Intended for healthcare professionals

Feature Breast Cancer Screening

3D mammography is on the upswing in the US, as experts argue about its value

BMJ 2019; 366 doi: (Published 17 July 2019) Cite this as: BMJ 2019;366:l4506
  1. Jeanne Lenzer, associate editor
  1. The BMJ
  1. jeanne.lenzer{at}

While some doctors are happily recommending digital breast tomosynthesis, Jeanne Lenzer reports that the value of adding an extra dimension to mammography remains decidedly unproven

Dr Sanderson (not her real name) stood half naked, about to undergo routine screening mammography at Brigham and Women’s Hospital in Boston, when a technician asked if she’d like to “upgrade” to a 3D mammogram. The physician told The BMJ that she was shocked to be put on the spot with virtually no information about the new and presumably superior technique. She was even more disturbed when the technician couldn’t answer her questions: what are the test’s sensitivity and specificity? What about radiation exposure? Later, Sanderson heard from other women who had been offered the new mammogram at the point of service, a form of “upselling.” “Surely, patients shouldn’t be put in the position of making a decision when they are vulnerable and without the opportunity to receive balanced information,” Sanderson says. “It doesn’t seem to meet basic standards of informed consent.”

The potential benefits and harms of 3D mammography, also called digital breast tomosynthesis (DBT), are the subject of a fiery controversy among experts12—but concerns about the test are virtually unknown to the women who are encouraged to have it. If they are given any information at all, it is generally derived from marketing materials that proclaim the technology as “lifesaving,” “a great step forward,” and “cutting edge.”34 Celebrity spokespeople, including Grammy award winning singer Sheryl Crow, amplify the message. Crow tells women that “early detection saved my life” and urges women to request Hologic’s 3D Genius system because it “detects cancer 15 months earlier than any other machine.”5 Crow is a paid spokesperson for Hologic.6

Uptake of the new technology has been based largely on superior detection rates in women with dense breasts and the reported test characteristics of superior sensitivity and specificity over traditional 2D mammography.7

But whether improved detection rates translate into clinical benefit is not clear. In 2016, citing a lack of clinical outcomes data, the United States Preventive Services Task Force issued an “I” or inconclusive grade to DBT—even for women with dense breasts. The task force concluded that “the current evidence is insufficient to assess the balance of benefits and harms of adjunctive screening for breast cancer” in a woman with an “otherwise negative screening mammogram.”89

A recent analysis of insured women showed wide variation around the US, with greater use of DBT in wealthier areas.10 An accompanying commentary noted that if DBT was eventually found to be helpful, the initial, skewed uptake could worsen socioeconomic disparities in breast cancer care.11 If DBT was found on balance to be harmful, it would be the wealthier, fully insured people who were harmed.

Due to uncertainties about DBT, the US National Cancer Institute in conjunction with the Canadian Cancer Trials Group launched the tomosynthesis mammographic imaging screening trial (TMIST) in 2017. TMIST is a randomised controlled trial that will enrol 165 000 women at 100 centres. The primary outcome is whether DBT reduces the incidence of advanced cancers compared with 2D mammography.12 The trial is slated to be completed in 2030.

Parts of the picture

DBT was granted a CE Mark in 2008 for use in the European Union. In 2011 the US Food and Drug Administration approved its use. The approvals were based on the promise of improved sensitivity and specificity over 2D mammography.21314 An evidence review by the US Agency for Healthcare Research and Quality found that sensitivity rose from 0.54 (95% confidence interval 0.42 to 0.65) with 2D mammography to 0.85 (0.74 to 0.92) with DBT. But some reports show DBT having superior specificity with only marginal or no increase in sensitivity1516 and others have found no increase in specificity compared with 2D mammography.17

Worta McCaskill-Stevens, chief of the Division of Cancer Prevention of the National Cancer Institute, told The BMJ that the evidence for DBT “is inconsistent” and that most of the studies were “single institution and retrospective with design problems.” Another limitation, she said, is that “many of the trials have been [conducted] by the vendor” or were conducted in European countries that relied on registries—not randomised controlled trials.

Michael Baum, emeritus professor of surgery at King’s College and University College London, who specialises in breast cancer screening and treatment, told The BMJ that the risk of overdiagnosis with DBT is substantial and likely to cause more women to receive unnecessary treatment for lesions such as ductal carcinoma in situ and small invasive cancers that are not destined to cause symptoms or death. Treating these women with surgery, chemotherapy, and radiation, he says, could cause far more harm than good.

But some proponents of DBT say that its benefits are so clear that TMIST isn’t even needed. Daniel Kopans, emeritus professor of radiology at Harvard Medical School and founder of the Breast Imaging Division at Massachusetts General Hospital, told The BMJ that the superiority of DBT was a “no brainer” and that TMIST “is an enormous and potentially dangerous waste of money.” Citing studies that show DBT has superior sensitivity and specificity compared with 2D mammography—a situation he calls a rare “win-win” as increases in test sensitivity generally come with a loss of specificity—he says that more women receiving 2D mammography will die from undetected cancers that could have been picked up by DBT.

Kopans dismisses the concerns of sceptics who say that DBT could cause a rise in harm due to overdiagnosis, saying that estimates of overdiagnosis are inflated and based on “scientific nonsense.” Kopans has received payments from manufacturers of DBT machines, including Siemens, Hologic, and GE, and has received royalties from several patents on various breast imaging and DBT devices. Kopans accuses critics of mammography screening of wanting to “deny” women access to save money.

PR campaigns trump caution

Journalists have generally reported industry claims uncritically.34 In 2016, the Telegraph published an article touting DBT as a possible solution to the 3500 tumours that are currently missed on NHS tests each year.6 The article was based on unpublished preliminary data showing that “current standard tests are not sensitive enough to pick up around one in six cases.” The study was completed in July 2018 but has not been published as of June 2019.

The success of promotional information is unmistakable, despite uncertainty about the evidence—an uncertainty that challenges true informed consent, even if facilities took the time to lay out what evidence there is. By 2011, DBT was commercially available in Africa, Asia, the Middle East, and South America. FDA data show that 61% of mammography centres in the US now offer DBT.18 Uptake has been far lower in the EU, where only 12% of centres use DBT,19 and Baum says there has been little uptake in the UK. Its use in the US has been expanded from women with dense breasts to a screening tool for all women. On 10 April 2019, the American Society of Breast Surgeons issued a consensus statement recommending that all women start routine mammography screening at age 40, with 3D mammography as the “preferred modality.”20

Radiologists caution that the operational constraints of DBT might stress institutional resources. Tomosynthesis units carry a steep price tag: even cheaper models cost about between $400 000 (£320 000; €350 000) and $430 000, whereas traditional 2D units cost between $80 000 and $100 000.21 Lag time to reporting is longer for DBT because it takes radiologists 1.5 to 2.5 times longer to interpret the 200 to 400 images of DBT than to interpret the two images of traditional 2D mammograms, which could require “adjustments to workflow and the addition of staff.”21 Storage capacity might also be strained, as DBT typically generates 450 megabytes per study but can reach three gigabytes, potentially causing systems to lock up and impede workflow.21

Radiation exposure from DBT is 38% higher on average (range 0-75%) than 2D mammography.22 Women with dense breasts, for whom DBT is most often recommended, risk greater radiation exposure as dense breast tissue absorbs substantially more radiation than fatty tissue.23 Finally, insurance companies might not reimburse for DBT, which has led manufacturers to fund lay organisations like Are You Dense and Are You Dense Advocacy, which encourage women to request DBT and help activists to lobby state and federal legislators to require insurers to cover the test.2425 Coverage is required in 36 US states, and Are You Dense has a global strategy to increase uptake of the test.


  • Competing interests: I have read and understood BMJ policy on declaration of interests and have no relevant interests to declare.

  • Provenance and peer review: Commissioned; not externally peer reviewed.


View Abstract