Swarm Learning for decentralized and confidential clinical machine learning
Stefanie Warnat‐Herresthal, Hartmut Schultze, Krishnaprasad Lingadahalli Shastry, Sathyanarayanan Manamohan, Saikat Mukherjee, Vishesh Garg, Ravi Sarveswara, Kristian Händler, Peter Pickkers, N. Ahmad Aziz, Sofia Ira Ktena, Florian Tran, Michael Bitzer, Stephan Ossowski, Nicolas Casadei, Christian Herr, Daniel Petersheim, Uta Behrends, Fabian Kern, Tobias Fehlmann, Philipp Schommers, Clara Lehmann, Max Augustin, Jan Rybniker, Janine Altmüller, Neha Mishra, Joana P. Bernardes, Benjamin Krämer, Lorenzo Bonaguro, Jonas Schulte-Schrepping, Elena De Domenico, Christian Siever, Michael Kraut, Milind Y. Desai, Bruno Monnet, Maria Saridaki, Charles Siegel, Anna Drews, Melanie Nuesch-Germano, Heidi Theis, Jan Heyckendorf, Stefan Schreiber, Sarah Kim-Hellmuth, COVID-19 Aachen Study (COVAS), Paul Balfanz, Thomas Eggermann, Peter Boor, Ralf Hausmann, Hannah Kuhn, Susanne Isfort, Julia Stingl, Günther Schmalzing, Christiane Kühl, Rainer Röhrig, Gernot Marx, Stefan Uhlig, Edgar Dahl, Dirk Müller‐Wieland, Michael Dreher, Nikolaus Marx, Jacob Nattermann, Dirk Skowasch, Ingo Kurth, Andreas Keller, Robert Bals, Peter Nürnberg, Olaf Rieß, Philip Rosenstiel, Mihai G. Netea, Fabian J. Theis, Sach Mukherjee, Michael Backes, Anna C. Aschenbrenner, Thomas Ulas, Deutsche COVID-19 Omics Initiative (DeCOI), Angel Angelov, Alexander Bartholomäus, Anke Becker, Daniela Bezdan, Conny Blumert, Ezio Bonifacio, Peer Bork, Boyke Bunk, Helmut Blum, Thomas Clavel, Maria Colomé‐Tatché, Markus Cornberg, Inti Alberto De La Rosa Velázquez, Andreas Diefenbach, Alexander Dilthey, Nicole Fischer, Konrad U. Förstner, Sören Franzenburg, Julia-Stefanie Frick, Gisela Gabernet, Julien Gagneur, Tina Ganzenmueller, Marie Gauder, Janina Geißert, Alexander Goesmann
Abstract
Abstract Fast and reliable detection of patients with severe and heterogeneous illnesses is a major goal of precision medicine 1,2 . Patients with leukaemia can be identified using machine learning on the basis of their blood transcriptomes 3 . However, there is an increasing divide between what is technically possible and what is allowed, because of privacy legislation 4,5 . Here, to facilitate the integration of any medical data from any data owner worldwide without violating privacy laws, we introduce Swarm Learning—a decentralized machine-learning approach that unites edge computing, blockchain-based peer-to-peer networking and coordination while maintaining confidentiality without the need for a central coordinator, thereby going beyond federated learning. To illustrate the feasibility of using Swarm Learning to develop disease classifiers using distributed data, we chose four use cases of heterogeneous diseases (COVID-19, tuberculosis, leukaemia and lung pathologies). With more than 16,400 blood transcriptomes derived from 127 clinical studies with non-uniform distributions of cases and controls and substantial study biases, as well as more than 95,000 chest X-ray images, we show that Swarm Learning classifiers outperform those developed at individual sites. In addition, Swarm Learning completely fulfils local confidentiality regulations by design. We believe that this approach will notably accelerate the introduction of precision medicine.