Distributed learning: Developing a predictive model based on data from multiple hospitals without data leaving the hospital - A real life proof of concept.

By A. Jochems, T. Deist, J. van Soest, M. Eble, P. Bulens, P. Coucke, W. Dries, Philippe Lambin, A. Dekker. 2016

186 citations recorded by the source corpus.

PURPOSE One of the major hurdles in enabling personalized medicine is obtaining sufficient patient data to feed into predictive models. Combining data originating from multiple hospitals is difficult because of ethical, legal, political, and administrative barriers associated with data sharing. In order to avoid these issues, a distributed learning approach can be used. Distributed learning is defined as learning from data without the data leaving the hospital. PATIENTS AND METHODS Clinical data from 287 lung cancer patients, treated with curative intent with chemoradiation (CRT) or radiotherapy (RT) alone were collected from and stored in 5 different medical institutes (123 patients at MAASTRO (Netherlands, Dutch), 24 at Jessa (Belgium, Dutch), 34 at Liege (Belgium, Dutch and French), 48 at Aachen (Germany, German) and 58 at Eindhoven (Netherlands, Dutch)). A Bayesian network model is adapted for distributed learning (watch the animation: http://youtu.be/nQpqMIuHyOk). The model predicts dyspnea, which is a common side effect after radiotherapy treatment of lung cancer. RESULTS We show that it is possible to use the distributed learning approach to train a Bayesian network model on patient data originating from multiple hospitals without these data leaving the individual hospital. The AUC of the model is 0.61 (95%CI, 0.51-0.70) on a 5-fold cross-validation and ranges from 0.59 to 0.71 on external validation sets. CONCLUSION Distributed learning can allow the learning of predictive models on data originating from multiple hospitals while avoiding many of the data sharing barriers. Furthermore, the distributed learning approach can be used to extract and employ knowledge from routine patient data from multiple hospitals while being compliant to the various national and European privacy laws.

Subfields: Artificial Intelligence, Pulmonary and Respiratory Medicine

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

Citations in the corpus

Deep Leakage from Gradients (2019, methodology): “This scheme is named as collaborative learning and widely used when the training set contains private information [18, 20].”
Machine Learning in Radiation Oncology: Opportunities, Requirements, and Needs (2018, background): “One potential solution to the challenges associated with multi-institutional data sharing is the use of distributed learning; the group at Maastricht University led by Philippe Lambin has been pioneering this approach (68, 69).”
Inference Attacks Against Collaborative Learning (2018, methodology): “ticipants between 4 and 30 to match the deployment scenarios and applications proposed for collaborative learning, e.g., hospitals or biomedical research institutions training on private medical data [29,30]. This is similar to prior work [25], which was evaluated on MNIST with 2 participants and fa”
Distributed Gradient Descent Algorithm Robust to an Arbitrary Number of Byzantine Attackers (2019, background): “Third, in certain scenarios, data is naturally collected at different locations, and it is too costly to move all data to a centralized location [19].”
Privacy Risks of General-Purpose Language Models (2020, background): “…the answer is afﬁrmative, then the future applications of state-of-the-art NLP tools in compelling learning paradigms like collaborative or federated learning [36], [37], [46] can be largely threatened and restricted, especially on privacy-critical domains including healthcare, genomics and ﬁnance.”
Developing and Validating a Survival Prediction Model for NSCLC Patients Through Distributed Learning Across 3 Countries (2017, unknown): “To realize distributed learning for this study, we have used a formerly adapted method (38, 39).”