Personalized Breast Cancer Screening

Overview

The purpose of the study is to measure short-term safety and efficacy of personalized vs. standard mammography screening among 50-67-year aged women. The CE-marked risk model incorporates genetic data, family history, lifestyle/hormonal factors and mammographic density. Consenting women will be 1:1 randomized to a control group receiving no risk measurement and continuing their normal biennial mammography, while women in the intervention group will receive risk measurement and an ensuing risk stratified screening programme. Questionnaire information on life quality, breast cancer worry and anxiety will be collected at baseline and different timepoints later from both groups. The primary endpoint - the fraction of low risk women rejecting the recommended extension of screening interval from 2 to 4 years, will be measured 2 years and 4 years after inclusion. PRSONAL will be a success if this fraction is lower than 30%. Secondary outcomes, include quality of life, breast cancer worry and anxiety. Commitment from the target group is key for success, and interview studies followed by a questionnaire survey among women will feed into construction of a citizen directed web-based Risk Communication Tool. This tool will collect risk information, present the risk estimate and provide individual risk communication, while monitoring involvement, acceptance, and psychosocial consequences of personalized screening. The large volume of individuals undergoing screening, necessitates automated, but individualized interaction with the screened individuals. The tool will constitute such a platform. In total, 2200 women will be randomized 1:1 without blinding to a control group assigned to the standard screening program, and an intervention group, which will be offered a risk measurement and risk stratified screening accordingly. Women in the intervention group are stratified into four risk groups. Depending on the risk group the women will be offered a mammography every 1-4 years. The control group are assigned to the standard national screening program with biennial screening. The primary outcome of the study will be the proportion of women in the low risk group, who choose to have the next mammography within two years from the enrollment, indicating that the women will have rejected the proposed de-escalated screening intensity. Moreover, potential harms such as increased anxiety, worry or reduced quality of life will be measured via self-report questionnaires.

The screening program in Denmark has a very high quality, and overall Danish women\'s satisfaction with breast cancer screening is high. While breast cancer screening serves its purpose, its implementation in a population in which 7 of 8 never develop breast cancer also produces harms in terms of anxiety, hassle and loss of working time of the visit, just under 2% risk of false positive findings prompting unnecessary additional examinations. The one-size-fits-all approach by definition invariably leads to additional drawbacks: 1) women with unrecognized low risk are examined too frequently and thus get disproportionate more harm, 2) women with unrecognized high risk are not examined frequently enough, which leads to delayed detection more treatment and long term harms, 3) a very big fraction of the very scarce resource of trained mamma radiologists is currently used to evaluate normal screening mammograms. Simulations suggest that risk stratified breast cancer screening would detect more breast cancers at an early stage while reducing the number of unnecessary biopsies among healthy women. This trial aims to personalize breast cancer screening by using the CE approved breast cancer risk model, BOADICEA, which is the most comprehensive model currently available for breast cancer risk prediction. It is the only model that incorporates the most up to date polygenic risk score for breast cancer, based on 313 single nucleotide polymorphisms (SNP), as well as familial breast cancer history, reproductive history, lifestyle/hormonal risk factors and mammographic density, obtained from image analysis of the mammogram. The model will calculate the individual absolute 10-year risk of breast cancer. Notably, only the mathematical risk prediction of BOADICEA was validated, and not its value in the real world. Applying the model to women from Copenhagen produces a very wide risk distribution, which would be a valuable basis for screening decisions. The objective of the study is to measure short-term efficacy, acceptability, and psychological safety of personalized vs. standard mammography screening among 50-67-year aged women. The women will be randomized 1:1 to a control group (without blinding) assigned to the standard screening program, and an intervention group, which will be offered a risk measurement and risk stratified screening accordingly. The control group are assigned to the standard national screening program with biennial screening. The study will investigate the proportion of women in the low risk group, who choose to have the next mammography within two years from the enrollment, indicating that the women will have rejected the proposed de-escalated screening intensity. Moreover, potential harms such as increased anxiety, worry or reduced quality of life will be measured via self-report questionnaires. At baseline, data concerning risk of breast cancer will be collected. This includes questionnaires on family history of breast cancer, lifestyle, reproductive history. Breast density from the mammogram will be recorded and height and weight will be measured. Blood will be drawn from women in the intervention group. Germline DNA from leukocytes will be extracted from the blood test from the women included in the intervention group. The ensuing genetic analysis will consist of determining the genotypes of 313 common pre-specified single nucleotide polymorphisms (SNPs) of the genome with an array especially developed for PRSONAL. No other SNPs or genes will be examined without renewed consent from the women. Safety measures on quality of life, breast cancer worry and anxiety will be recorded continuously with questionnaires. Data management: Basic characteristics of the women, in- and exclusion criteria, relevant medical history, risk and safety measures will be entered in the electronic case report form (CRF) in a project database. Prospectively, all contacts, their nature and content, between the women and PRSONAL and register information will be stored. All necessary data processing agreements have been approved and signed. Personal information and blood samples or derivatives thereof will remain in Denmark. The general data protection regulation and the data protection act will be kept. Sample size assessment: If more than 30% of the women in the low-risk group will opt out of the de-escalated screening and have their mammography within 800 days from baseline, PRSONAL will not have achieved its goals, in terms of acceptance and economic sustainability. When calculating the necessary sample size according to the primary outcome, we apply the following assumptions: 95% of the women attending screening mammography will be eligible. 50% of those will participate in the randomization. Of those randomized to the intervention group, 10% will regret their decision before risk is communicated, and of the remaining, 46% will be classified as low risk women. To achieve 90% statistical power and significance level of 0.05 to detect a drop out frequency among low-risk women of 30% or below, we will need to randomize 2200 women, when stratifying analyses according to median age.