Assessment of Kidney Volume Measurement Techniques for Ultrasound Images

Abstract

This study intends to assess and compare the accuracy of different methods for estimating the kidney volume of ultrasound images consist of volume measurement from length-based, area-based and surface-based. For length-based method, the ellipsoid formula was used and for surface-based method, the volume can be automatically obtained from 3D ultrasound system after some manual contouring. For area-based method, sets of ultrasound images with different number of slices were used. After manual contouring, the slices were reduced to 2D representation and the enclosed area between slices is calculated as the volume of the kidney. For a better assessment, experiment was also performed using egg phantom. As results, for egg phantom, by using water displacement method as gold standard volume, length-based method underestimates the volume for about 6% and surface-based method overestimate the volume for about 4%. For area-based method, the volume is also underestimated but varies with total number of slices used and can be as low as 2.6% and as much as 4.5%. By applying the same analysis to the kidney, it is thus concluded that for kidney volume measurement, area-based and surface based methods are more accurate but an automatic technique for contour detection need to be developed for repeatability usage of the methods. Length-based method on the other hand needs to have a new correctional factor implemented to ellipsoid formula for more accurate volume measurement.