NEW YORK and PITTSBURGH, March 05, 2020 (GLOBE NEWSWIRE) — Owkin, a startup that deploys AI and Federated Learning technologies to augment medical research and enable scientific discoveries, announces a collaboration with the University of Pittsburgh. This pilot leverages the high-quality datasets and world-class medical research within Pitt’s Departments of Biomedical Informatics and Pathology, as well as Owkin’s pioneering technologies and research platform. Collaborations such as these have potential to advance clinical research and drug development.
Pitt researchers led by Michael Becich, MD, PhD, Associate Vice Chancellor for Informatics in the Health Sciences and Chairman and Distinguished University Professor of the Department of Biomedical Informatics (DBMI), will team up with Owkin to develop and validate prognostic machine learning models. The pilot project will then have the potential to expand into several key therapeutic areas for the University.
“The Pitt Department of Biomedical Informatics in partnership with the Department of Pathology is committed to improving biomedical research and clinical care through the innovative application of informatics and best practices in next generation data sharing. This collaboration with Owkin will expand our innovations in the computational pathology space,” Dr. Becich said. “Our currently funded projects explore areas such as the intersection of genomics and machine learning applied to histopathologic imaging (computational pathology) to broaden our understanding of the role of the tumor microenvironment for precision immune-oncology.”
This partnership makes it possible for Pitt to join the Owkin Loop, a federated network of US and European academic medical centers that collaborate with Owkin to generate new insights from high-quality, curated, research-grade, multi-modal patient data captured in clinical trials or research cohorts. Loop generated insights can inform pharmaceutical drug development strategy, from biomarker discovery to clinical trial design, and product differentiation. Owkin seeks to create a movement in medicine by establishing federated learning at the core of future research.
Federated learning technologies enable researchers in different institutions and different geographies to collaborate and train multicentric AI models on heterogeneous datasets, resulting in better predictive performance and higher generalizability. Data does not move, only the algorithms travel, thus protecting an institution’s data governance and privacy. Furthermore, Owkin’s data use is compliant with local ethical body consent processes and data compliance regulations such as HIPAA and GDPR.
“We’re thrilled to launch this project with Dr. Becich and his team at Pitt. The quality and size of the University’s research cohorts in combination with the DBMI’s mandate to bring together healthcare physicians and innovative academics to work on some of the most cutting-edge science, makes this collaboration a great opportunity to develop predictive AI models and to scale other research in the future. Owkin is proud to bring their expertise in machine learning technologies and data scientists to the table to foment new clinical insights,” Meriem Sefta, Owkin Head of Partnerships said.
Owkin is presenting their recent discoveries at HIMSS 2020 in Orlando, Florida on March 10th at 2:45 pm EDT. From the Innovation Live Stage, Meriem Sefta and Charles Maussion will explore opportunities to transform medical research with a Federated AI Learning platform. Rooted in institutional relationships that optimize high-quality research cohorts fit for AI, they will illustrate these breakthrough methods with a Nature Medicine published case study developed in close collaboration with one of our partners. Several other Owkin members will also be on hand to discuss the organization’s innovative research platform. Get in touch with Owkin to learn more or meet them at their booth during the conference.
Owkin specializes in artificial intelligence technologies applied to clinical research. It was Co-founded in 2016 by Thomas Clozel, a hematologist oncologist and researcher, and Gilles Wainrib, a computer science teacher-researcher at the École Normale Supérieure, and a Stanford University PostDoc. Owkin enables researchers to use data from health care or research activities to train interpretable machine learning models. These models allow better prediction of patient prognosis and response to treatment, developing new generations of biomarkers. In October 2019, Owkin published its breakthrough analysis of tumor biology using an interpretable deep-learning model, called MesoNet, in Nature Medicine.
Owkin is among the first companies to use Federated Learning technologies, a decentralised analysis approach that protects patient data by ensuring that it never leaves the hospital. For more information, visit www.owkin.com and follow @OWKINscience on Twitter.
About the University of Pittsburgh:
A nonsectarian, coeducational, state-related, public research university founded in 1787, the University of Pittsburgh (Pitt) is a member of the prestigious by-invitation-only Association of American Universities and internationally renowned as a leading center of learning and research in the arts, sciences, humanities, professions, and health sciences. Comprising a Pittsburgh campus, which is home to 16 undergraduate, graduate, and professional schools, and four Western Pennsylvania regional campuses, Pitt offers nearly 500 distinct degree programs and confers more than 8,500 degrees annually.
Pitt has ranked among the top 10 recipients of funding from the National Institutes of Health since 1998 and is ranked among the top 10 American research universities nationally in terms of total federal science and engineering research and development obligations. For more information, visit www.pitt.edu. For more information about Pitt’s Department of Biomedical Informatics, visit https://www.dbmi.pitt.edu/
CONTACT: Press Contacts: Owkin: Talia Lliteras – Tél. +33 7 87 21 81 90; email@example.com Pitt: Patrick McMahon – Tel. 412-624-4148; firstname.lastname@example.org