Abstract
1607
Objectives Recently reported large variabilities in the administered activity (AA) for pediatric nuclear medicine imaging studies suggest the need for well-founded AA guidelines. Such guidelines must balance image quality with the requirement that the absorbed dose be as low as reasonably possible. Towards this end, we have investigated the effect of body habitus on the relationship between AA and defect detectability for Tc-99m DMSA SPECT.
Methods We generated two phantoms, both with a body weight of 32 kg, but heights of 124 and 146 cm, corresponding to 10th and 90th height percentiles for this weight. Organ uptakes were determined based on literature values to simulate Tc-99m DMSA SPECT. Projection data were simulated modeling attenuation, scatter, and resolution of a LEHR collimator. Defects were simulated as uptake of 15-25% at 3 positions in the renal cortex. Projections were simulated with AAs of 20, 50, 75, 100, 125, and 150% of a standard activity of 1.85 MBq; noise was simulated corresponding to a 30 minute acquisition for each AA. Projections were reconstructed using filtered backprojection and a 3D Butterworth filter. We evaluated defect detection performance using the Channelized Hoteling Observer (CHO), which has previously been shown to model human performance.
Results The rating values from the CHO were analyzed using receiver operating characteristics (ROC) analysis to compute the area under the ROC curve (AUC). The AUC values for the taller patient were larger (indicating better performance), with p values greater than 0.05 in all but one cases. The data suggest that for a taller body habitus, the same AUC may be achieved with an ~50% decrease in AA relative to that for the shorter patient.
Conclusions Image quality is affected by body habitus and not simply weight. This suggests that more complete dosing rules are required to achieve minimum radiation exposures for pediatric patients