Radiomics Enabler®, an ETL (Extract-Transform-Load) for biomedical imaging in big-data projectsDownload PDF

11 Apr 2018 (modified: 16 May 2018)MIDL 2018 Conference SubmissionReaders: Everyone
  • Abstract: Clinical data warehouses are now routinely used by large hospitals. They allow researchers to pool medical data from millions of patients to create unified and harmonized patient cohorts for clinical trials and epidemiology studies. In bio-medical imaging, although there is a need to extract large amounts of data to help develop decision tools based on artificial intelligence, the identification, ex-traction and anonymization of relevant sequences is traditionally very time-consuming, as it is done patient by patient. We developed Radiomics Enabler® to streamline this process. We compared the extraction and anonymization pro-cess with and without Radiomics Enabler® on 118 imaging exams. We found that the operator’s time required using his PACS (Picture Archiving and Commu-nication System) workstation was 5 hours and 37 minutes. The time was reduced to only 17 minutes with Radiomics Enabler®. Radiomics Enabler® is available as open-source and can be integrated with any clinical data warehouse.
  • Keywords: ETL, DICOM, PACS
  • Author Affiliation: Medexprim, Centre Hospitalier Universitaire de Toulouse
4 Replies