Comprehensive comparative study of computed tomography-based estimates of split renal function for potential renal donors: Modified ellipsoid method and other CT-based methods

Objective: To investigate a new modified computed tomographic (CT) ellipsoid method of split renal function and to compare results from this method with other CT-derived metrics. Methods: Thirty-eight potential renal donors with both CT and nuclear renography were retrospectively evaluated for estimated split function using 6 CT methods to determine accuracy. For the CT methods, correlation, reproducibility, ease in image post-processing, and the ability of CT-derived methods to determine the dominant kidney before renal transplantation were evaluated using a nuclear renography reference standard. Results: Four of the 6 CT methods (split renal volume, modified ellipsoid method, parenchymal area, attenuation capacity) showed similar strong correlation (r = 0.84-0.79). Bland-Altman analysis revealed similar performance in differences (SDs <3.0%) between those CT measures and reference standard, as well as good interobserver agreement for the modified ellipsoid and parenchymal area methods. The technically simpler methods had inferior performance. Post-processing time for the modified ellipsoid method was significantly shorter than semiautomated split renal volume or parenchymal area method (P < 0.01). Each CT-based method showed excellent agreement (100% or 97.4%) with renography regarding the determination of dominant kidney. Conclusions: Excellent correlation with nuclear split renal function supports the use of CT alone for the imaging assessment for many potential renal donors, including the decision of which kidney to harvest. Among the CT-based methods, the modified ellipsoid method can be performed rapidly with high accuracy and reproducibility.