27% to 28% in National Performance Benchmarking Study from the Breast Cancer Surveillance Consortium. This lower-than-expected positive predictive value might be due to chance or some imbalance in the randomization.
Second, in subset analyses in the cohort of about 5,000 women aged 45 to 49 years, there were no statistically significant differences in detection rates and positive predictive value. These results are likely due to the increased density of breast tissue. Here, the results of primary endpoints, interval, and advanced cancers are the endpoints of most interest.
Third is the question of extra costs, in dollars, and including time, of adding DBT to digital mammograms when the value (ie, the detection of interval and advanced cancers over the total costs) is unknown.
Finally, it is instructive to look at what breast cancers are detected. A higher percentage DCIS, as well as early-stage invasive breast cancers, was observed in the digital mammogram plus DBT arm. Overall, DCIS and these lower-stage, lower-grade and lower Ki-67 proliferation index breast cancers are more likely detectable, raising the question of overdiagnosis. Understanding the contributions of DBT to digital mammogram awaits reporting of the primary endpoints.
Sprague BL, et al. Radiology. 2017;doi:10.1148/radiol.2017161519.
Collapse
Pattacini P, et al. Radiology. 2018;doi:10.1148/radiol.2018172119.
Click Here to Manage Email Alerts