Prostate cancer is the most common cancer among men in the UK. Although survival is high compared to other cancers common in men, prospects for older patients and those with more advanced disease are poor. The treatment of such patients has been challenging, with few effective options available and many patients becoming non-responsive to therapy such that their condition gets worse.
Janssen is interested in understanding the prostate cancer treatment pathway in England, analysing both primary and secondary care data to describe patient characteristics, the progression of their illness over time, the treatments they receive and their survival. Additionally, by combining these data with information from trials, researchers will be able to establish how well new treatments may perform if made available in England.
This research will provide valuable insights into prostate cancer care in England and the potential benefits to patients of new treatments for prostate cancer, particularly for prostate cancer patients at high risk of progression to more serious disease.
The study will provide insight into the clinical management and health outcomes of prostate cancer (PC) patients (ICD-10 C61) in England using cancer registrations from the National Cancer Registration Dataset (NCRD). Patients will be linked to data from the Clinical Practice Research Datalink (CPRD), Systemic Anti-Cancer Therapy dataset (SACT), Radiotherapy Dataset (RTDS), and Hospital Episode Statistics (HES) to map their progression through the course of disease.
Two trial-like cohorts will be created, consistent with the selection criteria for two PC trials:
* ATLAS (NCT02531516, EudraCT2015-003007-38)(1,2): a randomized, double-blind, placebo-controlled phase 3 study of apalutamide in subjects with high-risk, localised (LPC) or locally advanced prostate cancer (LAPC) treated with primary radiation therapy.
* PROTEUS (NCT03767244, EudraCT2018-001746-34)(3,4): a randomized, double-blind, placebo-controlled phase 3 study of apalutamide in subjects with high-risk, LPC or LAPC prostate cancer who are candidates for radical prostatectomy.
The baseline characteristics of patients in these cohorts will be described at an aggregate level, plus their front-line therapy, longitudinal disease assessment, salvage therapy following biochemical recurrence (BCR), and time-to-event outcomes. In addition to these aggregate outputs, a subset of non-identifiable row-level data items analysed by HDI will be forwarded securely to Janssen-Cilag Limited to support three further analyses related to the development of health technology assessment (HTA) dossiers for the aforementioned trials: (i) the extrapolation of long-term survival outcomes observed within the two trial-like cohorts, (ii) an informed fits analysis to reduce uncertainty around separate extrapolations of time-to-event data for trial participants, and (iii) two indirect treatment comparisons. These will be described in more detail within later sections.
Together, these insights will establish a detailed picture of the treatment pathway and patient outcomes of two trial-like populations, aiding the optimisation of existing treatments and providing comparative effectiveness data for novel therapies undergoing trial.
A series of revised time-to-event outputs will be reported for this study, as summarised in Appendix 1 and bullet-pointed below.
* Overall survival (OS): defined from index to all-cause death or censoring.
* Progression-free survival (PFS): defined from index to the earliest of all-cause death, movement to a subsequent disease state, or censoring.
* Metastatic-free survival (MFS): defined from index to the earliest of all-cause death, movement to a metastatic disease state, or censoring.
* Event-free survival (EFS), ATLAS: defined from index to the earliest of all-cause death, PSA failure (2ng/ml increase in PSA over nadir), direct movement to a metastatic disease state, or censoring.
* Event-free survival (EFS), PROTEUS: defined from index to the earliest of all-cause death, biochemical failure (BCF), direct movement to a metastatic disease state, or censoring.
For the PROTEUS-like cohort, one of the failure events for the EFS outcome is BCF. BCF will be defined per the following criteria, with the date of BCF being equal to the final applicable PSA reading:
* Had two pairs of consecutive PSA readings dated within the disease state, with each pair of consecutive readings spanning an interval of at least one week.
* A rise in PSA between the consecutive readings of both pairs of PSA measures (one to two, and three to four).
* The final reading (reading four) yielded an absolute PSA of ≥0.2 ng/mL.
Rules for the capture of movements between disease states are described later in the application, illustrated in Appendix 3, and with definitions tabulated in Appendix 2.
As a descriptive study, there are no a priori hypotheses.
As noted elsewhere in the submission, it is acknowledged that longitudinal measures of PSA will be fundamental for the capture of disease dynamics relevant to the some of the outcomes of interest. These include movements to subsequent disease states (such as are relevant for the estimation of PFS; see Appendix 2 for criteria for movements between disease states), or the capture of BCF within the EFS measure for the PROTEUS-like cohort. Based on NICE recommendations (i.e., NG131), it is hoped that PSA tests will have been repeated every three to six months within the population under study. However, it is expected that the frequency and completeness within the administrative datasets may fall short of this ideal. Per the planned feasibility exercise, outcomes dependent on these repeated measures of PSA will not proceed to analysis if PSA measures are deemed too sparsely or unreliably recorded for key dimensions of the disease pathway to be captured for the cohort under study.
Maximiliaan Nievaart - Chief Investigator - Janssen Pharmaceuticals UK
Craig Knott - Corresponding Applicant - Health Data Insight CIC
Alessandro Dos Santos - Collaborator - Janssen Pharmaceuticals UK
Amy Zalin-Miller - Collaborator - Health Data Insight CIC
Bhavini Patel - Collaborator - Janssen Pharmaceuticals UK
Simran Gill - Collaborator - Janssen Pharmaceuticals UK
Sophie Jose - Collaborator - Health Data Insight CIC
HES Admitted Patient Care;HES Diagnostic Imaging Dataset;HES Outpatient;NCRAS Cancer Registration Data;NCRAS National Radiotherapy Dataset (RTDS) data;NCRAS Systemic Anti-Cancer Treatment (SACT) data;ONS Death Registration Data