Federated Learning of Electronic Health Records Improves Mortality Prediction in Patients Hospitalized with COVID-19

Machine learning (ML) models require large datasets which may be siloed across different healthcare institutions. Using federated learning, a ML technique that avoids locally aggregating raw clinical data across multiple institutions, we predict mortality within seven days in hospitalized COVID-19 patients. Patient data was collected from Electronic Health Records (EHRs) from five hospitals within the Mount Sinai Health System (MSHS). Logistic Regression with L1 regularization (LASSO) and Multilayer Perceptron (MLP) models were trained using local data at each site, a pooled model with combined data from all five sites, and a federated model that only shared parameters with a central aggregator. Both the federated LASSO and federated MLP models performed better than their local model counterparts at four hospitals. The federated MLP model also outperformed the federated LASSO model at all hospitals. Federated learning shows promise in COVID-19 EHR data to develop robust predictive models without compromising patient privacy.

Comparison of 2.6- and 1.4-mm imaging probes for confocal laser endomicroscopy of the urinary tract.

INTRODUCTION: Probe-based confocal laser endomicroscopy (pCLE) is an emerging technology for dynamic, in vivo imaging of the urinary tract with micron-scale resolution. We conducted a comparative analysis of pCLE with a 2.6-mm probe and a 1.4-mm probe that is compatible with flexible endoscopes. MATERIALS AND METHODS: Sixty-seven patients scheduled for bladder tumor resection were recruited. pCLE imaging was performed using 2.6- and 1.4-mm probes. Image quality with the different probes was examined and further compared with standard histopathology. RESULTS: Images with the 2.6-mm probe have better resolution of cell morphology. The 1.4-mm probe has a wider field of view and better view of microarchitecture. While image quality with the 2.6-mm probe is superior, the 1.4-mm probe is compatible with flexible cystoscopy and maneuverability is maintained, enabling imaging of areas of the bladder that were previously challenging to access with the larger probe. CONCLUSIONS: The optical specifications of the 2.6-mm probe are more suitable for distinguishing urinary tract histopathology. Further design optimization to improve resolution and additional validation of the diagnostic accuracy of the smaller probe are needed to help extend application of pCLE for optical biopsy of the upper and lower urinary tract.