[Tailored Breast Screening Trial (TBST)]

mammography screening in premenopausal women is still the object of controversy and the cost and harms-benefit balance is lower than the one observed for women aged 50 years or more. The reasons are the lower screening sensitivity and the lower risk of occurrence of breast cancer at younger ages. For these reasons, an annual interval is suggested for this age group, and a lower positive predictive value of the recall rate is observed in screening practice. Harms of screening are false positive rates and overdiagnosis (that is an exceeding number of cases over the women's lifetime ascribable to early diagnosis). A tailored approach to screening could both contribute to a reduction of possible harms and reduce the costs of the service screening programmes, which started to offer screening mammography to 45- 49-year-old women also in Italy, according with the suggestions of the Italian Group for Mammography Screening. Higher breast density is considered a marker of risk and, at the same time, it brings about a masking effect that decreases the screening mammography sensitivity at younger ages.

the Tailored Breast Screening Trial (TBST) is a population-based, non-inferiority randomised trial aimed at evaluating the impact of a change in the screening protocol in a service screening practice. "Tailored" means that the use of a breast density classification allocates women to a longer interval, decreasing the number of screening rounds in the 45- 50-year age range. Density of the breast at the baseline is considered as an indicator of risk and also as a masking factor. The aim of this study is to assess the impact of a longer interval and the reduction of side effects for women allocated to the intervention group. In this study, any further intervention is offered to high-density women who are followed-up according to the usual care interval, since the study aimed at decreasing the screening burden.

45-year-old women resident in the screening centre catchment area will be invited to attend for mammography screening and will be asked for informed consent in order to be included in the study. After the enrolment, they will receive a high quality digital mammography; two views and breast density will be classified according with the BI-RADS classification. Women are randomly allocated either to an usual care group or to the intervention group. In the intervention group, women with a dense breast (3-4 categories in BI-RADS)will be invited again after 1 year, while the lower-density group in the intervention arm will be invited after 2 years. After the age of 50, all women will continue to be screened in the usual service screening programme. Density of the breast will be read by 2 readers, and controversies will be solved by a consensus. Allocation of women is blinded to the mammography readers.

the outcomes are: A. cumulative incidence of interval-cancer cases by intention to treat (ITT) grouping and by density group, aimed at assessing the non inferiority of screening performance; B. cumulative incidence of T2+/node-positive status breast cancer cases between arms and by protocol. Screening performance parameters will be evaluated at each screening round. Interim and outcome analysis are expected at 3 and 6 years average follow- up, respectively, starting from the beginning of the screening.

the non-inferiority limit is derived from the accepted level of interval cancers in women 50-69 years, which was considered acceptable by the European Community Guidelines. Assuming a 70%BI-RADS 1-2 at the baseline on the basis of the digital mammography experience, the estimated sample size with a power of 90% is 16,596 women per arm.

the background of the study, that uses data from cancer registries and screening centres, describes the most relevant characteristics of breast cancer occurring in the Italian regions where the study will be conducted. The description of stage distribution in the 40-54-year-old group and the coverage by spontaneous screening, as available by outpatient current data, showed the modification of stage at presentation, according with the implementation of service screening in women 50-54 years. In Emilia-Romagna and Piedmont Regions, service screening is already active for women aged 45 years and more, even though different invitation protocols were used. In Tuscany Region the extension of screening has already been planned for the next few years. Performance data are reported: they show the recall and the detection rates at first and at repeated screening rounds in Emilia-Romagna Region. In order to assess the feasibility and the interobserver reproducibility of breast density visual classification by BI-RADS, two testing sets of 80 digital mammograms for screened women aged 50-52 years were made available by the ISPO (Institute for oncological study and prevention) service screening tests. The test was performed by 11 radiologists, which are members of the working group. Concordance was moderate with a BI-RADS classification in 4 categories (kappa=0.46;95%CI 0.41-0.53), and substantial using 1-2 vs.3-4 (kappa=0.68;95%CI 0.59-0.79). The retest performed using the second mammography test did not change the results.

the TBST was authorised by the Ethical Committee in Florence (Tuscany Region, Central Italy) and the recruitment is still in progress through the invitation of the target population in the participating screening centres. In Florence and Veneto Region, the first round is almost completed and it confirmed the feasibility of the study and of the randomisation process.