Artificial Intelligence in Nephrology—State of the Art on Theoretical Background, Molecular Applications, and Clinical Interpretation

Artificial intelligence (AI) has transformed the clinical approach to analysis of large datasets, introducing the possibility of verifying long-term observations. AI tools ease the analysis of connections between multiple variable parameters and are particularly useful in the field of nephrology. These solutions enable the search for early diagnostic markers and predictors of renal function deterioration, both in acute and chronic conditions. Furthermore, AI techniques can be used as data mining tools, paving the way for future theories regarding the pathomechanisms of disease. Moreover, recently published papers focus on building models that facilitate decision-making, thus predicting renal involvement, its progression, and systemic complications. This review aims to demonstrate the multifunctionality of various AI methods from an omics perspective. To increase the power of argumentation, a mathematical background of each method is presented, followed by examples of molecular applications and anchorage in the nephrological clinical context. Our aim was to demonstrate the potential of AI tools in addressing diagnostic, prognostic, and therapeutic challenges, as well as to initiate the discussion on the pros and cons of future AI applications in nephrology.