Xinhui Duan PhD, Presenter: Nothing to Disclose

Jia Wang PhD, Abstract Co-Author: Nothing to Disclose

Mingliang Qu MD, Abstract Co-Author: Nothing to Disclose

Cynthia H. McCollough PhD, Abstract Co-Author: Research grant, Siemens AG

To improve the accuracy with which kidney stone volumes can be measured from clinical CT images

We developed a model-based method to estimate the volume of small objects, such as kidney stones, from CT images. First, the point spread function (PSF) was estimated using a thin tungsten wire in air and our clinical stone scan protocol. The 3D PSF was approximated as an isotropic 3D Gaussian function. Next, the PSF was convolved with a uniform spherical object with variable radius. Then, a correction factor was determined, which equaled the ratio of the object diameter to the diameter of the full width at half of the maximum CT number within the object (FWHM) as the CT number segmentation threshold (the corresponding volume is referred to as the FWHM volume). To evaluate the accuracy of the technique when applied to the measurement of kidney stone volumes, the FWHM volume determined using CT images of various stone types was multiplied by three correction factors, which were corresponding to the correction factor for the length, the width and the height of the stone respectively, to obtain corrected volumes. Scans were acquired of 61 human kidney stones (10 hydroxyl or carbonate apatite, 14 calcium oxalate, 10 cystine, 27 uric acid) that were placed in water-filled tubes within a 30-cm wide water phantom. Our clinical dual-energy CT protocol (80 and 140 kVp with tin filtration) was used, and images were reconstructed with 0.6-mm thickness. Reference stone volumes were measured using a water replacement method (mean stone volume = 97 mm3, range=28 - 317 mm3). 立方毫米   Searching...      

Compared with the FWHM volume, the average reduction of absolute error relative to the reference volume was 7.9% (p<0.0001). The error reduction had a moderate negative correlation with stone volume (Spearman coefficient ρ=-0.65) and weak correlation with CT number (ρ = 0.07). There was no significant difference between the results from 80 and 140 kVp images (p=0.42).      

The proposed method improved the accuracy of kidney stone volume measurement compared with use of a FWHM segmentation method.      

Accurate measurement of kidney stone volumes may allow improved tracking of stone growth or regression, improving patient management.      

Duan, X, Wang, J, Qu, M, McCollough, C, Kidney Stone Volume Estimation from CT Images Using a Model-based Method of Point Spread Function.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11009491.html