Tomosynthesis cuts interval breast cancers by a quarter in major screening trial

Digital breast tomosynthesis could prevent around a quarter of invasive interval breast cancers that appear between two routine screening examinations, according to the final follow-up analysis of the TOSYMA study, the largest comparative trial of the two screening methods conducted worldwide to date. In the first year after a normal screening result, the rate of invasive interval cancer diagnoses was 29 percent lower among women screened with tomosynthesis than among women screened with conventional digital mammography. The findings were published in The Lancet Oncology by a research team from the University of Münster and the University Hospital Münster in collaboration with other partners.

Interval cancers are tumours diagnosed between regular screening appointments, after a woman has received a normal result at her last examination. Experts regard these cancers as a key benchmark for the effectiveness of early detection, because interval tumours tend to be more aggressive and carry a poorer prognosis than cancers found during screening itself. A reduction in interval cancers therefore suggests that a screening technology is finding more cancers at the time of the examination, before they have the chance to grow and spread unnoticed.

Tomosynthesis works differently from standard mammography. Instead of a single two-dimensional X-ray image, several images are taken from different angles using low-dose X-rays. From these images, cross-sectional pictures of the breast with a thickness of up to one millimetre can be reconstructed. This three-dimensional reconstruction means that signs of a tumour are less likely to be obscured by overlying tissue, one of the main limitations of conventional mammography, in which dense glandular tissue can hide small lesions.

The TOSYMA trial enrolled almost 100,000 women aged 50 to 70 from 17 screening centres in the German federal states of North Rhine-Westphalia and Lower Saxony. Between 2018 and 2020, participants were randomly assigned to one of two groups. One group underwent tomosynthesis, supplemented by a two-dimensional image calculated from the tomographic images. The other group underwent digital mammography. Because the study was embedded within the nationwide, quality-assured German mammography screening programme, it took place under real-world conditions rather than in a specially constructed research setting.

To identify interval cancers, the research team checked participants’ data against the cancer registries of the two federal states, which cover more than 95 percent of new breast cancer cases. Suspected cases were reviewed using a multi-step process, and throughout the review staff at the cancer registries and reference centres, as well as the independent data safety committee, remained unaware of which screening group each woman had been assigned to. This blinding was designed to protect the comparison between the two methods from bias.

The first analysis of the trial, published in 2022, had already shown that tomosynthesis detects a significantly higher number of invasive breast cancers at screening: 7.1 versus 4.8 per 1,000 women screened. What remained unclear at that point was whether this higher detection rate translated into genuine clinical benefit. The new 24-month follow-up analysis addresses that question by comparing interval cancer rates in the two groups over the two years following screening.

In the first year after a normal screening result, 4.5 out of 10,000 women in the tomosynthesis group were diagnosed with invasive breast cancer, compared with 6.3 per 10,000 in the mammography group, a 29 percent difference. Over the full two-year follow-up period, the rate in the tomosynthesis group was 14.2 per 10,000 women, which was 24 percent lower than the 18.6 per 10,000 women in the mammography group. Unlike the first-year difference, however, the two-year difference did not reach statistical significance. The authors cite the impact of the pandemic, amongst other factors, as a possible reason why some 20,000 fewer women were recruited than originally planned, which limited the statistical power of the longer follow-up comparison.

The diagnostic effect was particularly pronounced among women who took part in a second screening round. Among this group, the interval cancer rate after two years was 13.5 per 10,000 women, whereas among women undergoing their first screening it was almost identical to the mammography group at 18.8 per 10,000 women. Furthermore, interval cancers in the tomosynthesis group were less likely to be locally advanced or to have spread to the axillary lymph nodes, characteristics that are generally regarded as indicators of effective early detection. The study also measured programme sensitivity, the proportion of all invasive tumours diagnosed within two years that were detected during the screening examination itself: 83.5 percent in the tomosynthesis group versus 72.3 percent in the mammography group. The reliability with which women without breast cancer were classified as normal was similar in both groups, indicating that the higher detection rate was not achieved at the cost of substantially more false alarms.

With around 75,000 new cases per year, breast cancer is by far the most common cancer among women in Germany, and the results are likely to inform the ongoing debate about which imaging technology should be used in population screening programmes. ‘Our study decisively strengthens the scientific basis for tomosynthesis in breast cancer screening,’ said Professor Stefanie Weigel, from the Department of Radiology at the University of Münster and the University Hospital Münster. ‘It expands and enhances the previously limited body of research on follow-up after tomosynthesis screening.’ Professor Walter Heindel, leader of the study from the same department, outlined the next step: the carcinomas detected during the regular follow-up screening of women in both groups using mammography will now be analysed, with the aim of determining whether tomosynthesis reduces advanced tumour stages more effectively than mammography in screening. As a randomised, controlled trial conducted within a real screening programme, the study offers some of the strongest evidence yet available on the comparative performance of the two methods, while the authors note that the reduced recruitment and the non-significant two-year comparison warrant continued evaluation of long-term outcomes.

Subject of Research: Comparison of digital breast tomosynthesis with digital mammography for breast cancer screening in the TOSYMA randomised trial

Article Title: Tomosynthesis detects breast cancer more frequently during screening

Article References: Tomosynthesis detects breast cancer more frequently during screening. Original publication (Bibliographic details unavailable.)

Image Credits: AI Generated

DOI: Not provided

Keywords: breast cancer, digital breast tomosynthesis, digital mammography, cancer screening, interval cancer, TOSYMA study, early detection, randomised controlled trial, The Lancet Oncology, programme sensitivity, screening programme, University of Münster