CT-based attenuation correction in 82Rb-myocardial perfusion PET–CT: incidence of misalignment and effect on regional tracer distribution

Abstract

Misalignment of low-dose-CT used for attenuation correction (AC) may cause artifacts in cardiac-PET-CT. The aim was to evaluate incidence and severity of misalignment and its quantitative effects on regional myocardial (82)Rb-distribution. Rest/dipyridamole (82)Rb-perfusion-PET-CT studies of 92 consecutive patients were analyzed for misalignment. Two different scanning protocols were employed: the first 57 patients had separate CTs for rest and stress PET. The following 35 patients had one CT at rest, used for AC of rest and stress PET. Misalignment was visually scored on a five-point scale (0 = no, 1 = minimal, 2 = mild, 3 = moderate, and 4 = severe). In five representative patients with normal perfusion and low probability of disease, 95 polarmaps were created by shifting CT vs PET prior to reconstruction of attenuation-corrected data sets using dedicated software (three dimensions of space; magnitude of shifts, 5, 10, 14 mm). PET/CT -misalignment was detected in 60% of rest and 67% of stress studies. Alignment for rest was better than that for stress (0.7 +/- 0.7 vs 1.0 +/- 0.9, P = 0.03). Comparison of the two protocols revealed no effect on the alignment of the stress study (1.0 +/- 0.9 vs 1.0 +/- 0.9, P = 0.9). Quantitatively, the largest individual effect of any artificial misalignment was a 25% reduction of relative (82)Rb uptake. With a shift of 1 cm, the largest effect in an individual was a 19% decrease. Anterior wall was most frequently influenced by misalignment, but changes of uptake also occurred in all other segments. Misalignment between CT and PET in cardiac-PET-CT influences regional tracer distribution in multiple segments. Repeated CT imaging after dipyridamole does not improve alignment. These results emphasize the need for strategies to improve coregistration in clinical imaging protocols.