Breast cancer screening with digital breast tomosynthesis (DBT) offered advantages over digital mammography (DM), including improved cancer detection and lower false-negative rates, researchers reported.
Among 380,641 screening examinations (n=183,989 DBT), DBT rates trended lower for overall false-negative (FN) examinations at 0.6 per 1,000 screens versus 0.7 per 1,000 screens for DM (P=0.20), according to Melissa Durand, MD, of the Yale University School of Medicine and Smilow Cancer Hospital in New Haven, Connecticut.
Also, symptomatic FN examinations were 0.4 per 1,000 screens for DBT versus 0.5 per 1,000 screens for DM (P=0.21), they wrote in Radiology. The study was presented during the Radiological Society of North America (RSNA) virtual meeting.
“Our results build on past studies that have shown that digital breast tomosynthesis improves performance outcomes for breast cancer screening,” Durand said in an RSNA press release. “With DBT, we show we are detecting more invasive cancers, but they are cancers with favorable prognostic criteria, which means these patients would have more treatment options.”
However, Durand explained that in order to have found a statistically significant difference in FNs, the researchers would have needed more than 2 million screens. Nonetheless, the results follow trends observed in previous studies comparing DBT and DM, she stated.
“With DBT, we show we are detecting more invasive cancers, but they are cancers with favorable prognostic criteria, which means these patients would have more treatment options,” Durand said. “Using FN cancer rates is a way to get an idea of how a tool can affect morbidity/mortality in a more reasonable time frame than a randomized clinical trial.”
In an accompanying editorial, Elaine Schattner, MD, of Weill Medical College in New York City, wrote, “Simply put, both methods are excellent. While a radiologist might deem this a negative finding, a woman contemplating mammography might be reassured that, if she has early-stage breast cancer, either digital method (DM or DBT) is likely to reveal it.”
But Schattner, who is a breast cancer survivor, cautioned that “My concern, as a women’s health advocate, is that such high-quality screening be available to all women…having state-of-the-art equipment is not sufficient to guarantee valid interpretation of images. Apart from the issue of some facilities having newer or older equipment, heterogeneity of radiologists’ training, experience, and skills will influence the accuracy of this test.”
The authors reported that race or ethnicity information was available for 82.5% of women with screen-detected cancers, and that “With DBT, the percentage of symptomatic FN cancers was significantly lower in White women than in Black women, compared with those screened with DM. This could suggest differences in access to DBT between these populations, although without knowing the race or ethnicity of all study participants, this cannot be definitively stated.”
Durand pointed out that “digital breast tomosynthesis equipment does cost more but reimbursement is also higher by most insurances, so the return on investment is rapid.” She added that screening outcomes are “significantly improved with DBT, which translates to more efficient workflow. Because there are fewer recalls, fewer additional studies need to be done, and when an abnormality is found on DBT, the work-up or evaluation of the finding usually needs less additional imaging, or can go straight to ultrasound, compared to abnormalities detected on 2D digital mammography alone.”
The authors conducted the retrospective study at 10 academic and community practices. DM screening examinations 1 year before DBT implementation and DBT screening examinations from the start date until June 30, 2013, were linked with cancers through June 30, 2014. Data was collected in 2016 and analyzed in 2018-2019.
“Cancers after FN examinations were characterized by presentation, either symptomatic or asymptomatic. FN rates, sensitivity, specificity, cancer detection and recall rates…were compared,” they wrote.
The authors also reported that asymptomatic FN rates trended higher in women with dense breasts at 0.14 per 1,000 screens for DBT versus 0.07 per 1,000 screens for DM (P=0.07).
Additionally, DBT improved sensitivity compared with DM (89.8% vs 85.6%, P=0.004) and specificity (90.7% vs 89.1%, P<0.001).
Also, Durand and colleagues found that cancers identified with DBT were more frequently invasive (P<0.001), had fewer positive lymph nodes (P=0.04) and distant metastases (P=0.01), as well as lower odds of an FN finding of advanced cancer (odds ratio 0.9, 95% CI 0.5, 1.5).
As for recall rates, they were lower for all breast densities with DBT except for the predominantly fatty category for other density categories (P<0.001 vs P=0.47).
Study limitations included the retrospective design, variations across screening cohorts, “which were enough to show significance but likely were not clinically meaningful,” and lack of data “on an individual level whether some women may have been represented in both the DM cohort and the DBT cohort, which could introduce bias,” the authors noted.
Also, the impact of DBT on patient survival has not been established and “Mortality data is beyond the scope of the present study,” according to Durand.
The study was funded by Hologic.
Durand disclosed relevant relationships with Hologic.
Schattner disclosed no relevant relationships with industry.