Pediatric 99mTc-DMSA Renal SPECT: Electronic Lesion Insertion of Focal Renal Defects for Dose Optimization

Objectives The ultimate objective is to evaluate the tradeoff between image quality (IQ), as measured by defect detectability, and administered activity (AA) in 99mTc-DMSA SPECT as a function of age, patient height and weight. We have previously developed a database of pediatric phantoms for both genders (Newborn to 15-year-old). The goal of this work was to model renal defects of known size, position and intensity using this database.

Methods We created focal renal lesions consisting of volumes of reduced uptake simulating focal acute pyelonephritis at various locations of the renal cortex. The defects were created by the intersection of an ellipsoid with the cortex. The center of the ellipsoid was positioned at the outer cortical surface. The major axis was along the line passing through the center of the kidney and had a length equal to twice the cortical thickness. This gave a defect shape that was smaller at the inner and larger at the outer cortical surfaces, consistent with clinically observed defects. Defects with gradual or sharp decreases of activity at the defect edges were created to model infection or scar, respectively. Partial functional reduction was modeled by scaling the defect by a factor <1.0 before subtracting from the normal cortex. The volumes of these defects were chosen to be near the limit of clinical detectability and yet clinically relevant. The defect sizes are different for each phantom: the defect volume was set to be 0.3 cm3 for a newborn phantom and such that the defect had the same contrast for other ages and sizes. Defects were created for the UF NHANES-based phantom series, which is comprised of phantoms representing five ages (newborn, 1, 5, 10, 15 years), both genders and three height percentiles (10th, 50th, and 90th) for each age. For each age, gender and height there are 3 phantoms with 3 kidney volumes: reference and ±15%. Organ uptakes were based on a recently developed pharmacokinetic model for DMSA that has been validated against clinical DMSA SPECT scans. Phantoms were generated to model uptake at 3-hours post-injection. Realistic uptake variations based on the clinical images were also modeled. Projections of the phantoms were generated using an analytic projection code that models attenuation, scatter and the collimator-detector response. Noisy projections for AAs were generated. Combined with the defect model these will serve as the basis for mathematical observer studies to determine the tradeoff between AA and IQ for the phantoms in the database.

Results A model and a set of defects at 3 positions for each of the 90 phantoms in the database have been developed. The defects were judged to be clinically realistic.

Conclusions The defects and phantom database are being used in a study to evaluate the tradeoff between IQ and AA. Combined with data on relative radiation risk, these results will be useful in developing administered activity guidelines for pediatric 99mTc-DMSA SPECT.