Prospective randomized comparison of digital breast tomosynthesis plus synthesized images versus standard full-field digital mammography (TOSYMA)

• Applicant: Professor Dr. Walter Heindel
• Subject Area: Nuclear Medicine, Radiotherapy, Radiobiology
• Term: since 2016
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 286538777

The randomised controlled TOSYMA study examines the hypothesis that the combination of Digital Breast Tomosynthesis (DBT) and synthesised 2D mammograms leads to a clinically relevant increase in the screening detection rate of invasive breast cancer compared to the standard, digital 2D mammography.However, the expected increase in cancer detection caused by DBT could also be due to overdiagnosis. For this reason, it is important to assess whether the use of DBT not only increases cancer detection but also reduces the occurrence of interval cancers.To balance the potential benefit of DBT in terms of raised cancer detection against the potential for overdiagnosis, two primary endpoints have been defined in the study protocol: the detection rate of invasive breast cancers and the cumulative 24-month incidence of invasive interval cancers. A hierarchical test procedure will be applied to account for multiplicity, evaluating the detection rate in the first and the interval cancer rate in the second place of the sequence of hierarchically ordered hypotheses. The pre-defined order of hypotheses corresponds to the natural hierarchy in the medical context: the issue of a reduction in the interval cancer rate is only raised in the case of increased cancer detection. According to the pre-defined order of hypotheses and the primary objective of the study in the planning phase, the initial sample size calculation was based solely on the first primary endpoint (detection rate). However, as the investigation of interval cancers to assess the impact of potential overdiagnosis caused by tomosynthesis has gained increasing national and international attention, it seems most important to achieve a reasonable statistical power for the evaluation of both primary endpoints. This requires an increase of the initially calculated sample size of 80,000 to 120,000 women.The revised sample size calculation takes into account both, the detection rate and the interval cancer rate. It was carried out without knowledge of the data from the currently recruiting TOSYMA study, i.e. all planning assumptions are based on external data that do not belong to ongoing study. The assumptions regarding the first primary endpoint - on which the initial calculation was based - have been updated and revised in accordance with data that became accessible in the meantime. The multiple significance level of 5% is still maintained.The intended sample size increase offers the unique potential to evaluate both primary hypotheses and may help to provide currently universally lacking insights into the effectiveness of breast cancer detection with tomosynthesis. A sample size increase is much more efficient than designing a new confirmatory trial for the assessment of interval cancer rates.

• DFG Programme: Clinical Trials
• Co-Investigator: Professorin Dr. Stefanie Weigel