Mammographic surveillance in breast cancer patients aged 50 years or older

Design: Multi-centre, randomised, controlled, phase III trial of annual mammography versus 2 yearly for conservation surgery patients and 3 yearly for mastectomy patients. There will also be an observational cohort study of those eligible patients whom the specialist or patient opts for continued annual mammography or immediate discharge to the screening programme or stopping mammography altogether. An integrated feasibility study will assess the willingness for centres and patients to participate. Randomisation will be stratified by type of surgery (conservation, mastectomy), DCIS or invasive disease, age at randomisation ( 75yrs), hormone therapy at randomisation (continued, stopped, N/A as ER-ve). In addition mandatory information on known risk factors for recurrence/new primaries/ breast specific survival such as NPI (and individual components such as lymph node status, histological grade, invasive size), vascular invasion, margin involvement, ER status as well as treatment history, screen detected tumour and breast density will be collected. Setting: Breast cancer units in the United Kingdom Target population: Female patients previously treated surgically with curative intent for invasive or non invasive (DCIS) breast cancer in breast cancer units in the UK. Patients will be 50 years or over and 3 years post surgery, undergoing surveillance mammography and self assessment support in line with the cancer reform strategy. Health Technologies being assessed: Centres will utilise digital imaging in line with the National roll-out of digital screening mammography. Digital mammography offers advantages in terms of image quality, increased sensitivity in important subgroups (e.g. women with dense breasts) and reduced examination time. It is also reported to deliver significantly less radiation than conventional mammography. In addition to the mammogram eligible patients would be invited to respond to a brief questionnaire regarding their general health and treatment effects, delivered at their choice either, by a radiographer, at the time of mammogram (Cardiff model, Vaile 2006) or by your breast care nurse (Manchester Model, Beaver 2005) if in clinical follow-up. The questionnaire will also assess whether those patients who should be taking a hormone blocking drug such as Tamoxifen or Arimidex are receiving the correct prescription and advice about maintaining bone health. The General Practitioners of all patients will be aware of their recruitment and will be asked to advise the trial centre of any breast related consultations or referral to breast clinic during the period of the study. All patients with breast cancer, both invasive and DCIS will be given clear and concise information regarding their diagnosis and treatment and will be aware that if they get a new symptom they will be seen again urgently (within 2 weeks) in a specialist clinic (Cambridge model, Chapman 2007). Patients expressing concern regarding any aspect of the study will be offered more detailed follow-up or subsequent treatment. In addition the PPI component of the study will evaluate adequacy of current information offered as standard to patients. Measurement of cost and outcomes: The economic analysis will focus on (1) within trial cost effectiveness analyses, incremental cost effectiveness ratios will be presented using (i) disease specific survival and (ii) QALYs, (2) a lifetime decision analytical cost effectiveness model. Both analyses will be from the perspective of health and social care sector. Primary outcomes: 1. Disease specific survival 2. Cost effectiveness Secondary outcomes: 1.Recurrence (time to recurrence, type of recurrence and features of recurrence plus new primary in ipsi-lateral breast and new primary in contra-lateral breast) using questionnaire and data linkage 2.Number of referrals back to the hospital system using HES data linkage 3.Long term survival (20 years post surgery) using ONS flagging Outcomes for sub-studies: 1.Fear of recurrence (FoR)and Quality of Life using the ACS FoR and distress thermometer and EQ-5D-5L, respectively 2.Compliance with guidelines (bone health, adjuvant hormone blocking therapy) 3.Experience with the support provided for rapid assessment of new symptoms 4.Experience with follow-up strategies assessed by patient,specialist & nurse Sample size: Assuming that patients who have survived and are recurrence free at 3 years post surgery have a disease specific survival for the next 5 years of 89%, then recruiting 5000 patients (2500 per arm) will allow detection of non-inferiority with a 2.5% 1-sided significance and 90% power, defining non-inferiority as no worse than 3% below the standard arm. This sample size would also allow the detection of non-inferiority in the time to recurrence with 2.5% 1-sided significance and a 90% power. These calculations assume that patients who have survived and are recurrence free at 3 years post-surgery have a rate of recurrence after 5 years on study of 8% or less, non-inferiority is defined as no worse than 2% below the standard arm. This would also allow differences of 2% to be detected in the time to recurrence rates at 5 years with 90% at the 5% significance level. A Quality of Life sub-study of 600 patients (300 in each arm) would be sufficient to detect a relatively small standardised difference of 0.3 with 90% power and 5% 2-sided significance level. For example, this would equate to a 7.5 change in quality of life assuming a conservative standard deviation of 25. Statistical analysis: Kaplan-Meier survival curves will be constructed for the primary outcome of disease-specific survival and Cox proportional hazards models will be used to compare trial arms after adjustment for stratification variables as well as exploring important prognostic factors. All analyses will be carried out on an intention-to-treat basis. For the secondary outcome measures, appropriate analyses will be undertaken as specified in the research plan. Project Timetable: Start July 2013, 6 months set up of 30 centres, first patient randomised January 2014. Feasibility assessed at 24 months. Recruitment of 5000 patients over the next 42 months given 100 centres randomising in the first 24 months. Interim analysis May 2020 when all patients have 3 year follow-up with minimum of 5 year follow-up.