Planar scan vs. SPECT/low-dose CT for estimating split renal function by ^99mTc-DMSA scintigraphy in children

Abstract

In the present study, we compared estimates of split renal function (SRF) in paediatric patients of various diagnostic subgroups by ^99mTc-dimercaptosuccinic acid (DMSA) scintigraphy using either geometric mean (GM) based on planar scans or a volume of interest (VOI)-based analysis on single photon emission tomography combined with low-dose CT (SPECT/ldCT). Two experienced physicians blinded to patient diagnosis retrospectively analysed all paediatric ^99mTc-DMSA scintigraphies that were conducted in our department between 2011 and 2016 and which included both a planar scan and SPECT/ldCT. All scintigraphies were performed on either a Phillips Precedence 16 slice CT or a Siemens Symbia 16 slice CT. SRF was estimated from planar scintigraphy using the geometric mean (GM), while the VOI-based analysis (VBA) was used for kidney segmentation on SPECT/ldCT.

Results

A total of 68 scintigraphies were included. A Bland-Altman plot-based analysis showed a bias for SRF of 2.1% with limits of agreement from − 7.5 to + 11.7% for the whole data set but showed larger differences between the two methods outside the normal range of 45–55%. In the GM-based SRF analyses, 29 cases were found to be outside the normal range, and in seven of these, VBA showed normal SRF. In the remaining 39 cases, VBA showed an abnormal SRF in only one case.

Conclusion

Approximately a quarter of planar DMSA scintigraphies that show an abnormal SRF in paediatric patients may be normal when assessed by SPECT/ldCT, which likely reflects underestimation of the kidney with the poorest function when assessed by GM due to the lack of attenuation correction. Planar scans that show an abnormal SRF in paediatric patients should thus preferably be supplemented by SPECT/ldCT.