Agreement for the analysis of clinical records based on artificial intelligence
Institut Curie has set up an electronic medical records system for more than 10 years, gathering more than 10 million texts for hundreds of thousands of patients. This unstructured database contains all the consultation, surgery and radiotherapy reports as well as the chemotherapy protocols administered to patients. The use of advanced technologies is necessary for the automatic and real-time analysis of these "Big Data" of exceptional richness. The analysis of these data could open the way to a better understanding of the development of the complex pathologies that cancers and to allow an improvement of the therapeutic decision-making, especially in the case of treatments whose response is very heterogeneous.
To meet this challenge, Institut Curie and OWKIN have signed a partnership agreement for the development of a new clinical case analysis software based on the most modern methods of artificial intelligence and natural language analysis. The objective is the development of a tool allowing doctors to consult the files of similar patients to refine the therapeutic decision making by predicting the toxicities and the response to the treatments with the greatest respect for the confidentiality of the data.
exploit the data potential accumulated over more than 15 years: extracting knowledge, predicting evolution to better cure and advance research
"This agreement was the first agreement at the Institut Curie with a start-up specializing in mass data analysis. With the arrival of a Data Director, the Institut Curie, through its Carnot Institute, is determined to position itself in this key area for tomorrow's personalized medicine", explains Amaury Martin, Director of the Technology transfer and Industrial Partnerships Department of Institut Curie and Director of the Carnot Curie Cancer Institute. For his part, Dr. Alain Livartowski, Director of Data at the Institut Curie hospital, welcomes this cooperation: "This partnership enables Institut Curie to exploit the data potential accumulated over more than 15 years: extracting knowledge, predicting evolution to better cure and advance research".
Owkin's advanced Machine Learning technologies allow the automatic analysis of unstructured data of several types, combining free text, images and biological data. The artificial intelligence algorithm developed by OWKIN could also allow the discovery of new biomarkers predictive of toxicity and response to treatments. It guarantees absolute respect for the confidentiality of patient data.
"The technologies of natural language and image analysis processing have grown without precedent in recent years, notably with the explosion of deep learning. This opens up completely new perspectives for bringing innovative tools to doctors at the Institut Curie and allowing them to exploit and interpret as much as possible all the massive data accumulated within the hospital for several years, "says Gilles Wainrib, President of OWKIN.