J Nucl Med. 2007; 48 (Supplement 2):120P
General Clinical Specialties: Pediatrics
Pediatrics II |
The utility and properties of the geometric mean in the assessment of differential renal function in pediatric lasix renography
James Warrington1,
Martin Charron1 and
Pipi Salle1
1 Department of Diagnostic Imaging, Division of Nuclear Medicine, Hospital for Sick Children, Toronto, Ontario, Canada
407
Objectives: The geometric mean (GM) method is believed to improve the accuracy of assessment of differential renal function compared with the determination of differential renal function from posterior differential (PD) images alone. The aim of this work is to improve understanding of the nature of the GM correction in pediatric lasix renography and to determine how often the GM changes the categorization of renal function between impairment and normal function compared with routine PD assessment. Methods: Lasix renograms for 67 patients (median age 3 years old) with suspected renal obstruction were assessed for differential renal function by both the posterior differential method (PD) and the geometric mean method (GM) requiring the acquisition of both anterior and posterior renal images. Fifty-five patients had hydronephrotic renal enlargement and eleven patients had renal ectopia, horseshoe kidney configuration, duplicated collecting systems or bilateral ureteric enlargement. Results: When the differential renal function of the abnormal kidney was less than 50%, the mean PD value was 38.3% and the GM increased this value on average by 1.7% (p=0.01, n=41). When the differential renal function of the abnormal kidney was greater than 50% the mean PD value was 56.1% and the GM decreased this value on average by 2.2% (p=0.02, n=20). PD values less than 30% were associated with a greater increase by GM assessment (4.3%) compared with PD values within the 30% to 50% range (0.9%) (p=0.04, n=47). The geometric mean changed the classification of differential renal function from abnormal to normal in 13% of cases and from normal to abnormal in 7% of cases. Conclusions: GM correction tends to normalize the differential renal function by shifting it closer toward 50% compared with the PD measurement. The magnitude of the difference between the GM and the PD values tends to be greater when renal function demonstrates a greater degree of impairment. One fifth of cases resulted in a change in categorization of renal function by use of the GM to correct the PD.
