Re: Automated renal stone volume measurement by noncontrast computerized tomography is more reproducible than manual linear size measurement: S. R. Patel, P. Stanton, N. Zelinski, E. J. Borman, M. A. Pozniak, S. Y. Nakada and P. J. Pickhardt J Urol 2011; 186: 2275-2279

摘要

To the Editor: In their discussion Patel et al note that the accuracy of automated volume measurements was verified in other models using phantoms. However, they were using coronary artery calcium scoring tools, which had a fixed attenuation threshold of 130 HU. Marginal voxels, which are engaged partly by stone and partly by surrounding media, are difficult to segment accurately with a fixed attenuation threshold. The ensuing volume measurement typically is either underestimated or overestimated due to threshold levels that are too high or too low, respectively. In their study using phantoms Demehri et al used variable threshold as the segmentation threshold with OsiriX software for volume estimation of urinary calculi.