A multicenter evaluation of commercial attenuation compensation techniques in cardiac SPECT using phantom models

Background. Our goal was to evaluate the ability of current commercially available attenuation compensation (AC) techniques to correct for the effects of attenuation, scatter, and loss of resolution with depth, in a series of standardized phantom experiments. Methods and Results. The following systems were evaluated: Hawkeye (GE Medical Systems), Profile (Siemens Medical Systems), Vantage Pro (ADAC Laboratories), TAC (SMV America), M-STEP (University of Michigan, Ann Arbor), TransAct (Elscint), Beacon (Marconi Systems), and PET Advance (GE Medical Systems). Studies were performed with the use of normal and ischemic myocardium in air and with technetium 99m or fluorine 18 in torso phantoms. There was considerable variation in the uniformity of short-axis slices and defect contrast for images acquired in air (no AC). AC improved the uniformity of activity throughout normal myocardium. A simulated hot liver resulted in severe artifacts and was only partially corrected by many systems. A high-quality attenuation map appears to be an important determinant of image quality. No system produced AC images of comparable quality to those obtained in the absence of scatter/attenuating media. Conclusions. There are significant differences in the ability of commercial AC systems to reduce artifacts due to attenuation and scatter. These differences are partly the result of non-AC factors (collimation, orbit, etc). In general, systems that generated high-quality attenuation maps yielded the best results.