Digital mammography has been the standard of care in breast cancer screening since about 2005, when a National Cancer Institute-funded study of nearly 50,000 women across 33 sites in North America demonstrated the technology’s power to improve breast cancer detection—while neither negatively impacting accuracy nor increasing false positives for women with dense breasts, those in the pre- and peri-menopausal stages and those under age 50. There is, however, another rising contender in the ring: the three-dimensional imaging x-ray modality known as breast tomosynthesis.
In a tomosynthesis scenario, multiple limited-angle projections are acquired from the standard cranio-caudal and medial lateral oblique mammographic screening positions under compression. The images subsequently are reconstructed into slices via algorithm for z directional viewing of mammographic breast tissue. The technology is seen as a means of peering through layers of overlapping breast tissue that may be obscuring the visibility of breast lesions.
Five years after the Food and Drug Administration (FDA) approved the first tomosynthesis device for breast cancer screening, the adoption of tomosynthesis units (with those approved for use in the U.S. featuring the ability to acquire images in both two-dimensional digital form and as three-dimensional tomosynthesis slices) has become increasingly widespread. Some players—the breast imaging division at Massachusetts General Hospital in Boston among them—have made a combination of tomosynthesis and digital mammography a standard protocol for all mammograms. Despite the apparent advantages to be gained from such an approach, the path to migration has posed, and will continue to pose, its challenges.
“It’s not straightforward as it might seem,” sums up Debra Monticciolo, MD, FACR, professor of radiology, Texas A&M University Health Sciences, College Station, Tex.; vice chair for research and section chief, breast imaging, Baylor Scott & White Healthcare department of radiology, Temple, Tex.; and chair, American College of Radiology (ACR) Commission on Breast Imaging. “There is a lot to consider about how the technology will be applied and how it will be integrated into the current workflow.”
Recent studies appear to make a strong case for tomosynthesis. Etta Pisano, MD, vice chair, department of radiology, Beth Israel Deaconess Medical Center, Harvard Medical School, Boston; and senior director of research development at the ACR, says the most compelling clinical evidence for tomosynthesis is for its use as an adjunct to two-view, two-dimensional digital mammography. In an interview with Radiology Business Journal and a paper (1) published in Clinical Imaging this past fall, Pisano points to several studies that demonstrate the effectiveness of tomosynthesis in decreasing patient recalls and increasing cancer detection when combined with digital mammography.
One such study (2), a retrospective performance comparison of tomosynthesis partnered with digital mammography and digital mammography alone, involved 454,850 screening examinations—173,663 using the combined technologies and 281,187, only digital mammography—at 13 academic and non-academic breast imaging facilities in the U.S. A “significant” reduction in patient recall rate—16.1 (P<.001) per 1,000 subjects (3)—was recorded across all sites for tomosynthesis plus digital mammography versus digital mammography alone. The study also revealed a “significantly higher” cancer detection rate of 0.12%—with a 95% level of confidence that the rate falls between 0.08%, 0.16%—for tomosynthesis utilized in conjunction with digital mammography.
Pisano also cites trials conducted in Norway (Oslo Trial)(3) and Italy (STORM)(4). The former entailed the screening of 12,621 women at a single facility. Results of this trial encompassed a 27% increase in breast cancer detection for tomosynthesis used in tandem with digital mammography; the ratio of false positive rates of digital mammography and tomosynthesis plus digital mammography was 0.85, with just a 1 in 1,000 chance that this number is inaccurate. All studies were reviewed by four independent radiologists, with a requirement for arbitration in instances where one or more clinicians classified a case as BIRADS 2 or higher.
The STORM trial included 7,292 participants from multiple imaging centers and spanned two independent parallel mammography interpretations (an approach that is common in Europe, according to Pisano). The overall cancer detection rate for the two trials was 8.1 per 1,000 women screened via two-view tomosynthesis and two-view digital mammography and 5.3 per 1,000 patients screened using only two-view digital mammography. The differential of 2.7 per 1,000 study participants can be considered very telling, Pisano states.
Monticciolo corroborates Pisano’s comments about the general value of breast tomosynthesis in cancer detection and reducing false positives that result in recalls, adding that patients with “heterogeneously or extremely dense” breasts are most likely to benefit from exams conducted using the technology. “The rationale here is that dense breast tissue can mask tumors, so if we slice through, we may be able to ‘get around’ the density,” she explains. “However, the ACR has gotten away from applying percentages—in other words, x percent density, as a measure of when tomosynthesis is a better idea and when it isn’t. Everything needs to be considered on a case-by-case basis, with what is best for the patient at the center.”
To further complicate decision-making, the definition of density is changing.
“The ACR BI-RADS no longer defines density as percentage of dense tissue in quartiles,” Monticciolo says. “This has changed how we look at density for each patient. It may affect how an individual practice applies tomosynthesis. Ultimately, we are trying to apply it to those patients who will benefit most.”
At present, most clinicians who perform tomosynthesis-based exams do so as an adjunct to digital mammography and obtain both two-dimensional (standard) and three-dimensional views of the breast, thereby doubling the necessary radiation dosage. That dosage “is still under the 0.6 limit, as required by the Food & Drug Administration (FDA), and it is safe, but we still have to be mindful of radiation exposure,” Monticciolo observes.
Some practitioners, she says, work around this issue by replacing standard images with synthesized views of breast tissue, made up of reconstructed tomosynthesis slices. However, Monticciolo favors the “standard view approach” when opting for tomosynthesis; she believes it affords a higher resolution and a better exam all around.
Clinical trial results and practitioner preferences notwithstanding, there has been a recent push for a transition to the use of tomosynthesis alone when patient parameters warrant, reports Laurie Fajardo, MD, MBA, professor of radiology, University of Utah, Salt Lake City. Fajardo has been involved in two tomosynthesis studies, the results of which will soon be submitted to the FDA.
One trial centers on a combined two-dimensional/three-dimensional breast-imaging unit; the other is aimed at assessing an “improved” tomosynthesis system that produces only three-dimensional images. The latter is said to leverage “better” image reconstruction algorithms that would take the place of synthesized images and eliminate the need for a digital mammography component in exams.
“The data from this trial will determine if radiologists interpret tomosynthesis studies without 2D dose-acquired or synthetic images as well as mammograms,” Fajardo states. “I think it is an effective and efficient way to read. It also addresses the double radiation dosage issue quite well.”
Not surprisingly, economics comprises another moving part of the tomosynthesis puzzle. On October 31, 2014, the Centers for Medicare and Medicaid Services (CMS) approved Medicare reimbursement for women undergoing a tomosynthesis exam in conjunction with digital mammography beginning in 2015. The reimbursement is approximately $57, in addition to the fee for the two-view digital mammogram. Laws mandating that private payors cover tomosynthesis have been passed in Illinois and Pennsylvania.
“That’s the good news,” reports Geraldine McGinty, MD, MBA, assistant professor of radiology at Weill Cornell Medical College in New York City and chair of the ACR Commission on Economics. “The less positive news is that many private payors continue to call tomosynthesis ‘experimental’ and [therefore] not warranting coverage. Blue Cross is in this camp, and so is Cigna, for example. This is all despite the ACR’s position on tomosynthesis, which is that it is a highly promising technology.
“Evidence collected to date suggests that it reduces additional imaging and may improve cancer detection in some women. The ACR continues its focus on using imaging technology as effectively and as appropriately as possible, and we need additional research to determine for which cases it is most helpful. Unless payors provide reimbursement for tomosynthesis, it will be difficult to further facilitate the outcomes research and gather the supporting data.”
Complicating matters, McGinty notes, payors’ refusal to cover tomosynthesis has forced some practices to swallow the additional cost with which it comes. “This is not a sustainable model for the long term; radiology practices and practitioners cannot be expected to give widespread care for free,” she asserts.
The ACR has been attempting to turn the tide and persuade private payors to pick up the tab for tomosynthesis; as such, it has engaged in calls with these entities to help them understand why and how tomosynthesis does not fall into the category of experimental technology. An “advocacy packet” containing the ACR’s position statement on tomosynthesis, along with explanatory articles, has also been prepared and is being distributed to the private payor community.
Similar initiatives by the ACR to educate referring physicians about tomosynthesis and its benefits are underway as well. “We have made our position quite clear,” McGinty emphasizes.
Just as the economics of tomosynthesis appear to be a moving target, so, too, do practice patterns when tomosynthesis becomes part of the mix. “It’s an evolution,” states Mark Alson, MD, FACR, of Women’s Imaging Specialists in Healthcare (WISH). WISH operates three women’s imaging centers in Fresno and Clovis, Calif., including the WISH Fresno Breast Center. The latter installed its first tomosynthesis machine in October 2012, after attendance at courses and extensive reviews of literature led Alson and his colleagues to believe that going the tomosynthesis route would decrease recall rates and increase the likelihood of cancer detection. A second unit was added the following year; the center also has two digital mammography units.
Although he could not quantify the changes tomosynthesis has wrought for his practice, Alson says it is fulfilling its decreased recall rate and enhanced detection promise, and that any of its digital mammography equipment being retired going forward will therefore be replaced with a tomosynthesis unit. Currently, approximately 70% of screenings conducted at the WISH Fresno Breast Center are executed via tomosynthesis, as are an estimated 80% of diagnostic procedures.
In keeping with the transition, “we have had to develop our own internal criteria for when to utilize tomosynthesis and when not to do so,” Alson explains. All baseline mammograms fall into the tomosynthesis category, since these exams do not have the benefit of prior exams for comparison; and in order to give Alson and his colleagues what they consider the most precise basis of comparison going forward. Tomosynthesis also is used in screening women with dense or “complex” breasts. Alson defines the latter as “nodular” or displaying any unusual pattern (e.g., overlapping tissue) that would, if detected via digital mammography, spark a recall.
Additionally, the imaging center’s protocol holds that tomosynthesis screenings be performed on patients who have undergone a lumpectomy, as digital mammography renders it difficult to distinguish between a lumpectomy scar and a new malignancy that is too far away from the scar to be scar tissue. Women with a strong family history of breast cancer and those whose referring clinicians (primary care physicians, OB/GYNs, and surgeons) request a tomosynthesis exam are on the list as well, Alson notes. There are certain exemptions; for example, older women and those who cannot comfortably stand and hold their breath while remaining steady on their feet as dictated by the tomosynthesis procedure are screened using digital mammography alone.
“Seventy is really our relative cut-off age-wise,” Alson observes. “This is not to say that there aren’t 85-year-old women out there who can tolerate the longer period of standing for the exam, but in general, it’s more the women difficult for women over 70.”
As for the diagnostic side, the WISH Fresno Breast Center employs tomosynthesis on any patient with what Alson refers to as a “focal problem”—i.e., a woman who has felt
a lump in her breast during a self-
examination, or is experiencing repeat occurrences of a phenomena (e.g., pain) in an identical area of the same breast. “We favor tomosynthesis in these focal cases because it allows us to visually dissect the tissue,” Alson asserts. “Discrete slices make it so that overlapping tissue doesn’t mask the abnormality.”
In her practice, McGinty utilizes tomosynthesis in a majority of screening exams (although some patients decline it, primarily for payment-related reasons) and on the diagnostic side, as a “problem-solution” tool (e.g., if conventional digital mammography is not offering enough imaging data to enable a definitive diagnosis). Monticciolo says her department utilizes tomosynthesis primarily as a screening tool.
Then, there are the workflow and staffing challenges. The fact that tomosynthesis studies that incorporate digital mammography requiremore time to conduct and interpret than do digital mammography studies alone (the interpretation step alone can take up to twice as long, by most estimates) led to some initial changes in workflow for Alson and his team. At first, technologists could conduct 30% fewer exams, and practitioners read significantly fewer studies, over an eight-hour period.
This situation remedied itself as all parties grew accustomed to the technology. “I need to read the same number of exams as I always have, and with experience I am able to do so,” Alson says. “Experience has allowed me to reduce the time from two-and-a-half to double.”
Monticciolo, meanwhile, thinks that workflow and what some sources call the “2.5 problem” can be mitigated somewhat by optimizing image hanging protocols that facilitate the interpretation of images produced in tomosynthesis studies “in the most logical order possible”; the protocol will be refined if necessary down the road.
Only so much that can be done, however, as tomosynthesis results in so many more images than a standard exam. “Even working as efficiently as possible, it just takes longer to read the additional images,” she says.
Fajardo has noticed that the more radiologists learn to rely primarily on the tomosynthesis portion of combined studies (Figure, page 31), the smoother workflow becomes. “We are at the point where we don’t scrutinize the two-dimensional images as much anymore,” she states. “Instead, our strategy is to use digital mammography as a roadmap to see what we need to look at in more detail on the tomosynthesis images.
“Over time, we’ve learned to read the combined [studies] faster. Will we ever read these as fast as we could straight digital mammography exams? Probably so. And if we ever get to a point where we are utilizing only the three-dimensional images, reading will be faster.”
In addition, Fajardo says, there is research being done to evaluate new reconstruction algorithms that reduce the number of images in a tomosynthesis study. For some tomosynthesis
studies, there there can be 70-80 images per projection. “If this can be reduced by 50%, reading times and workflow culd be dramatically improved,” she believes.
Questioning the future
While significant progress surrounding the use of tomosynthesis appears to have been made, the modality is, as sources point out, still in its infancy. Consequently, there remain many questions to be answered and issues to be resolved.
Pisano observes that there exists little published data on the use of the two tomosynthesis acquisition methods available on systems currently on the market—one allows for a combination exam comprising tomosynthesis and allows for configuration for synthetic digital mammography or conventional direct acquisition; and the other acquires a single-view tomosynthesis in medial-lateral oblique view plus a conventional cranio-caudal digital mammographic view with no synthetic option. At least three other manufacturers, she says, are seeking FDA approval for other tomosynthesis systems, but little is known about their clinical performance.
Even if tomosynthesis does detect more breast cancer occurrences than digital mammography, this in itself does not lead to reductions in mortality from the disease. A significant effort, sources assert, needs to be made to ascertain if the benefits afforded by tomosynthesis justify its use and potential drawbacks.
Within this determination, Pisano states, there exists the question of whether or not the cancers detected through tomosynthesis screening are of the category that progress and eventually “kill women,” and whether embracing the technology will lead to marked increases in over-diagnosis and, hence, overtreatment of cancers that are unlikely to progress and harm patients. Also meriting further examination are whether increased radiation exposure during tomosynthesis, as well as the cost of the procedure, are justified; and if all women—or only those with dense breasts and/or in certain age categories—should undergo tomosynthesis in the first place.
“Tomosynthesis is still in its infancy, and significant research needs to be completed to come up with the answers to these questions,” Pisano emphasizes. When it comes to screening selected sub-populations of women and “problem-solving specific conditions,” she deems it critical to compare the impact of tomosynthesis as a companion to digital mammography against the impact of other adjunctive breast imaging technologies to digital mammography. Such adjunctive breast imaging technologies might include ultrasound for women with dense breasts and magnetic resonance for high-risk cases.
To address some of these issues, Pisano has, under the aegis of the ECOG-ACRIN Cancer Research Group, requested funding for a five-year study from the National Cancer Institute arm of the National Institutes of Health (NIH). The randomized study, of which Pisano will serve as principal investigator, is intended to draw a definitive conclusion as to whether or not tomosynthesis “detects the cancers that kill women, saves lives, and much more,” Pisano explains.
It will entail tomosynthesis trials at 70 or more imaging centers in the U.S. and Canada. “The more concrete benefits of tomosynthesis we can demonstrate and promote, the more the specialty as a whole, as well as payors, will buy into it,” Pisano says. “The end will justify the means.”
- Cole EB, Pisano ED; Tomosynthesis for breast cancer screening. Clin Imaging. 2015, Sept 24. Pii S0899-7071(15)00244-2. Doi:10.1016/j.clinimag.2015.09.014 [Epub ahead of print].
- Friedewald SM, Rafferty EA, Rose SL et al. Breast cancer screening using tomosynthesis in combination with digital mammography. JAMA. 2014;311 (24): 2499-507.
- Skaane P, Bandos Al, Gullien R et al. Comparison of digital mammography alone and digital mammography plus tomosynthesis in a population-based screening program. Radiology. 2013; 267(1):47-56.
- Ciatto S, Houssami N, Bernardi D et al. Integration of 3D digital mammography with tomosynthesis for population breast-cancer screening (STORM): a prospective comparison stud study. Lancet Oncol.