STORM, a new dimension for mammography screening

The STORM (Screening with Tomosynthesis or Standard Mammography) trial, from a team of Italian and Australian researchers, reports the world’s first completed population breast screening trial of integrated 2-dimensional (2D) and 3-dimensional (3D) mammography.1 Using breast tomosynthesis, a pseudo-3D derivative of digital mammography, in combination with 2D mammography, STORM researchers found that breast cancer detection was significantly increased when screen-reading was performed using integrated 2D and 3D mammography compared with 2D mammography alone (Box).1 The trial also showed that this screening method could significantly reduce false-positive results (Box). These data provide evidence that integrated 2D and 3D mammography would enhance breast screening performance measures, and are consistent with the interim findings of another population screening study conducted in Oslo, Norway, using the same mammographic technology.2

Should the present evidence on 3D mammography (tomosynthesis) be used now to introduce changes to breast screening practice in Australia? There are likely to be two divergent views: one, that existing evidence on detection measures is sufficient evidence for changing practice; and another, that additional evidence is needed to show that improved detection with integrated 2D and 3D mammography would translate into improved screening efficacy. It is important to place the emerging evidence on breast tomosynthesis into context, recognising that integrated 2D and 3D mammography breast screening has been examined in only two population-based screening trials (one of which has, to date, reported only an interim analysis), and that another trial is in progress in Malmö, Sweden, that will produce relevant evidence.3 Consistent evidence from these screening trials and ideally an additional large trial in an Australian screening environment would be valuable. Considering the broader evidence on breast screening, its benefits and harms, and in particular the possibility that detection of additional cancers could contribute to overdiagnosis, is also relevant when deciding on the type of evidence needed to recommend this method of screening.1,4

Mammography screening has had its fair share of criticism over several decades; however, pooled analyses of all the breast screening randomised controlled trials (RCTs) conclude that mammography reduces breast cancer deaths.5,6 Evidence of mammography screening benefit has also been shown in evaluations of population-based programs, including that evaluating the BreastScreen Australia program.5-7 There are however potential harms from mammography screening — specifically, false-positive results and the highly debated risk of overdiagnosis.4-6 Overdiagnosis refers to detection of cancers that would not have become clinically manifest during the woman’s lifetime were it not for screening.4-6 Breast screening detection measures alone do not equate to screening benefit, traditionally measured in terms of breast cancer mortality reduction. Because of uncertainties of the effect of increased breast cancer detection with integrated 2D and 3D mammography, relative to 2D mammography, on mortality benefit and on potential overdiagnosis, it cannot be assumed that the additional cancer detection observed with integrated 2D and 3D mammography will translate to a net incremental screening benefit.8 Therefore, to assess whether the increased cancer detection from integrated 2D and 3D mammography is likely to confer additional benefit, an RCT comparing integrated 2D and 3D mammography with standard 2D mammography that examines interval cancer rates as a surrogate end point for screening efficacy would provide critical evidence on this matter and would avoid the extended wait required to assess mortality outcomes.8 Demonstrating that integrated 2D and 3D mammography reduces interval cancers would show that this new breast screening technology detects cancers that would have biologically progressed as opposed to overdetecting indolent cancers. Measuring interval cancer rates does not require prolonged follow-up, and is routinely monitored in BreastScreen services as part of quality assurance. In addition, these RCT data could be used to assess cost-effectiveness and service delivery implications (such as the time burden from integrated 2D and 3D screen-reading) in the Australian context.

Australia, through its population-based BreastScreen program, and opportunities for collaborations between health services and research groups, is ideally positioned to take a lead role in executing such an RCT and, if results are favourable, to introduce integrated 2D and 3D mammography in a controlled environment with established processes for monitoring outcomes. Several approaches could be used to conduct trials, for example randomisation could be performed at the individual or at the service level, and would require staggered introduction of integrated 2D and 3D mammography. Pilot studies will be necessary to determine feasibility of integrated 2D and 3D mammography screening and to guide planning of larger trials in the Australian screening context. Such an approach would potentially see Australia take a lead role in providing high-quality evidence on whether or not a change from standard to integrated 2D and 3D mammography is likely to confer additional screening benefit to women, and would have the advantage of systematically introducing integrated 2D and 3D mammography into screening practice through formal evaluation with feedback into practice. It would also allow women attending population screening to access the new mammography technology by participating in screening trials. Adopting such an approach to integrated 2D and 3D mammography would accord with the McKeon report’s key recommendation to “embed research in the health system” and to “facilitate translation of research into evidence-based healthcare”.9 Without a prompt but carefully planned evaluation embedded in practice, we risk premature adoption of integrated 2D and 3D mammography based on relatively limited evidence of its effect, or unduly delayed adoption with loss of early benefit to Australian women and loss of an opportunity for Australia to lead in breast cancer translational trials.