Peter Nürnberg

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.

Lung cancer remains one of the leading causes of cancer-related death in developed countries. Although lung adenocarcinomas with EGFR mutations or EML4-ALK fusions respond to treatment by epidermal growth factor receptor (EGFR) and anaplastic lymphoma kinase (ALK) inhibition, respectively, squamous cell lung cancer currently lacks therapeutically exploitable genetic alterations. We conducted a systematic search in a set of 232 lung cancer specimens for genetic alterations that were therapeutically amenable and then performed high-resolution gene copy number analyses. We identified frequent and focal fibroblast growth factor receptor 1 (FGFR1) amplification in squamous cell lung cancer (n = 155), but not in other lung cancer subtypes, and, by fluorescence in situ hybridization, confirmed the presence of FGFR1 amplifications in an independent cohort of squamous cell lung cancer samples (22% of cases). Using cell-based screening with the FGFR inhibitor PD173074 in a large (n = 83) panel of lung cancer cell lines, we demonstrated that this compound inhibited growth and induced apoptosis specifically in those lung cancer cells carrying amplified FGFR1. We validated the FGFR1 dependence of FGFR1-amplified cell lines by FGFR1 knockdown and by ectopic expression of an FGFR1-resistant allele (FGFR1(V561M)), which rescued FGFR1-amplified cells from PD173074-mediated cytotoxicity. Finally, we showed that inhibition of FGFR1 with a small molecule led to significant tumor shrinkage in vivo. Thus, focal FGFR1 amplification is common in squamous cell lung cancer and associated with tumor growth and survival, suggesting that FGFR inhibitors may be a viable therapeutic option in this cohort of patients.