Back to search

BEHANDLING-God og treffsikker diagnostikk, behandling og rehabilitering

Digital breast tomosynthesis, the future screening tool for breast cancer?

Alternative title: Tredimensjonal digital mammografi - fremtidens screeningmetode for brystkreft?

Awarded: NOK 12.2 mill.

Digital breast tomosynthesis (DBT) is a new technique in mammographic screening for breast cancer. In September 2017, the Norwegian Institute of Public Health published a report stating that there was currently insufficient evidence to support implementation of DBT as a screening tool in Norway. This conclusion is in line with recommendations from the European Commission Initiative on Breast Cancer (EU) and from the International Agency for Research on Cancer (IARC). The randomized controlled trial, the Tomosynthesis Trial in Bergen (the To-Be trial), started January 14, 2016 and finished December 31, 2017, after one screening round. The main aim was to compare breast cancers detected using DBT versus DM, while the secondary aim was to compare early performance measures. The trial was performed as a part of BreastScreen Norway. Of the approximately 33,000 women who attended screening in Bergen during the study period, 89% consented to participate in the trial. These women were randomized to screening with either DBT (n=14 734, the study arm) or DM (14 719, the control arm). Screening was performed with the standard procedures in BreastScreen Norway. Preliminary results showed a longer average reading time for DBT (71 seconds) than DM (41 seconds) and fewer cases discussed at consensus for DBT (6.4%) versus DM (7.4%). The time used at consensus was longer for DBT than for DM (doi: 10.1007/s00330-018-5690-x). The main results from the To-Be trial were published in Lancet Oncology (doi: 10.1016/S1470-2045(19)30161-5) in June 2019. In the study arm, 95 women (0.66 %) were diagnosed with breast cancer, compared to 87 women (0.61 %) in the control arm. The recall rate was 3.1 % in the study arm, compared to 4.0 % in the control arm. This difference was statistically significant (p<0.0001). There were minimal differences in histopathological tumor characteristics. The results were somewhat surprising, considering that the majority of previous studies have shown a considerable increase in the cancer detection rate and that DBT detects the small and likely slowly growing tumors. The cause of different results might be related to the study design, but also the radiologists' experience in screen-reading DBT. The results from the To-Be trial were also investigated for different categories of mammographic density (doi: 10.1148/radiol.2020201150). The study found that women in low-density categories yielded lower recall rates for DBT than for DM. However, the adjusted relative risk of recall and screen-detected breast cancer increased by increasing density for DBT, but not for DM, suggesting that screening with DBT is less useful for women with high mammographic density. Analyses of compression force and compression pressure showed that women screened with DBT received lower compression force and pressure compared to the women screened with DM. Radiation dose was low for both techniques, and somewhat lower for DM versus DBT (doi: 10.1177/0284185119863989). Mammographic features among women recalled for further assessment following DBT versus DM showed that asymmetry was the most common feature for both methods, but more common for DM (38.9%) versus DBT (24.3%) (doi: 10.1016/j.ejrad.2021.109753). Spiculated mass was the most common feature for breast cancer after screening with DBT, while calcification was most frequent after DM. The study concluded that the distribution of mammographic features differed among recalled women after screening with DBT versus DM. Economic costs of DBT and DM were compared, and after including the costs of equipment, examination, and reading time, the cost of DBT was estimated to be 8.5 Euros more per screened woman than with DM (doi: 10.1007/s10198-019-01094-7). After adjusting for costs of recall, the estimate was somewhat lower. The study found significantly higher costs of DBT than DM, but the costs were lower than previous cost analyses on the use of DBT. A retrospective review was performed on the mammograms from screening and interval cancer from the To-Be study by radiologists at the breast center at Haukeland University Hospital. External radiologists will also perform a review. DM mammograms have been analyzed using artificial intelligence (AI), and when the algorithm for DBT is finished, these mammograms will also be analyzed with AI. We followed the women who participated in the trial for two years. They were offered screening with DBT in the next screening round in the follow-up study, To-Be 2, to compare the effects of DBT after screening with DBT vs DM. No statistically significant differences in rates of interval cancer, breast cancer, or histopathological tumor characteristics between the two groups were observed (doi: 10.1148/radiol.2021203936). The trial is registered at ClinicalTrials.gov. Information about the trial is published on the web pages of Cancer Registry of Norway.

To-Be studien var den første randomiserte kontrollerte studien som sammenlignet tidligindikatorer for tomosyntese (DBT) og vanlig digital mammografi (DM) i verden. Studien ble publisert i Lancet Oncology og har fått stor oppmerksomhet internasjonalt. Våre funn, ingen signifikant økning i deteksjon av brystkreft har skapt oppsikt og vi vurderer det som en tungtveiende årsak til at EU i dag konkluderer med at DBT er minst like bra som DM, og ikke bedre (https://healthcare-quality.jrc.ec.europa.eu/). Kostnadsanalysene støtter konklusjonen. Oppfølgende studier på intervallkreft viste heller ingen forskjell i deteksjon i de to armene, noe som er med på å underbygge dagens retningslinjer. I ettertid av studien er det opprettet en gruppe med radiologer og forskere fra Europa og USA som har regelmessige møter bruk av randomiserte kontrollerte studier innen radiologi diskuteres. Studiedesignet og våre erfaringer fra To-Be blir brukt som utgangspunkt for design av andre randomiserte studier.

Digital breast tomosynthesis (DBT) is a new screening and diagnostic technology for breast cancer. Studies from the U.S. have shown that screening with DBT results in a lower recall rate and a higher detection rate compared with digital mammography (DM). Whether a higher detection rate will lead to less false negative screening exams (less interval cancer) and thus a lower mortality, or if it is of influence for other aspects of screening is not known as of today. The economic aspects of DBT are only reported from opportunistic screening in the U.S. In other words, the number of studies on DBT as performed in organized screening programs in Europe and Norway is limited. More studies on early performance measures, including prognostic and predictive tumor characteristics, radiation doses, and the economic aspects are needed to obtain sufficient evidence to assess whether DBT can be recommended as a screening tool in Norway and Europe. This trial is aimed at testing synthetic 2D + DBT (tomo) versus DM as screening tools in women aged 50-69, who are invited to participate in the Norwegian Breast Cancer Screening Program at Danmarksplass in Bergen. A four months pilot is aimed at testing feasibility, get experiences with, and learning how to read the DBT, and to test practicalities of the project. The trial will start January 1, 2006 and run as a part of the screening program, for two years. Two years follow-up of the participating women will ensure complete data on breast cancer detected before and in the next screening round, in addition to treatment received among women with breast cancer. The effect of tomo will be evaluated by comparing the achieved early performance measures and economical aspects with those achieved by DM. The comparison of the performance measures will be performed as a part of a PhD while a researcher will perform analyses related to the economic aspects.

Funding scheme:

BEHANDLING-God og treffsikker diagnostikk, behandling og rehabilitering