Digital Breast Tomosynthesis and the Challenges of Implementing an Emerging Breast Cancer Screening Technology Into Clinical Practice Christoph I. Lee, MD, MSHSa,b,c, Constance D. Lehman, MD, PhDa,c

Emerging imaging technologies, including digital breast tomosynthesis, have the potential to transform breast cancer screening. However, the rapid adoption of these new technologies outpaces the evidence of their clinical and cost-effectiveness. The authors describe the forces driving the rapid diffusion of tomosynthesis into clinical practice, comparing it with the rapid diffusion of digital mammography shortly after its introduction. They outline the potential positive and negative effects that adoption can have on imaging workflow and describe the practice management challenges when incorporating tomosynthesis. The authors also provide recommendations for collecting evidence supporting the development of policies and best practices. Key Words: Digital breast tomosynthesis, technology adoption, adjunct screening J Am Coll Radiol 2013;10:913-917. Copyright © 2013 American College of Radiology

INTRODUCTION

According to the American Cancer Society, 130,000 US lives have been saved in the past 20 years through early detection and treatment of breast cancer, with substantial credit given to screening [1]. However, mammography remains an imperfect modality, with concerns for potential harms outweighing benefits among certain subgroups of women, including those aged 40 to 49 years [2]. An estimated 10% of screening women with no cancer undergo unnecessary diagnostic imaging and/or biopsy [3]. The potential harms from false-positive screening results provided the impetus for the US Preventive Service Task Force to revise its guidelines in 2009 to no longer recommend routine screening for women 40 to 50 years of age. Over the past decade, screen-film mammography (SFM) has been rapidly replaced by digital mammography (DM), which has comparable accuracy but greater workflow efficiency compared with SFM. In 2011, the FDA approved digital breast tomosynthesis (DBT) for a Department of Radiology, University of Washington School of Medicine, Seattle, Washington. b Department of Health Services, University of Washington School of Public Health, Seattle, Washington. c Hutchinson Institute for Cancer Outcomes Research, Fred Hutchinson Cancer Research Center, Seattle, Washington.

Corresponding author and reprints: Christoph I. Lee, MD, MSHS, Seattle Cancer Care Alliance, 825 Eastlake Avenue East, G3-200, Seattle, WA 98109-1023; e-mail: [email protected]. Dr Lee is supported in part by the GE AUR Radiology Research Academic Fellowship, which is cosponsored by GE Healthcare (Milwaukee, Wisconsin) and the Association of University Radiologists (Oak Brook, Illinois). ª 2013 American College of Radiology 1546-1440/13/$36.00  http://dx.doi.org/10.1016/j.jacr.2013.09.010

all mammographic clinical indications. With FDA endorsement, DBT is diffusing into routine clinical practice with the promise of decreasing false-positives and increasing cancer detection by eliminating DM’s interpretive limitations caused by superimposed breast tissue. Yet, the adoption of DBT is outpacing the collection of clinical effectiveness data and reimbursement policies, leaving individual radiology groups with little guidance regarding whether and how to implement this emerging technology into their practices. In this article, we review the drivers for rapid DBT adoption, compared with the drivers of DM; evaluate the potential impact of DBT on breast imaging workflow and practice management; and provide recommendations for evidence gathering to guide DBT policy and best practices. INITIAL EVIDENCE AND PROMISE

DBT garnered great enthusiasm on the basis of early observer performance studies that showed its equal or better accuracy compared with standard DM [4-6]. Single-institution studies showed adjunct DBT in addition to standard DM to improve diagnostic accuracy, largely because of a reduction in false-positives [7-9]. More recently, two population-based screening studies demonstrated substantial, statistically significant gains in screening performance when DBT is added to DM. Both an interim analysis from the Oslo screening study (n ¼ 12,631 women) and final results from the Italian Screening With Tomosynthesis or Standard Mammography trial (n ¼ 7,292 women) confirmed significant 913

914 Journal of the American College of Radiology/Vol. 10 No. 12 December 2013

reductions in recall rates (15%e17%) and improvements in cancer detection rates (33%e53%) with adjunct DBT in national screening populations [10,11]. As an adjunct tool, DBT has many workflow advantages compared with screening ultrasound and MRI. Because it is built into newer generation mammographic units and is obtained during the same breast compression as standard digital mammographic projections, DBT is associated with little extra time investment for patients and technologists. The addition of screening ultrasound or MRI, in contrast, requires the transfer of patients between examination rooms and both patient and technologist time for image acquisition. Thus, compared with ultrasound and MRI, DBT has the advantage of increased patient throughput, streamlined equipment needs (including purchasing, maintenance, certification, and quality assurance), reduced physical space needs, and reduced training of technical staff members and physicians across modalities. These advantages have led many practices to adopt DBT at an early stage, before the acquisition of sufficient clinical effectiveness data. DRIVERS OF EARLY ADOPTION

To identify the key drivers of DBT’s early adoption, comparison with DM’s adoption over the past decade may be helpful. DM received FDA approval in 2000 for the same screening and diagnostic indications as traditional SFM. However, the major clinical trial (ACRIN’s Digital Mammographic Imaging Screening Trial) demonstrating that DM had similar overall accuracy to SFM was conducted from 2001 to 2003, with results published in 2005 [3]. The study found no statistically significant difference in overall diagnostic accuracy between SFM and DM but did find improved accuracy with DM in premenopausal women and in those with dense breasts. Regardless, DM had already diffused into general radiology practices at the time of reporting. Moreover, a cost-effectiveness analysis comparing DM with SFM for screening was not published until 2008 and demonstrated that using DM for all screening was not more cost-effective than using SFM [12]. However, by that time, DM had firmly supplanted SFM in the majority of US radiology practices. FDA approval is only one step in allowing technology adoption. The main purpose of FDA approval is to determine that new imaging technologies are safe and effective. However, the threshold level of evidence required for FDA approval of new or modified imaging modalities does not necessarily require demonstration of improved patient outcomes [13]. Moreover, this subtlety of FDA approval is not clear to most patients or many health care providers. After FDA approval, rapid diffusion of DM coincided with reimbursement, in accordance with the Benefits Improvement and Protection Act of 2000 [14]. Congress enacted DM reimbursement for Medicare beneficiaries,

with private insurers following suit shortly thereafter. Similar to DM, adjunct computer-aided detection (CAD) programs (which received FDA approval in 1998) obtained Medicare coverage in 2000 to assist radiologists in mammographic interpretation, despite limited evidence that CAD improved accuracy compared with routine mammography alone [15,16]. Subsequently, a 2007 study of > 400,000 mammograms from > 40 US facilities found overall reduced screening accuracy with CAD versus without CAD [17]. Other recent analyses suggest uncertainty regarding whether CAD has made any positive impact on patient outcomes [18,19]. Similar to DM, rapid diffusion of CAD was highly associated with Medicare coverage, with prevalence of CAD increasing from 4.8% in 2001 to 26.9% in 2003 [20]. In contrast to DM and CAD, DBT is currently not reimbursed by Medicare, and yet the technology continues to diffuse into community settings [21]. Therefore, financial remuneration from third-party payers, although critical for technological adoption, is not the sole driving force. Instead, device manufacturers are using direct-to-consumer marketing to target women who may be interested in paying out of pocket for potentially improved screening outcomes. DBT is being touted as “3-D mammography” in community settings, and practices are adopting the new technology to differentiate themselves from their regional competitors and gain a higher proportion of the available imaging market share. Other factors, including patient-driven legislative efforts, may influence diffusion. As of July 2013, nearly 40% of the US screening population resides in states with breast density reporting laws [22], with many more states likely to enact similar laws in the future. These laws require radiology practices to directly inform women with heterogeneously or extremely dense breasts by mammography that they are at increased risk for breast cancer and may benefit from supplementary screening [23]. The laws do not, however, specify the recommended modality for adjunct screening, and legislation for payment of supplemental screening varies widely by state. Thus, patients and providers must make challenging decisions with sparse data on whether supplemental screening may be beneficial and, if so, if patients can afford it. The legislation in Connecticut, Indiana, and Illinois mandates reimbursement for ultrasound screening among women with dense breasts. However, no other states with enacted laws have incorporated such a clause. Offering adjunct ultrasound to nearly half of the screening population (the proportion with dense breasts) would be logistically cumbersome with a dearth of qualified sonographers [24]. Although automated whole-breast ultrasound could eliminate the manpower shortage, the added workflow costs with regard to physical space, equipment, technologist time, and patient time still puts ultrasound at a disadvantage to DBT [25]. As it is currently used, adjunct DBT confers additional radiation

Lee, Lehman/Digital Breast Tomosynthesis 915

with a dose equivalent to that of standard screening DM [26,27]. However, in May 2013, the FDA approved software that would allow 2-D synthetic views to be reconstructed from DBT projections, eliminating the need for standard digital mammographic views and negating the increased radiation risks associated with DBT [28]. This FDA approval of software, similar to CAD programs for DM, will likely further fuel DBT’s adoption. As a consequence of legislation, there may be increasing demand among patients for supplemental screening. In response, individual radiology practices with the resources to implement DBT are taking advantage of the opportunity with the possibility of garnering a greater market share of patients. Practices could argue that, from the patient perspective, a small, initial out-of-pocket expense for DBT could prevent more expensive copayments and anxiety associated with unnecessary diagnostic evaluation of false-positive screening findings. PRACTICE MANAGEMENT IMPLICATIONS

The financial implications of acquiring and offering DBT are not trivial. At this time, there is no guaranteed third-party reimbursement. Some payers may reimburse adjunct DBT through an accessory Current Procedural Terminology code (76499, unlisted diagnostic radiographic procedure), but the additional reimbursement remains relatively low (averaging $50 above the standard DM reimbursement of $140). Newer digital mammographic units with DBT capability have a commercial price tag of about $750,000, which represents a large, upfront capital investment. With the current average outof-pocket cost for adjunct DBT at about $50, practices offering this new technology will make up only a portion of their initial capital investments in new equipment. Other associated costs for implementing DBT include additional space for and purchase of dedicated workstations to interpret DBT studies. Moreover, there will be required expansion of IT support for the substantially larger amounts of imaging data that must be archived. Of great concern among radiologists with regard to changing workflow is increased interpretation time. The Oslo population-based screening study demonstrated a statistically significant increase in radiologist interpretation time when adding DBT to routine digital mammographic screening studies, with a mean interpretation time of 90 seconds compared with 45 seconds for DM alone [10]. Other studies have also noted longer interpretation times regardless of whether DBT is performed as a standalone or an adjunct study [4,6,29,30]. Increased radiologist experience with DBT does not seem to significantly decrease interpretation times, suggesting that the increased number of images acquired and needing review is the limiting factor [6]. Although doubling interpretation time in the screening setting may negatively affect workflow, it is uncertain whether radiologists’ time can be saved by eliminating unnecessary diagnostic workup. Given that

a main advantage of DBT is a marked reduction in false-positives, the reduction in diagnostic volume in a practice can then be replaced by larger batched screening interpretations. Because diagnostic breast imaging is conventionally more time consuming, with relatively less financial support for physician time compared with screening, shifting practice volumes in favor of more screening and less diagnostic imaging could be advantageous from both financial and workflow management perspectives. Including DBT in diagnostic breast imaging increases the potential value of this emerging technology. Evidence suggests the obviation of additional diagnostic digital mammographic views after abnormal DBT screening, given that DBT is associated with similar or better diagnostic accuracy than digital spot compression views [31-33]. Moreover, diagnostic DBT projections after abnormal results on screening DM could eliminate the time required for radiologists and technologists to obtain adequate diagnostic digital mammographic views and potentially decrease unnecessary additional radiation exposure. With the ability to obtain 3-D diagnostic sweeps in a matter of seconds, DBT could eliminate the need for conventional digital mammographic diagnostic views, further increasing workflow efficiency. Finally, practices will face a large time investment to increase interaction and communication directly with patients. Already, practices in states with density reporting laws must inform patients about their increased cancer risk. Invariably, patients and referring physicians will turn to radiologists for guidance on whether supplemental screening is warranted. From both the ethical and medicolegal perspectives, practices will have to make a concerted effort to provide up-to-date information on the scientific evidence, fully disclosing the limits of the current data, and inform patients about associated longterm benefits, risks, and costs of performing additional screening. For example, in women with dense breasts and no other risk factors for breast cancer, there is currently no evidence to suggest that supplemental screening studies are associated with any long-term survival benefit [23]. Individual practices will have to determine which patient subpopulations in their communities will have access to adjunct screening, realizing that charging an additional fee may put emerging screening technologies out of the reach of patients who do not have the ability to pay out of pocket. POLICY IMPLICATIONS AND FUTURE DIRECTIONS

Both the advocacy push for supplemental screening and the adoption of DBT into clinical practice are occurring at the community level. Therefore, the evidence that will ultimately support policy decisions with regard to DBT best practices will likely also originate from the community level. Individual radiology practices, regional radiology alliances, and local and state radiology societies

916 Journal of the American College of Radiology/Vol. 10 No. 12 December 2013

must becoming willing partners in informing patients, referring physicians, third-party payers, and policymakers about their early experiences adopting DBT into practice. One of the key lessons learned from DM’s diffusion is that evidence synthesis and reporting did not keep pace with its real-time adoption. It is likely that DBT will diffuse broadly before a definitive, US-based clinical trial can be completed to inform stakeholders. Furthermore, by the time a multiyear trial is complete, the technology itself will likely have evolved through multiple iterations, with results no longer applicable to what is in current use [34]. Given that the breast imaging continuum is becoming increasingly complex, with various modalities used at different stages of care, the task of assigning direct causation of patient outcomes from DBT in the midst of multiple confounding variables will be difficult to ascertain through standard clinical trials [35]. Therefore, comparative effectiveness research methodology should play a major role in developing DBT policies and best practices moving forward. Comparative effectiveness research engages multiple stakeholders to gather already accessible data to estimate the effects of new imaging modalities [13]. It requires far less time and resources to complete than traditional trials and offers real-world evidence of the incremental effects of new imaging modalities on patient outcomes and costs. For DBT, radiologists should engage multiple stakeholders, including patients, providers, payers, and policymakers, to determine how best to collect observational data at the local practice setting in real time, aggregate data across practices, and harness data analyses into evidence that can affect patient care. The success of the National Oncologic PET Registry (NOPR) is one model for collecting evidence at the practice level with direct implications on clinical guidelines and financial reimbursement [36]. NOPR was created in 2006 to serve as an infrastructure for collecting data on PET’s effectiveness in changing the clinical management of patients with solid tumors. In return for participating in the registry, imaging facilities received Medicare reimbursement for PET scans. Not only did registry data demonstrate that PET had added benefit for monitoring patient response to therapies, they also helped solidify Medicare coverage for PET scans, with facilities no longer needing to report clinical data to NOPR. Because of the Mammography Quality Standards Act, breast imaging is uniquely positioned for cross-sectional analyses of real-world, observational data. In the future, practices may have to report DBT performance measures for auditing purposes, similar to screening mammography. These performance measures may include DBT’s role in decreasing a facility’s recall rate and biopsy rate, while potentially increasing cancer detection rate and positive predictive values. Requiring medical outcomes

audit reporting for DBT can be incentivized with contingent reimbursement similar to NOPR or to the Medicare Practice Quality Reporting System. These registries may show improved detection of early cancer and substantial decreases in false-positives resulting from DBT. Such evidence would make a strong case for creating permanent DBT reimbursement policies. Other types of comparative effectiveness research methodology including health economic modeling can be used to estimate the potential cost-effectiveness of implementing DBT from a number of perspectives. From the practice management perspective, it will be important to determine the cost-effectiveness of implementing DBT into screening programs. Factors such as clinical workflow and physician time may be important parameters to examine. Such an analysis did not occur for DM but would likely have shown increases in overall practice management efficiency over SFM. Just as important, radiologists should estimate potential savings from downstream health care costs avoided by implementing adjunct DBT. Such analyses will help demonstrate our added value in today’s costconscious health care environment. Finally, radiology practices must engage with local patients, patient advocacy groups, and referring physicians to better communicate and educate the public regarding supplemental screening technologies. We must be honest brokers of emerging imaging technologies, with clear presentation of potential benefits and limitations during early stages of evidence collection. Local radiology groups and societies should partner with patient advocacy groups pushing mandatory density reporting and argue for incorporating funds for concomitant data collection and reimbursement for supplemental screening technologies such as DBT into proposed legislation. To move early adoption toward more appropriate adoption of DBT, we must become engaged stakeholders to help guide future policies and best practices. TAKE-HOME POINTS

 Similar to the diffusion of DM in the past decade, the diffusion of DBT outpaces the evidence of its effectiveness.  The key driving forces for the rapid adoption of DBT are the advocacy-driven breast density legislation movement and increasing patient demand for supplemental screening.  Adjunct tomosynthesis, which can decrease falsepositive rates, has the potential to shift breast imaging practice away from diagnostic imaging workup and toward higher volumes of screening.  Radiology practices that are early adopters of tomosynthesis need to partner with other stakeholders to collect observational data prospectively at the local level to inform decisions on reimbursement and practice guidelines.

Lee, Lehman/Digital Breast Tomosynthesis 917

REFERENCES 1. Advances in breast imaging. Although mammography remains standard for breast cancer screening, several newer technologies are helping to finetune diagnosis. Harv Womens Health Watch 2010;17:1-3. 2. van Ravesteyn NT, Miglioretti DL, Stout NK, et al. Tipping the balance of benefits and harms to favor screening mammography starting at age 40 years: a comparative modeling study of risk. Ann Intern Med 2012;156:609-17. 3. Pisano ED, Gatsonis C, Hendrick E, et al. Diagnostic performance of digital versus film mammography for breast-cancer screening. N Engl J Med 2005;353:1773-83. 4. Wallis MG, Moa E, Zanca F, Leifland K, Danielsson M. Two-view and single-view tomosynthesis versus full-field digital mammography: highresolution X-ray imaging observer study. Radiology 2012;262:788-96. 5. Skaane P, Gullien R, Bjorndal H, Eben EB, Ekseth U, Haakenaasen U, et al. Digital breast tomosynthesis (DBT): initial experience in a clinical setting. Acta Radiol 2012;53:524-9. 6. Gur D, Abrams GS, Chough DM, et al. Digital breast tomosynthesis: observer performance study. AJR Am J Roentgenol 2009;193:586-91. 7. Michell MJ, Iqbal A, Wasan RK, et al. A comparison of the accuracy of film-screen mammography, full-field digital mammography, and digital breast tomosynthesis. Clin Radiol 2012;67:976-81. 8. Bernardi D, Ciatto S, Pellegrini M, et al. Prospective study of breast tomosynthesis as a triage to assessment in screening. Breast Cancer Res Treat 2012;133:267-71. 9. Svahn T, Andersson I, Chakraborty D, et al. The diagnostic accuracy of dual-view digital mammography, single-view breast tomosynthesis and a dual-view combination of breast tomosynthesis and digital mammography in a free-response observer performance study. Radiat Prot Dosimetry 2010;139:113-7. 10. Skaane P, Bandos AI, Gullien R, et al. Comparison of digital mammography alone and digital mammography plus tomosynthesis in a population-based screening program. Radiology 2013;267:47-56. 11. Ciatto S, Houssami N, Bernardi D, et al. Integration of 3D digital mammography with tomosynthesis for population breast-cancer screening (STORM): a prospective comparison study. Lancet Oncol 2013;14:583-9. 12. Tosteson AN, Stout NK, Fryback DG, et al. Cost-effectiveness of digital mammography breast cancer screening. Ann Intern Med 2008;148:1-10. 13. Gold LS, Klein G, Carr L, Kessler L, Sullivan SD. The emergence of diagnostic imaging technologies in breast cancer: discovery, regulatory approval, reimbursement, and adoption in clinical guidelines. Cancer Imaging 2012;12:13-24. 14. Centers for Medicare and Medicaid Services. Medicare, Medicaid, and SCHIP Benefits Improvement and Protection Act of 2000. Ruling No. 01-01. Available at: http://www.cms.gov/Regulations-and-Guidance/ Guidance/Rulings/downloads/cmsr0101.pdf. Accessed June 14, 2013. 15. Warren Burhenne LJ, Wood SA, D’Orsi CJ, et al. Potential contribution of computer-aided detection to the sensitivity of screening mammography. Radiology 2000;215:554-62. 16. Birdwell RL, Ikeda DM, O’Shaughnessy KF, Sickles EA. Mammographic characteristics of 115 missed cancers later detected with screening mammography and the potential utility of computer-aided detection. Radiology 2001;219:192-202. 17. Fenton JJ, Taplin SH, Carney PA, et al. Influence of computer-aided detection on performance of screening mammography. N Engl J Med 2007;356:1399-409.

18. Noble M, Bruening W, Uhl S, Schoelles K. Computer-aided detection mammography for breast cancer screening: systematic review and metaanalysis. Arch Gynecol Obstet 2009;279:881-90. 19. Taylor P, Potts HW. Computer aids and human second reading as interventions in screening mammography: two systematic reviews to compare effects on cancer detection and recall rate. Eur J Cancer 2008;44:798-807. 20. Fenton JJ, Foote SB, Green P, Baldwin LM. Diffusion of computeraided mammography after mandated Medicare coverage. Arch Intern Med 2010;170:987-9. 21. Szabo L. New 3-D mammograms have benefits, risks. USA Today. October 9, 2012. Available at: http://www.usatoday.com/story/news/ nation/2012/10/09/3d-mammogram-tomosynthesis/1615719/. Accessed August 19, 2013. 22. Are You Dense Advocacy. D.E.N.S.E. state efforts. Available at: http:// Areyoudenseadvocacy.org/dense/. Accessed June 16, 2013. 23. Lee CI, Bassett LW, Lehman CD. Breast density legislation and opportunities for patient-centered outcomes research. Radiology 2012;264:632-6. 24. Berg WA, Blume JD, Cormack JB, et al. Combined screening with ultrasound and mammography vs mammography alone in women at elevated risk of breast cancer. JAMA 2008;299:2151-63. 25. Kelly KM, Dean J, Lee SJ, Comulada WS. Breast cancer detection: radiologists’ performance using mammography with and without automated whole-breast ultrasound. Eur Radiol 2010;20:2557-64. 26. Poplack SP, Tosteson TD, Kogel CA, Nagy HM. Digital breast tomosynthesis: initial experience in 98 women with abnormal digital screening mammography. AJR Am J Roentgenol 2007;189:616-23. 27. Feng SS, Sechopoulos I. Clinical digital breast tomosynthesis system: dosimetric characterization. Radiology 2012;263:35-42. 28. Gur D, Zuley ML, Anello MI, et al. Dose reduction in digital breast tomosynthesis (DBT) screening using synthetically reconstructed projection images: an observer performance study. Acad Radiol 2012;19: 166-71. 29. Good WF, Abrams GS, Catullo VJ, et al. Digital breast tomosynthesis: a pilot observer study. AJR Am J Roentgenol 2008;190:865-9. 30. Zuley ML, Bandos AI, Abrams GS, et al. Time to diagnosis and performance levels during repeat interpretations of digital breast tomosynthesis: preliminary observations. Acad Radiol 2010;17:450-5. 31. Tagliafico A, Tagliafico G, Astengo D, et al. Mammographic density estimation: one-to-one comparison of digital mammography and digital breast tomosynthesis using fully automated software. Eur Radiol 2012;22:1265-70. 32. Tagliafico A, Astengo D, Cavagnetto F, et al. One-to-one comparison between digital spot compression view and digital breast tomosynthesis. Eur Radiol 2012;22:539-44. 33. Noroozian M, Hadjiiski L, Rahnama-Moghadam S, et al. Digital breast tomosynthesis is comparable to mammographic spot views for mass characterization. Radiology 2012;262:61-8. 34. Pearson SD, Knudsen AB, Scherer RW, Weissberg J, Gazelle GS. Assessing the comparative effectiveness of a diagnostic technology: CT colonography. Health Aff (Millwood) 2008;27:1503-14. 35. Pandharipande PV, Gazelle GS. Comparative effectiveness research: what it means for radiology. Radiology 2009;253:600-5. 36. Tunis S, Whicher D. The National Oncologic PET Registry: lessons learned for coverage with evidence development. J Am Coll Radiol 2009;6:360-5.