Artificial intelligence approaches to improve kidney care

Key advances A deep recurrent neural network model using data from electronic health records enables the prediction of inpatient episodes of acute kidney injury (AKI) with lead times of up to 48 hours5. Integrating intraoperative physiological signals into an AKI risk model that dynamically integrates preoperative and intraoperative data improves the prediction of postoperative AKI6. A convolutional deep learning model enables the noninvasive classification of chronic kidney disease stage and estimated glomerular filtration rate using kidney ultrasound images8. A convolutional neural network trained for multiclass segmentation enables automated analysis of transplant biopsy and nephrectomy samples9. Artificial intelligence is increasingly being used to improve diagnosis and prognostication for acute and chronic kidney diseases. Studies with this objective published in 2019 relied on a variety of available data sources, including electronic health records, intraoperative physiological signals, kidney ultrasound imaging, and digitized biopsy specimens.