Systematic Review of Medical Imaging and Clinical Data in CKD Prediction Using Deep Neural Networks

The global healthcare problem of chronic kidney disease (CKD) requires accurate predictions in the early stages to achieve appropriate treatment and management strategies. Deep learning through deep neural networks (DNNs) has received rising interest for CKD prediction systems because it effectively combines medical images with clinical information. The review investigates the implementation of deep neural networks for predicting CKD by uniting medical imaging approaches with clinical database information. Medical imaging techniques using ultrasound as well as CT scans and MRI, together with clinical datasets, generate important information regarding the initial development stages of CKD. The enormous computing ability of DNNs makes them perfect for detecting intricate patterns that standard evaluation techniques typically overlook. The article investigates multiple DNN structures predicting CKD, system performance levels, obstacles, and possible future enhancements.