por
John R. Fischer, Senior Reporter | May 30, 2018
Medexprim and Softek Illuminate are
creating a unified solution to simplify
data mining in medical imaging
for U.S. and EU researchers
The full potential of unstructured information within clinical and imaging databases may soon be accessible to researchers on both sides of the Atlantic thanks to a partnership between U.S.-based Softek Illuminate and France’s Medexprim.
Combining Softek Illuminate’s InSight and PatientView platforms with Medexprim’s Radiomics Enabler, the two are seeking to create a uniform solution for the extraction and application of information that will further development of imaging biomarkers and the ability of U.S. and EU researchers to match eligible patients with clinical trials using artificial intelligence.
“Hospitals have millions of medical imaging exams and associated data that are very useful for researchers. However, the data they need must be found, cleaned, contextualized, and annotated,” Karine Seymour, CEO of Medexprim, told HCB News. “Our combined technologies will enable researchers to get all the data they require to advance radiomics research and artificial intelligence in medical imaging.”
Conventional PACS and EMR systems do not possess the research functionalities needed to extract enough data for meaningful understandings and findings, forcing many researchers to mine through vast amounts of information to find and match eligible patients with clinical trials.
Utilizing InSight and PatientView, users can apply unprecedented data mining capabilities to create patient groups from unstructured data derived from EMRs, RIS and LIS systems, extracting corresponding reports and imaging exams from across multiple silos.
Radiologists can then apply the Radiomics Enabler to automate the selection, extraction, pseudonymization, and secure routing of larger numbers of image sequences from a PACS for secondary research use. Its performance is based on a predefined set of rules.
By preparing and utilizing data to determine clinical trial matches, Seymour says the combined technologies will streamline the data collection process, leading to faster and higher quality results that will help in determining correlations between imaging features and other phenotypes and genotypes, as well as in the development of training data sets for machine learning algorithms.
“All that the hospital will need to do is use our combined solution to extract existing images of a potential cohort and process them to identify the patients that match,” she said.
The two will feature one another’s solutions at the Society for Imaging Informatics in Medicine (SIIM) Annual Meeting this week in National Harbor, Maryland.
Back to HCB News
