May 30, 2018 — U.S.-based Softek Illuminate and the entrepreneurial French firm Medexprim will be combining, distributing and supporting each other’s product offerings as a unified solution for hospital researchers interested in exploiting previously unstructured clinical and image databases. The two companies will be featuring the other’s solutions at the Society for Imaging Informatics in Medicine (SIIM) annual meeting, May 31-June 2 in National Harbor, Md.
Illuminate provides its InSight and PatientView solutions to assist radiology practices in unlocking the value of patient records stored across multiple silos of data to instantly deliver actionable intelligence.
Medexprim’s Radiomics Enabler is an Extract/Transform/Load (ETL) solution. The software automates the selection, extraction, pseudonymization/de-identification and secured routing of larger numbers of image sequences from a picture archiving and communication system (PACS) for secondary use in research.
Softek Illuminate CEO Matt McLenon said the partnership will allow his company to enrich its text mining suite with image extraction capabilities — a need with the exponential growth of radiomics and machine learning in radiology. Medexprim CEO Karine Seymour said the partnership gives Medexprim access to Illuminate’s installed base of major, mostly U.S. research hospitals that are very active in artificial intelligence (AI) projects in medical imaging. Additionally, Illuminate’s solutions complement Medexprim’s software by eliminating the latter’s customers’ challenge of mining through massive amounts of data to find the cohort of patients they wish to study.
The combined solutions from Illuminate and Medexprim allow researchers to instantly create patient groups through data-mining of unstructured data from a variety of systems (EMR, RIS, LIS), extract corresponding reports and imaging exams, curate and pseudonymize the results. The cleansed contextualized data can then be accessed by data scientists and other researchers interested in finding correlations between imaging features and other phenotypes and/or genotypes, as well as developing training data sets for machine learning algorithms.
Researchers will also be able to streamline the process of collecting data in clinical trials involving medical imaging, with faster, more qualitative results. Alerts on new events on a cohort of patients configured by Illuminate automatically triggers the extraction of corresponding imaging exams, while automated quality and control and pseudonymization is supplied by the Medexprim solution.