Methodological error and spatial variability of organ blood flow measurements using radiolabeled microspheres

The quantitative analysis of the spatial variability of organ blood flow by means of radiolabeled microspheres (MS) requires that the methodological variability ("error") of the technique (RDmeth.) is known in each individual organ. Therefore, RDmeth. was quantified (eight to nine nuclides) in 6941 tissue samples from 13 organs of three anesthetized dogs, and the relative importance of errors originating from both the stochastic nature of MS distribution (RDtheo.) and the process of quantitation of MS radioactivity (RDcounting) was assessed under varying conditions (high/low specific MS activity (SAMS); inaccurate separation of gamma spectra; large sample size). At "minimized" methodological error (experiment 2), RDmeth. of samples trapping approximately 375 MS/nuclide was 5.8% and only slightly exceeded RDtheo. (5%). RDmeth. varied in the range 2.7-7.8% in individual organs and contributed little (3.5%) to the organs' observed spatial variability of flow. In contrast, RDmeth.--due to increased RDcounting--considerably exceeded RDtheo. when SAMS was low (experiment 3), overlap of two nuclides' main photopeaks was critical (experiment 1), or counting geometry was inappropriate (pulmonary tissue samples). At the same time, the contribution of RDmeth. to spatial flow variability rose to 7.9% (experiment 3), 26.9% (experiment 1), and 15-23% (lungs). Completely artifactual measurements, as indicated by an extremely high RDmeth. of sample flow, were rarely observed (less than 0.1%). In general, our data suggest that blood flow can be measured reproducibly and with low methodological error using up to 8 nuclides, RDmeth. does not essentially contribute to the observed spatial variability of organ blood flow, and, hence, organ flow variability may be accurately quantified using the MS technique. However, if sources of error as indicated above are present, the practice of using RDtheo. as a measure of RDmeth. (thereby neglecting RDcounting) may notably underestimate true MS error and result in an overestimation of spatial heterogeneity of organ blood flow. RDmeth., therefore, should be quantified separately in each region of interest prior to the onset of a new study.