A Better Read on DBT
By Beth W. Orenstein
Radiology Today
Vol. 21 No. 6 P. 10

A five-year study shows digital breast tomosynthesis outperforms digital mammography.

The evidence is strong that screening mammography helps detect many breast cancers in their early stages, when they are most treatable. While debate continues over the appropriate age to start screening and at what intervals women need to be screened, digital breast tomosynthesis (DBT), sometimes called 3D mammography, is emerging as the best screening tool for average-risk women and even those with dense breasts. A number of studies have found that DBT is superior to standard 2D mammography; multiple studies have shown DBT decreases recall rates and increases cancer detection rates.

“Tomosynthesis allows viewing of the breast in multiple layers or slices,” says Emily F. Conant, MD, a professor and chief of breast imaging in the department of radiology at the Perelman School of Medicine at the University of Pennsylvania in Philadelphia. “This ability to scroll though slices of otherwise overlapping breast tissue helps us not only detect more cancers but also better characterize benign or normal areas of the breast.” With DBT, Conant adds, “you can remove some of the overlapping or obscuring breast tissue so that both normal and abnormal findings are better seen. That provides both improved cancer detection and decreased false-positives.”

Most studies of DBT are based on data from a single initial screening round—that is, until now. Conant is lead author of a study published earlier this year in the journal Radiology that examined multiple years and rounds of screening. The issue with the first round of screening is that the incidence of cancer detection rates and recall rates are higher than with subsequent rounds of screening. Conant’s study looks at the performance of DBT over time.

“The longer follow-up is a big deal and is the longest follow-up with cancer registry matching that has been published to date,” Conant says.

Conant and her colleagues began their study in fall 2011, after the University of Pennsylvania was able to convert all of its patients from 2D screening to DBT screening on one day. “We wanted to do that one-day conversion so that we could analyze the ‘before tomosynthesis’ and ‘after tomosynthesis’ outcomes at our university,” Conant says. The study set included more than 56,000 DBT and 10,500 prior exams done with 2D digital mammography. The research team used local and state cancer registries to track the results over time.

“We wanted to have cancer registry matching, so we would know that every woman we screened had at least a one-year follow-up and often more,” Conant says. “This way we could see if she developed a cancer that we did not detect, and we would have the ability to look at the sensitivity of tomosynthesis.”

The researchers found that DBT not only improved performance but also that the improved performance was maintained over multiple years. Cancer detection rates were 6 per 1,000 for DBT compared with 5.1 per 1,000 for 2D digital mammography alone. Screening recall rates were 8% for DBT, compared with 10.4% for 2D digital mammography alone.

Another strength of the study, Conant says, is that it relied on a diverse population of women. Black women, who are known to develop more aggressive breast cancer subtypes at an earlier age, made up about one-half of the study group. The researchers used the definition of advanced cancers from the National Cancer Institute–funded Tomosynthesis Mammographic Imaging Screening Trial (TMIST), the first randomized trial to compare two types of digital mammography—2D and 3D—for breast cancer screening.

“What we found is that we were finding very significant cancers—poor-prognosis cancers defined by the TMIST study,” Conant says. “We were finding more in our tomosynthesis group than our 2D group, and that’s a difference from prior studies.” Conant attributes the difference, at least in part, to previous studies having less diverse populations. Their study reported that a higher proportion of DBT- vs digital mammography–identified cancers were invasive (70% vs 68.5%, respectively). They also found that a greater percentage of DBT-identified tumors were advanced cancers (32.6% vs 25%, respectively).

Added Value
Samantha L. Heller, PhD, MD, an associate professor of radiology and program director of the breast imaging fellowship at New York University School of Medicine, says many previous studies have shown that DBT can help find more cancers than 2D mammography. In addition, multiple studies have shown that the use of DBT can decrease additional imaging and follow-up tests because radiologists can more easily distinguish between normal and abnormal mammographic findings, she says. What’s important about Conant and colleagues’ study, Heller says, is that the favorable cancer detection rates and decreased recall rates of DBT are sustained over multiple rounds of breast cancer screening.

Also, of note, Heller says, is that Conant and colleagues’ study shows that the cancers detected by sustained DBT screening may be more aggressive than cancers detected by 2D mammography. “These are the kinds of breast cancers that we would like to be able to detect early, so that they do not spread, and this finding is therefore of major interest,” she says.

Stamatia Destounis, MD, FACR, FSBI, FAIUM, of Elizabeth Wende Breast Care and a clinical professor at the University of Rochester Imaging Sciences in New York, says the Conant team’s study answers many of the criticisms of groups and organizations that oppose breast cancer screening for average-risk women. Those who oppose screening say it finds cancers that are slow growing, causing women unnecessary anxiety when they are called back for additional imaging or told they need biopsies that turn out to be negative. This study shows that DBT not only finds cancers that are of greater concern, but also reduces callbacks and false-positives, according to Destounis, who is also a member of Radiology Today’s Editorial Advisory Board.

This is especially true in women who have dense breast tissue. Women with dense breasts are more difficult to screen, and dense breast tissue can mask cancers. Conant says it has been well established in the literature that DBT is better for women with dense breasts.

“It finds more things and decreases false-positives,” Conant says. “But as we get to extremes of breast density, in very, very dense women, it can’t find everything.” That’s why, she adds, every woman needs to be assessed according to breast density and personal demographic risk. “All women should be screened with tomosynthesis, but women who have increased risk, whether because of breast density or something in their health history that tips the balance toward high risk, may need additional screenings with different modalities, such as ultrasound or MRI,” Conant says.

Destounis says DBT is better for most women with dense breasts, but she would also recommend ultrasound for women with very dense breasts because “smoothly outlined masses may hide in very dense breasts, even with DBT.” The best test for women with a family history of breast cancer is MRI, regardless of breast density, Destounis says. “MRI is the best test for high-risk women.”

Availability and Cost
So, what is the availability of DBT in the United States? Although it was approved by the FDA in 2011, “Not every facility has transitioned to DBT completely, but many have some DBT units,” Destounis says. “Women have to call the facilities near them to see whether they have the DBT technology at their center. Many facilities do have DBT now. The Mammography Quality Standards Act Program tells us there are a total of 8,695 certified facilities across the country and, of those, 5,989 have DBT.”

Robin Shermis, MD, MPH, FACR, director of ProMedica Breast Care in Toledo, Ohio, says his facility has been using DBT exclusively for the past five years because the image quality is significantly better. Shermis is also a member of Radiology Today’s Editorial Advisory Board.

Conant says a paper by researchers at Yale University published in JAMA Internal Medicine in June 2019 sheds light on the issue. The authors, Conant says, found that “it is very variable where DBT is being adopted. It seems to be more available in more privileged socioeconomic areas than poorer areas.”

DBT comes with a higher cost, as well. Destounis estimates that it costs approximately $65 more per study than 2D mammography for insured patients. However, she says, the extra cost is justifiable.

“It is a minimal increase when you consider that every woman gets hundreds of images between the two breasts that the radiologist interprets, instead of the standard four [two per breast with 2D],” Destounis says. “In addition, this is much more time intensive for the doctor interpreting the DBT mammogram. The files of these images are large and require additional storage and recovery capabilities and archiving. There also are additional costs of network bandwidth to transfer images to workstations for interpretation, and much more expensive mammography equipment to produce the images. For many facilities, it requires changes to your PACS systems to be able to perform the interpretation.”

Most insurance companies and Medicare cover DBT, Destounis says. However, she has found that some insurances underwritten outside of New York State, where she works, have been extremely resistant to changing their reimbursement and have not covered DBT, regardless of the significant amount of literature supporting it.

Conant believes her paper should help encourage reimbursement. The key, she says, is to demonstrate cost-effectiveness. “The reason tomosynthesis is worth more and should be reimbursed at a higher level is because we are finding more cancers and decreasing false-positives, and that saves money,” Conant says. “Not only are we finding more cancers, we are finding them more efficiently, cutting down on some of the extra imaging, and going directly to ultrasound to plan the necessary biopsy. The improved efficiency is because we have more information on the tomosynthesis image than on a 2D image.”

In an earlier paper, Conant says, “we published that tomosynthesis is a more efficient process to determine whether a biopsy is needed or not.” She adds that she’s interested in using long-term data to understand the cost of treatment and outcomes for patients. “Tomosynthesis is more cost-effective because it decreases false-positives and finds more cancers, but the question is, how much and what value does that add?” she says. “People are looking at this important question to determine what the appropriate reimbursement should be.”

In an editorial accompanying Conant and her colleagues’ study in Radiology, Heller and Linda Moy, MD, a professor of radiology at New York University, wrote that this study addresses major and unresolved questions that must be answered as the breast imaging community continues to adopt DBT. “Ultimately,” they conclude, “such work may help us address the most critical consideration of all: the impact of tomosynthesis on mortality.”

Conant agrees that DBT is “not going to find everything.” But it has shown its value, and it should become the standard of care, if it hasn’t already, she says.

— Beth W. Orenstein of Northampton, Pennsylvania, is a freelance medical writer and regular contributor to Radiology Today.