Shigeyoshi Soga MD, Presenter: Nothing to Disclose

Scott Britz-Cunningham, Abstract Co-Author: Nothing to Disclose

Kanako Kunishima Kumamaru MD, Abstract Co-Author: Nothing to Disclose

Michael Lally Steigner MD, Abstract Co-Author: Nothing to Disclose

Sayeed Khan Malek MD, Abstract Co-Author: Nothing to Disclose

Frank John Rybicki MD, PhD, Abstract Co-Author: Research grant, Toshiba Corporation Research grant, Bracco Group

Stefan G. Tullius MD,PHD, Abstract Co-Author: Nothing to Disclose

We investigated correlation between reference standard nuclear split renal function and six CT based methods, including previously reported methods plus a new modified ellipsoid method.

38 potential renal donors underwent 64-slice CT and nuclear renography. Split renal function calculated by a total of six CT methods with variable complexity were compared with reference standard renography results. CT based methods included split renal volume with (attenuation capacity) and without (split volume) renal attenuation correction, a modified ellipsoid method that used thick maximum intensity projection images, product of length and width, length, and a parenchymal area method defined as the product of average parenchymal thickness and renal length. Inter-observer agreement was evaluated for methods that required manual length measurements. Post processing time was recorded for split volume, modified ellipsoid, and parenchymal area methods. For nuclear split function, manual kidney depth correction was performed using depth measurements from CT. 

Four of the six CT based methods (split volume, ellipsoid method, parenchymal area, attenuation capacity) showed similarly strong correlation (r=0.84-0.79) with nuclear renography. Bland-Altman analysis revealed similar performance in differences (SDs less than 3.0%) between those CT measures and reference standard, as well as good inter-observer agreement (SDs less or equal to 3.0%) for the modified ellipsoid and parenchymal area methods. Other more technically simpler methods showed inferior performances. Post processing time for modified ellipsoid method was significantly shorter than semi-automated split renal volume or parenchymal area method (p<0.01).

Several CT-based measures showed excellent correlation with nuclear split renal function and support usefulness of CT to estimate split renal function. Among those CT based methods, the modified ellipsoid method is a good compromise, between post-processing time, high accuracy, and reproducibility.

Multiple CT based estimates have high correlation with nuclear split renal function. A modified ellipsoid method is an appealing compromise of high accuracy with limited image post-processing time.

Soga, S, Britz-Cunningham, S, Kumamaru, K, Steigner, M, Malek, S, Rybicki, F, Tullius, S, Comprehensive Assessment of CT-based Estimates for Split Renal Function for Potential Renal Donors: Modified Ellipsoid and Other CT-based Methods.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11034290.html