Predicting prostate cancer progression: protocol for a retrospective cohort study to identify prognostic factors for prostate cancer outcomes using routine primary care data

摘要: Introduction Prostate cancer is the most common in men UK, with nearly 40 000 diagnosed 2014; and it second cause of male cancer-related mortality. The clinical conundrum that live prostate rather than die from it, while existing treatments have significant associated morbidity. Recent studies shown very low mortality rates (1% after a median 10-year follow-up) no treatment-related reductions mortality, localised cancer. This study will identify prognostic factors progression to help differentiate aggressive more indolent tumours disease at diagnosis, so inform decision adopt conservative (active surveillance) or radical (surgery radiotherapy) management strategies. Methods analysis Clinical Practice Research Datalink (CPRD) contains 57 318 who were between 1 January 1987 31 December 2016. These be linked Office for National Statistics (ONS) Cancer Registration Analysis Service registry databases TNM stage, Gleason grade treatment data. Men diagnosis date prior lymph node distant metastases excluded. A priori determined potentially end point progression, measured baseline, participants followed through development death follow-up period (31 2016). Cox proportional hazards regression used estimate crude mutually adjusted HRs. Mortality risk predicted using flexible parametric survival models can accurately fit shape hazard function. Ethics dissemination protocol has approval Independent Scientific Advisory Committee UK Medicines Healthcare products Regulatory Agency Database (protocol 17_041). findings presented peer-reviewed journals local CPRD researcher meetings.