Moving to Individualized Screening: Personalizing Breast Care
Screening Will Become More Personalized
Irène Dietschi | Jun 14, 2017
MRI, ultrasound, tomosynthesis – there are multiple options to examine the breast besides mammography. At the 7th Siemens Breast Care Day at the European Congress of Radiology 2017 (ECR) it was commonly agreed that screening will become more personalized. One of the most promising tools is digital breast tomosynthesis.
Seven lessons learned from the research presented at Breast Care Day 2017
Breast health and breast care are burning issues which are of great interest for radiology professionals: That was the most obvious result of the 7th Breast Care Day at ECR 2017, a symposium organized by Siemens Healthineers in cooperation with Bayer. Highly recognized professionals from Europe and Australia explored new ground in advancing the diagnostics and treatment of breast cancer.
With the goal of raising awareness of new and improving diagnostic methods for breast cancer, Siemens Healthineers initiated the “Breast Care Day” in 2011. Six years later it has become a tradition, and for many ECR participants it is the primary event at this meeting since it hosts a program full of the latest insights in the multimodality diagnostics of breast cancer. The combination of daily routine reports, new studies, the outlook for the future and open discussions has attracted more than 7,000 participants over the last seven years. In addition, many radiologists worldwide followed the symposia online.
At the seventh event, one message which all speakers shared was that breast care must become even more personalized, not only when it comes to treatment, but much earlier, at the screening stage. Density, for example, is an issue, requiring an individualized approach in screening programs. One of the key questions was whether 3D digital breast tomosynthesis (DBT) is ready to replace 2D mammography in mammo screening in the future. All speakers agreed that the role of breast MRI is part of an ongoing discussion on how to improve breast cancer detection and treatment.
1. Preoperative breast MRI: More help than harm
Francesco Sardanelli, MD, Professor at the University of Milan, Italy presented the first results from the MIPA study1, a large multicenter study started in 2012 in which by July 2016 about 5,000 patients were enrolled. MIPA is a prospective, observational study, which aims at evaluating the effect of preoperative breast MRI. It was partly prompted by the “Choosing Wisely” campaign of the American Society of Breast Surgeons, which recommends to “not routinely order breast MRI in new breast cancer patients,”2 arguing that routine use of MRI can lead to overdiagnosis, overtreatment and increased mastectomy rates.
Sardanelli called Choosing Wisely a very smart initiative. And what is the significance of MRI if mastectomy is already planned? And in cases where conservative treatment is planned, does preoperative MRI alter the actual surgery? MIPA was designed to provide answers to these questions, and the first resultspresented by Sardanelli were indeed enlightening: Of 1,000 breasts conservatively treated after MRI, surgery was unchanged in 733 cases. The others had either more extensive surgery (143) or less extensive surgery (128). Sardanelli believes that those results clearly demonstrate the potential of MRI for personalized medicine. He is convinced that MRI is useful for patients because it allows for the tailoring of conservative treatment. The actual mastectomy rates after preoperative MRI differed only slightly compared to the numbers of planned mastectomies, regardless whether they were based on mammography and ultrasound (185 out of 1,201 cases) or diagnostic MRI (245 out of 1,224 cases). These results show that mastectomies in the planning prompt MRI – used as a confirmation tool, not vice versa.
2. Contrast-enhanced Breast MRI: No reason to stop
Concerns about the safety of gadolinium-based contrast agents (GBCA) are widespread in the global imaging community, and was an intensively discussed topic at ECR. Jörg Barkhausen, MD, professor at the University of Lübeck, Germany investigated its impact on breast MRI. Barkhausen explained that contrast-enhanced MRI in clinical breast imaging emerged more than 30 years ago, providing excellent results for the detection and characterization of breast lesions. However, in 2006 GBCA were suspected to cause nephrogenic systemic fibrosis. And in late 2013 Kanda and colleagues3 discovered signal alterations in different brain areas, which they ascribed to previous administrations of GBCA. Succeeding studies showed that gadolinium concentration in the brain does indeed correlate with repetitive injections of gadolinium. In Barkhausen’s view, the results of these studies must be taken seriously, even though no clinically relevant adverse events have yet been associated with gadolinium presence in the brain.
What does this mean for daily clinical routine of breast MRI? Barkhausen argued that for the majority of patients the impact is not significant, because they hardly undergo more than one or two MRI examinations. However, the situation is quite different for high-risk patients such as those burdened with BRCA mutation, which is associated with early-onset of breast cancer. These patients need multiple MRI’s. But even for this particular group Barkhausen recommended not to stop performing contrast-enhanced MRI, arguing that there are more benefits than risks and that cancers might be missed without MRI. Nonetheless he advised certain precautions, e.g. to switch from linear GBCA to macrocyclic GBCA – a type of gadolinium which reduces the level of residual signal intensity in the brain. Barkhausen also recommended to routinely get informed consent from the patients and to be open about the gadolinium issue.
3. Consider breast density
An extensive part of the symposium was focused on the topic of breast density. Luis Javier Pina, MD, professor at the University of Navarra, Pamplona in Spain called breast density “the radiologist’s problem child”. He referred to the well-known fact that dense breasts reduce the sensitivity of mammography by about 50 percent. The superimposed dense tissue can mask the lesion on 2D mammography. Moreover, density itself is a marker of risk to develop breast cancer. In Pina’s opinion, screening programs should consistently consider the density of the breast. He argued that over the past years, more objective technologies have been developed to classify the density patterns. These technologies could be useful to select the patients who have higher risk of developing breast cancer. For this group of patients, Pina claimed, biennial mammography alone is no longer a good option.
Pina’s views were backed by his Swedish colleague Dr. Hanna Sartor, MD from Lund University in Malmö, Sweden. She suggested that future screening programs must be personalized and individualized, taking breast density into account. Sartor also proposed that future workflows in assessing breast density could well rely on software measurements instead of radiologists’ assessment. In her presentation she highlighted the potential of the so-called volumetric breast density analysis (VBDA), a technique that has been studied extensively. Sartor judged it a proven tool to estimate density in mammography and tomosynthesis.
4. Does ultrasound or MRI help if breasts are dense?
PD Michael Golatta, MD, from the University of Heidelberg, Germany examined the role of ultrasound in breast screening. Ultrasound has a reputation of being especially helpful when patients have dense breasts. Austria, for example, even adopted ultrasound into its screening program. The outcome of this public health measure remains to be seen. Previous research in this field yielded ambivalent results, as Golatta described: A 2008 study by the American radiologist Wendie Berg4, MD examined combined screening with ultrasound and mammography vs. mammography alone in women at elevated risk of breast cancer. The results published in JAMA showed that adding a single screening ultrasound to mammography would yield an additional 6.1 to 7.2 cancers per 1,000 high-risk women, but it would also substantially increase the number of false positives. In other words: The gain in sensitivity is counterbalanced by a loss of specificity. A review published by John Scheel, MD and colleagues in 20155, where screening ultrasound as an adjunct to mammography in women with dense breast was examined, and revealed a high variability in the biopsy rate. Golatta concluded that conventional ultrasound improves breast cancer screening, but with certain weaknesses: It is time-consuming and cost-intensive, it has a limited field of view, it requires high level of skill and experience, it is hard to standardize, and it elevates biopsies. However, the acquisition of ultrasound data could be automated and thus standardized in the future. Golatta noted that at the University of Heidelberg a number of studies assessing the diagnostic value of automated breast volume scanning (ABVS) have been conducted6,7, and ABVS showed a comparable diagnostic performance to hand-held ultrasound. Golatta concluded that ABVS is an effective supplemental tool for mammography in breast cancer detection. A bigger screening study is still pending, and certain technical details ask for improvement.
Could additional screening with MRI provide a solution for women with dense breasts? This was the question asked by Carla Van Gils, MD, a professor from the University of Utrecht. It is also the basic question of DENSE8, a large, randomized controlled on-going trial in the Netherlands which Van Gils presented in Vienna. In the past years, a number of studies have been published which showed that indeed more cancers are detected if women with dense breasts have an MRI examination after mammography. In the United States, so-called breast density notification laws have been put into effect in 27 states. They are intended to inform women who have undergone mammography about the risks posed by breast density, and to help them decide on further action. Even so, MRI is not included in screening recommendations for women with dense breasts. As Carla Van Gils observed, MRI may be of value for supplemental screening, but needs further evidence in large randomized trials, with several screening rounds and relevant health outcomes. DENSE has been designed to fill this gap, Van Gils explained. The researchers want to find out the number of extra tumors detected with MRI and, more importantly, they want to see if the interval cancer rate is reduced compared to usual screening. To estimate mammographic density, a fully automatic and validated method is used. In 2018 all participants will have undergone a 2-year follow-up. The primary outcome is a difference in interval cancer rates between the two arms, the best proxy for a difference in breast cancer mortality, Van Gils noted.
5. Digital breast tomosynthesis: The most promising method
What Carla Van Gils said about MRI also holds true for digital breast tomosynthesis (DBT): More evidence is needed for its effectiveness, despite the enthusiasm the new technology evokes for its ability to increase cancer detection and reduce false-positive rates. In her presentation, Professor Sylvia Heywang-Köbrunner, MD, of Breast Diagnostics Center in Munich, Germany observed that tomosynthesis is not quite ready yet to replace 2D mammography for screening. Certain information is still lacking, particularly with regard to its interval cancer performance.
However, Heywang-Köbrunner also asserted that DBT is the most promising tool there is. First systematic reviews of screening trials and of mostly retrospective data from the United States confirm that it is clearly superior to digital mammography, allowing significantly improved sensitivity. In Europe, tomosynthesis is being examined in prospective studies. The Italian STORM study9,10, for example, has yielded an increased detection rate of 40 percent in the DBT only screening arm (without additional digital mammography).
Despite these promising figures Heywang-Köbrunner prompted that evaluation of follow-up rounds remains essential, and that a number of issues need further study, including whether DBT is finding clinically important cancers. Moreover, DBT at the present stage is faced with a number of logistic problems: It requires longer reading time – two to four times longer than with digital mammography –, and it entails huge amounts of digital storage capacities.
Nevertheless, Heywang-Köbrunner concluded that DBT is the better modality compared to digital mammography, and is thus more promising for mass screening than ultrasound or MRI.
6. The future of digital breast tomosynthesis
Kristina Lång, MD of Lund University presented the Swedish Malmö breast tomosynthesis screening trial11, a study which is getting a lot of attention in the scientific community. Lång thinks that DBT is more ready than not as a general screening method, pointing out that breast tomosynthesis increases the cancer detection rate in screening considerably. The Malmö trial is a prospective population-based single-arm study including randomly invited women 40 to 74 years old eligible for the screening program in the city of Malmö, Sweden. The women underwent one-view breast tomosynthesis and two-view digital mammography. The images were read and scored separately in a blinded double-reading procedure. Interim results of 7,500 women showed a significant increase in cancer detection rate, from 6.3 to 8.9 out of 1,000. The recall rate also increased significantly – from 2.6 percent to 3.8 percent – an elevation Lång called reasonable.
Despite her enthusiasm, Kristina Lång confirmed Heywang-Köbrunner’s appraisal that interval cancer rates need further analyses. And like her German colleague she mentioned the logistic challenges of DBT. Instead of a human approach to address the reading time problem she proposed a high-tech solution: machine learning. In future radiology – after the interval cancer rates have been analysed – Lång suggested that tomosynthesis in screening could be used with the aid of artificial intelligence.
7. Practical aspects: Standards needed
Tomosynthesis may be a very promising tool for breast imaging. However, for clinicians the technology raises practical problems: There are now several DBT systems on the market, many of which have applied differing technologies to acquire the data necessary for tomosynthesis image reconstruction. The various techniques all have advantages and some potential limitations, but altogether they produce one main obstacle, as Wayne Lemish, MD from Melbourne, Australia explained: These differences could potentially produce different clinical outcomes and the lack of uniformity may make the comparisons between clinical trials difficult. In the interest of patients it would ease the radiologists’ life if producers found ways to collaborate and integrate their innovations, Lemish said. It was an appeal shared by many of the clinicians present.
And another aspect became more than clear at the end of Breast Care Day 2017: In the future, it won’t do anymore to screen every woman and every woman’s breast the same way. The days of mass mammography will eventually be over. The future of breast diagnostics is personalized, multimodality imaging.
3Kanda T, Radiology 2014
4Berg, W.A., et al., Combined screening with ultrasound and mammography vs mammography
alone in women at elevated risk of breast cancer. JAMA, 2008. 299(18): p. 2151-63.
5Scheel JRet al., Screening ultrasound as an adjunct to mammography in women with mammographically dense breasts. Am J Obstet Gynecol. 2015 Jan; 212(1):9-17.
6Golatta M, et al. Interobserver reliability of automated breast volume scanner (ABVS) interpretation and agreement of ABVS findings with hand held breast ultrasound (HHUS), mammography and pathology results. Eur J Radiol (2013)
7Golatta M et. al. Evaluation of an automated breast 3D-ultrasound system by comparing it with hand-held ultrasound (HHUS) and mammography. Arch Gynecol Obstet. (2015)
The statements by Siemens customers described herein are based on results that were achieved in the customer's unique setting. Since there is no "typical" hospital and many variables exist (e.g., hospital size, case mix, level of IT adoption), there can be no guarantee that other customers will achieve the same results.