Study of obstructed kidney diagnosis using compartmental modelling & compare with the actual clinical evaluation

Abstract

The kidney is an organ in our body that filters blood to get rid of waste products, maintain the level of electrolytes in our body, stimulate the production of red blood cells and control our blood pressures. If a kidney is malfunctioning, waste products will accumulate and it will start to show signs of blockage. In order to detect the severity of blockage, renography technique is often used. However, there is no accurate technique and procedure used in clinical setting. In response to this issue, this project has the objective to search for a non-invasive method of analysing the renal obstruction and to develop a standard method of determining the severity of kidney blockage. In compartmental modelling analysis, the behaviour of the tracer was represented from the input into the kidney to the outflow from the kidney. The novel index developed U/V2, which represented normalized volumetric flow rate with volume, was used as the indicator in predicting condition of kidney obstruction by using support vector machine (SVM) and random forest (RF) classifier. The results were compared with actual clinical prediction by a certified nuclear medicine doctor and the results matched to a large extent.