Background: The objective of this study was to analyse the performance of several variants of kidney scintigraphy in children from the standpoint of: scar detection, an assessment of the rating of the pathology and an investigation of interobserver variability involved in the diagnostic procedure.
Material and methods: The analysis is based on results of a planar kidney scintigraphy and of a tomographic (SPECT) procedure. The latter was performed in two variants: 1) in which slices were obtained with axis of reconstruction identical with longitudinal axis of the body (SPECT I) and 2) in which axes were fitted to the long axis of each kidney separately (SPECT II). The rating of the diagnosed pathology was made using two scales, according to Goldraich and Howard. Evaluation of the images involved on the one hand, 150 individual kidneys and 75 patients on the other. The assessment was made by three independent observers, differing in experience in nuclear medicine and employed in three independent departments. In the statistical analysis, as a measure of observer agreement, a proportion of agreeing readings (%) was accepted; in addition, the kappa index of agreement was calculated.
Results: Better agreement among three observers was attained when planar images were read in contrast to SPECT (I and II) results. The reading of SPECT II images yielded a higher frequency of diagnosed pathology (scars) in kidneys and is characterized by better overall agreement in detection by individual observers than a similar evaluation of SPECT I images. The Goldraich scale secures better interobserver agreement of renal scar detection than is seen when the Howard scale was applied to acquire the rating.
Conclusions: The conclusion may be drawn that kidney scintigraphy is a method still burdened with a substantial subjectivism. Planar scintigraphy should be treated as a basic option for imaging post-inflammatory changes in kidneys.