Research Of Glomerular Filtration Rate Measurement Method Based On Medical Image Data

This work is supported by Medical IT and Health Service Division Neusoft Medical Limited.Assessment of renal function plays an important role both on diagnosis and treatment of kidney disease. Nowadays, there is not a method that both provides the precise anatomical information of kidney and assesses rental function. With the development of functional imaging, it is possible to determine unilateral renal glomerular filtration rate (GFR) through medical imaging data. However, it needs further study. In recent years, some scholars both at home and abroad begin to research the GFR measurement method using medical image data. This method is not only time consuming, but not high repeatability. Therefore, the automatic segment method using computers is important.In this thesis, based on MRI and CT images, the main contents are to achieve both GFR measurement method and automatic segmentation algorithm of kidney, abdominal aortic and rental cortex. There are three main researches as follows:①Assessment and comparison GFR determination method using different compartment model.②Modify the GFR determination automatic method base on UroCARE (urinary computer-aided diagnosis, Neusoft Medical) platform using two-compartment model.③Complete the automatic segmentation of rental cortex and medulla.In this thesis, the mainly contents and conclusions are as follows:(1) According to two-compartment model theory, complete the calculation of GFR based on divided kidney into tubular and vascular.25CT data sets of patients,8CT data sets of model pig are assessed. GFR measurement based on two-compartment model through CT data are relative accurate.(2) According to three-compartment model theory, complete the calculation of GFR base on divide kidney into tubular, vascular and interstitial cells. When the third compartment (interstitial cells) is large, measured GFR using three-compartment model is better.(3) This study has completed improvement of automatic measurement of GFR, which includes automatically segmentation of kidney, abdominal aorta and the calculation process of GFR determination. The whole determination process achieves high efficiency.25CT data sets of patient are assessed. GFR determination modules can segment kidney and abdominal aortic quickly and calculate GFR accurately. So it has prospect of clinical application.(4) Design and complete the rental cortex and medulla segmentation algorithm. The whole algorithm only requires20seconds.15MRI data sets of model pigs’unilateral renal are assessed. It shows good result of automatically segmentation.