Inter-observer reproducibility of relative Tc-99(m)-DMSA uptake

Tc-99(m)-DMSA planar images of 49 randomly selected patients (10 adults, 39 children) were sent to 15 physicians at various centres in Belgium. They w ere asked to calculate, using their own routine program, the relative uptak e (expressed as a percentage) of each kidney. The data were sent on disks f ormatted so that they could be read by all participants, using their own co mputer systems. For each scan, the interobserver er variability was express ed using the maximum difference and the standard deviation of left renal up take. Left renal uptake measured by the 15 observers in the 49 patients was 29.0-72.0% (mean +/-s=49.8+/-6.4%). The maximum differences in left renal uptake ranged between 1.7% and 12.0% (4.5+/-2.6%); however, the maximum dif ference did not exceed 8% in about 90% of the patients. The standard deviat ions of the individual left renal uptake were between 0.6 and 3.9 (1.3+/-0. 8). The standard deviations were significantly higher in adults (mean stand ard deviation = 2.05) than in children (mean standard deviation = 1.12) (P < 0.001); this was probably related to the high background observed in thre e adults with severe renal impairment. Indeed there was a significant corre lation (P<0.001) between the standard deviation and both the signal-to-nois e ratio and the degree of asymmetry between the right and left kidneys. The differences between right and left kidney uptake were systematically lower for some observers, suggesting an influence of the calculation programs. ( (C) 2000 Lippincott Williams & Wilkins).