Application of Distance Regularized Chan-Vese Method for Kidney Segmentation in Renal Scintigraphy

Abstract

The glomerular filtration rate (GFR) is a crucial index to measure renal function. In clinical practice, the Gates method is often used to obtain GFR and to evaluate renal function, but the results are easily affected by image quality and physicians' experience. When using the Gates method to calculate GFR, it is required to delineate the region of interest (ROI) of the kidney in dynamic renal scintigraphy firstly. In clinical environment, the ROIs are often delineated manually according to clinical experience. The manual process is subjective, complicated and low repeatable. To obtain the ROI of kidney accurately, we propose a distance regularization Chan-Vese (DRCV) method. This method effectively restrains the degradation of the symbol distance function and reduces the computational complexity in the process of segmentation. This fully automated ROI extraction method is proposed to avoid the errors caused by manual intervention in ROI mapping, to improve the accuracy of ROI extraction, and to reduce the time consumed in the whole detection process. In clinical experiments, the data of 213 subjects who underwent 99mTc-DTPA renal dynamic imaging were analyzed. Compared with previous ROI extraction methods, the proposed method reduces the computational time of segmentation by 30%. The results show that DRCV has great application prospects in accurately estimating GFR in renal dynamic imaging. The results of the clinical applications suggest that the proposed method provides a convenient approach to obtain accurate GFR estimation without manual interventions. We concluded that this automated ROI detection method is a promising method to automatically measure the GFR of patients with low-functioning kidneys.