You hear about it on the radio, on TV and see it in the health system newsletter. Your facility has bought in or is close to sending out an RFP. The buzz is all about digital breast tomosynthesis (DBT). Experts weigh in on its prowess, performance and power in cutting recalls. Is it ready for primetime?
Tomosynthesis’ uniqueness comes from its joint 2D and 3D imaging capabilities, producing a comprehensive image of the breast. With an increase in cancer detection rates, enthusiasm has spread rapidly, largely due to the results of two marquee studies: the Oslo Breast Screening Trial from Per Skaane, MD, PhD, of Oslo University Hospital Ullevaal in Norway, and colleagues, and the Italian STORM study from Francesca Caumo, MD, of the Centro di Prevenzione Senologica in Verona, Italy, and colleagues. Both studies demonstrated an increase in cancer detection rates and a decrease in false-positive recalls with DBT, leading to a fairly quick adoption of tomosynthesis in clinical practices.
The technology has proven effective as well. Brian M. Haas, MD, of Yale University in New Haven, Conn., and colleagues found a 30 percent reduction in recalls (Radiology 2013; 269:694-700). Stephen L. Rose, MD, of the TOPS Comprehensive Breast Center in Houston, and colleagues discovered a 37 percent recall reduction, as published in the June 2013 issue of American Journal of Roentgenology. As for cancer detection, Hass et al showcased a 9.5 percent increase in detection rate. Rose and colleagues found a 35 percent increase in detection.
“In our experience, invasive cancer detection is up 27 percent, which closely mirrors that of the large Norwegian study,” says Emily Conant, MD, of the Perelman School of Medicine at the University of Pennsylvania in Philadelphia. “This means that in our ongoing ‘natural experiment’ in which we have been imaging all of our over 22,000 screening patients with tomosynthesis, our cancer detection rate is up, particularly for the most clinically significant cancers: the invasive ones.”
Screening patients with dense breasts is one area experts wonder if tomosynthesis will have an impact, as mammography currently misses about half the cancers in dense breasts. As legislation requiring breast density notification, which is currently implemented in 14 states, informs more women about their breast composition, more are expected to request supplemental screening.
“We will need good data based on density, age and complexity to see if dense breast legislation will influence the use of tomosynthesis,” continues Conant. “Early literature shows that DBT reduces recall rates and increases detection rates across all densities, which suggests that these outcomes are due to more than density alone. Perhaps the improvements with tomosynthesis are due to other complementary factors, such as breast texture or the complexity of the image.”
“More and more practices are incorporating tomosynthesis into screening programs,” says Constance Lehman, MD, PhD, FACR, of the University of Washington in Seattle. “The technology went through a lot of iterations, and it took a lot of time to get to where it is today. However, the impact of this technology on clinical outcomes for our patients needs careful evaluation and the information to date is still preliminary.”
Some echo this idea and are more hesitant to fully embrace DBT just yet. David Dershaw, MD, FACR, of Memorial Sloan Kettering Cancer Center in New York City, says: “I know a lot of people have acquired tomosynthesis and are using it. There’s a moderate amount of enthusiasm about the whole thing, but in terms of published numbers there’s not very much out there.”
Going with the flow
Despite the many promises tomosynthesis has to offer for breast screening and detection, concerns remain. In particular, the technology presents challenges to radiology workflow due to the complexity of the images produced.
“Radiologists are spending twice as much time with tomosynthesis as a screening study, which is about 90 seconds of interpretation,” says Christoph Lee, MD, MSHS, of the University of Washington in Seattle. “However, the amount of time invested in screening interpretation may be compensated with the amount of time saved on the diagnostic end.” Lee suggests that a possible solution for workflow issues may be found in batched screening, allowing radiologists to use a quiet room to avoid interruption and read multiple studies at a time.
“Literature is saying that the interpretation time of tomosynthesis studies is decreasing,” Conant adds. “There is a learning curve to this. Moreover, the confidence we gain in seeing the breast in a 3D format far outweighs the additional time.”
Dershaw, however, offers a different perspective on the increased interpretation time. “There’s a real economic downside to tomosynthesis in that it’s currently simply a matter of radiologists doing it pro bono,” he says. “Its value is comparable to doing an additional view with mammography; the amount of time it takes to perform tomosynthesis on everybody is much greater than doing an additional view on 8 to 10 percent of patients. Those who argue that tomosynthesis is a time-saving procedure would have to prove it to convince me.”
DBT requires a time commitment even before a practice performs its first scan, as the FDA mandates eight hours of training specifically in tomosynthesis to adopt DBT into practice.
Beyond workflow challenges, reimbursement issues continue to loom large for the implementation of tomosynthesis. “There’s currently no direct code for reimbursement,” explains Lee. “Medicare is not really acknowledging the use of an accessory Current Procedural Terminology (CPT) Code for billing purposes. Some third parties will pay for an accessory CPT Code that can be added and some will pay a small amount of money in addition to mammography. More often than not, the service is provided for free or patients have to pay out of pocket to undergo tomosynthesis.”
“Reimbursement is a very real issue and until that gets dealt with, it’s going to be a difficult technology to spread around the country because patients have to pay for it out of pocket,” Dershaw says. Conant, on the other hand, seems to have a more positive take on the situation. “The remarkable thing is that tomosynthesis is spreading even though it’s not reimbursed. I think a push from patients is driving some of its continued adoption.”
Concerns about increased radiation dose also must be addressed before there’s wider adoption of tomosynthesis in clinical practice. The FDA has approved the use of the technology with mammography, in turn delivering double the dose to patients. Though this dosage remains within the acceptable range delineated by the FDA, it still poses a risk to patients.
“It’s been estimated that if we do tomosynthesis on women as a routine study, two cancers per 100,000 patients will be induced in women,” explains Dershaw. “This number is small, but it is real. The question is whether or not we should increase the amount of radiation patients are receiving because of a subsequent benefit. If the established benefit is simply to decrease call backs on screening, which is 8 to 10 percent of patients, then the remaining 90 percent are being penalized with additional radiation and no incremental advantage.”
While the increased dose of performing both a 2D and tomosynthesis study poses a risk for some, this is not a problem without a solution. “It’s going to get better certainly in terms of dose,” says Conant. “New software is now able to reconstruct a 2D image, or conventional ‘looking’ image from the 3D tomosynthesis projections, producing synthetic images from routine screenings and eliminating the doubling of the dose.” This software, which was approved by the FDA in May 2013, has already been used for several years in Europe. While there’s limited data to describe similarities between the performance of the synthetic 2D views and the standard, early evidence suggest comparability, according to Lee.
Aside from the prospect of synthetic image production, Conant also mentions the need to personalize screening. “Double dose isn’t sustainable for every woman every year. We need to assess if this screening needs to be done every year by looking at who benefits the most from the technology, how often they should be screened, how outcomes can be maximized and how exposure can be limited.”
From infancy to maturity
Given the obstacles that accompany the implementation of DBT in clinical practice, the details behind its use must be flushed out to ensure effectiveness as well as cost savings. “A lot of the technology’s future will depend on comparative effectiveness research that will be collected and analyzed by using large patient registries as sources of data. There is a lot of promise with tomosynthesis,” says Lee. Conant also asserts the need for more quality research to demonstrate improvements in patient outcomes and back evidence-based decisions about who will benefit the most from DBT.
Despite potential challenges, the breakthrough technology already offers promise for improved breast cancer screening and detection in women and is expected to continue to do so. “Upcoming data will drive the uptake of this evolving modality,” says Conant.
“Tomosynthesis is the next evolution of breast screening,” adds Lehman. “This provides a teaching moment to remind people that all mammography isn’t the same and that we are continuing to see improvements in the technology available.”
Not only will comparative effectiveness research better the discussion surrounding tomosynthesis, but engaging with patients and patient advocacy groups will aid in discerning the technology’s use too.
“You have to see a large volume of patients before you can truly appreciate how tomosynthesis will improve your efficiency and diagnostic work up of patients,” says Conant. “It took us a couple of months to really appreciate the impact, but this is indeed real. In my career as a breast imager, tomosynthesis has already had the most important impact on improving the outcomes of mammographic screening. We are only just beginning with this innovative technology.”