Oraculum - Big Data study
Research type
Research Study
Full title
A retrospective observational epidemiology study to characterise the disease and treatment pathway in prostate cancer, based on information in electronic health records (EHRs) and using natural language processing (NLP) and artificial intelligence (AI).
IRAS ID
268482
Contact name
Noel Clarke
Contact email
Sponsor organisation
Johnson and Johnson
Clinicaltrials.gov Identifier
CANC 40227 , NIHR CRN Reference
Duration of Study in the UK
1 years, 0 months, 30 days
Research summary
Prostate Cancer (PCa) is the 2nd most common disease amongst males worldwide, and the 3rd leading cause for men’s cancer death in developed countries. In Europe, prostate cancer is the most commonly diagnosed cancer in men and accounts for approximately one-quarter of newly diagnosed cancer cases per year.
To address the need for additional data outside of clinical trials and given the rapid changes in PCa management over the past few years, the aim of this study is to use data in electronic health records: treated as ‘big data’ to describe the management of PCa patients in a more precise and accurate way.
Oncology, with its combination of clinical, technological, biological data and long disease course, provides an ideal scenario for applying big data analytics. Artificial intelligence (AI) is a computer science technique which can perform operations such as pattern recognition and language prediction, mimicking human neural networks, on a more complex scale.
EHRs contain a rich source of medical data and are therefore an ideal data set on which to apply AI techniques. Unstructured data within EHRs contains 80% of significant medical information, which is currently inaccessible using conventional real-world evidence methodology.
Using one specific branch of AI: Natural Language Processing applied to unstructured and structured data contained in EHRs, this study aims to precisely describe the prostate cancer disease and treatment pathway with a level of breath and granularity which has so far been achievable.
Having a more complete and up-to-date snapshot of the disease evolution, clinical characteristics and treatment pathways within PCa will improve our ability to understand the potential implications of diagnostic and therapeutic approaches.
ORACULUM is a European-led study, which will include at least four countries: UK, France, Germany and Spain. Spain is a pilot country and is currently ongoing. In addition, 60 additional clinical sites across Europe will be invited to participate. Eligible sites will have EHRs in place and the study will prioritise sites with EHRs of 2 years or more. The study population consists of all prostate cancer patients, from localised to late stage metastatic disease.
Janssen is partnering with MedSavana S.L, a digital health technology company that has developed an NLP platform. The first interim analysis will include 10 sites as a proof of concept, including at least 1 site from each country. The final analysis will include all sixty pre-planned sites.REC name
South Central - Oxford A Research Ethics Committee
REC reference
19/SC/0642
Date of REC Opinion
6 Jan 2020
REC opinion
Further Information Favourable Opinion