Artificial Intelligence for the Diagnosis and Treatment of Diabetes Kidney Disease: a systematic review

Abrar, S. M.

Background: Diabetic nephropathy (DN) is a serious microvascular complication that affects 40% of diabetes patients. In the last decade, artificial intelligence (AI) has been widely used in both structured and unstructured clinical data to improve the treatment of patients/potential patients with DN. Methods: This systematic review aims to cover all applications of AI in the clinical use of DN or related topics. Studies were searched in four open-access databases (Pubmed, IEEE Xplore, DBLP Computer Science Bibliography, and ACM digital library). Finally, the author manually searched the reference list of included studies in the study for additional relevant articles. Results: Finally, a total of 24 original peers reviewed articles were included in this study. Through a manual data extraction, the summary of key information such as applied AI algorithm, main outcomes, performance evaluation etc. was taken. Then the included studies underwent a quality assessment criterion, assessing the reproducibility, generalizability etc. Most of the included studies revealed that the AI frameworks outperformed conventional statistical methods. A summary of the limitations, such as lack of data availability or external validation of the framework, in the included studies, was also included. Conclusion: The rapid advancement of the AI framework and the exponential data generation in healthcare can be utilized and applied in clinical practices. The aid of AI can be instrumental in the treatment of DN.