Direct Comparison of 2D Cardiac PET With 3D Cardiac PET Using a 3D Monte-Carlo Scatter COrrection Algorithm

This study re-examines 2D an 3D myocardial perfusion PET data from 19 subjects, mean age 65 years, mean BMI 30 kg/m2, 79% male. Rest and stress scans were acquired on a Siemens Accel™ PET scanner in 2D (septa extended) and 3D (septa retracted) modes. Rb-82 doses were rest 53+/-5 mCi and stress 53+/-6 mCi. Imaging times were 2D emission scan for 3 mins (90 sec delay post Rb-82 infusion), followed by a 3 minute, 3D gated emission scan (180 sec delay). Decay and shorter acquisition times led to a 62% reduction in the effective dosage 2D vs 3D datasets. 3D images were first pre-processed using a Monte-Carlo scatter and prompt gamma correction algorithm (Imagen3D™) then reconstructed using ImagenProTM (CVIT, Kansas City, MO). Reconstructed images were evaluated using relative, 17 segment raw scores (Cedars QPET). Studies were read by consensus of 2-blinded readers for: image quality (1-4, poor-excellent), interpretive certainty (1-3, low-high) and rest perfusion using a 17 segment model (0=normal; 1-3 = mild, moderate, or severe perfusion defects). Stress segmental scores were not evaluated due to differences in imaging start time post dipyridamole infusion.