Assessment of proteinuria level in nephrology patients using a machine learning approach

Proteinuria represents an increase in the urinary excretion of proteins. It could also follow kidney transplantation and affects more than 40% of kidney transplant patients per year. It results from protein increases in their filtered load, due to alterations in the selectivity of the glomerular capillary wall, or from defects in their tubular uptake. Different parameters are associated with the various stages of proteinuria and therefore allow characterizing of its severity. For this purpose, the variation of proteinuria was evaluated by loading input two parameters: glycemia and the blood level of the m-Tor inhibitor. Through combination of data with different machine learning algorithms, the goal of this research work was to evaluate how blood glucose values and the use of immunosuppressive drugs can lead to prediction proteinuria classification in patients.