Chronic Kidney Disease Detection using Deep Learning: A Systematic Review

Abstract

Chronic Kidney Disease (CKD) is a serious and rapidly increasing health disorder that affects kidney function and may lead to kidney failure if not detected at an early stage. Accurate and timely diagnosis of CKD is important for reducing mortality rates and improving patient care. In recent years, deep learning techniques have emerged as powerful tools in medical diagnosis due to their capability to automatically learn complex patterns from healthcare data. This systematic review provides a comprehensive study of deep learning methods applied for CKD detection and prediction. Various approaches including Convolutional Neural Networks (CNN), Deep Neural Networks (DNN), Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM), and hybrid deep learning models are analyzed based on datasets, preprocessing techniques, performance metrics, and classification accuracy. The review also examines the advantages and limitations of these models in clinical applications. Most studies reported high accuracy, sensitivity, and specificity, demonstrating the effectiveness of deep learning in CKD diagnosis. However, issues such as limited datasets, data imbalance, overfitting, and lack of interpretability continue to affect model performance. This review highlights current research trends, challenges, and future directions for improving CKD detection systems using deep learning. Overall, the study concludes that deep learning-based approaches have significant potential to support healthcare professionals in early and reliable diagnosis of CKD.