Effect of CT misalignment on attenuation — corrected myocardial perfusion SPECT

Abstract

BACKGROUND: Use of CT based attenuation correction (AC) for myocardial perfusion SPECT (MPS) is growing fast due to a rapid development of hybrid SPECT/CT systems. SPECT and CT studies are performed in a sequential way extending total study acquisition and making a patient movement more likely. The present work aims at answering the question how large misalignment between SPECT and CT studies should be considered significant and how often those misregistrations are observed.

MATERIAL AND METHODS: A retrospective study applying AC was performed in 107 patients who had coronary angiography (CA) performed within 3 months. Patients underwent a stress/rest Tc-99m MIBI 2 day SPECT/CT myocardial perfusion study. In case of SPECT and CT misalignment CT slices were shifted manually; shifts along 3 axes were recorded and after realignment a repeat reconstruction was performed. Euclidean distance of misalignment was also calculated. Images were analyzed by two experienced nuclear medicine specialists (consensus) applying visual semiquantitative method. Perfusion of three arteries was scored using a 5 grade scale. CA results were used as a reference for MPS findings.

RESULTS: In 47 patients (44%) CT realignment was necessary. CT was shifted mostly along x and y axes, and less often along z axis. Euclidean distance S exceeded 2 pixels in 3 stress and 2 rest studies. Only in 7 patients changes of scores assigned to coronary vessels were noted as a result of CT realignment. These changes concerned 9 vessel areas. In 7 out of 9 cases changes were noted toward a better agreement with results of CA. Only in one patient, with stress S > 3 pixels and negative result of CA, CT realignment changed vessel area score significantly, from probably abnormal to normal.

CONCLUSIONS: Only misalignments large enough, exceeding 2–3 pixels, have negative impact on attenuation corrected images. Such misalignments are rare, in our material were observed in 3 stress and 2 rest studies (3% and 2% of all studies, respectively). Only in one patient (below 1% of all studied patients) CT misalignment caused a significant study misinterpretation. Although alignment of SPECT and CT studies should be checked in every patient, small misalignments do not affect study interpretation.