The effect of truncation artifact to the myocardium on myocardial perfusion SPECT : Phantom Analysis

Objectives When performing myocardial perfusion SPECT, truncation may occur when a patient is too large to fit within the field of view (FOV) or when using the large zooming acquisition. Some studies have claimed that the truncation can lead to various artifacts and distortions to SPECT image. However, there has been little study done concerning the effect of the degree of truncation and reconstruction method. We investigated the effect of truncation artifact to the myocardium.

Methods Using a cylinder phantom and an anthropomorphic torso phantom with extra cardiac activity, the projection data was acquired with dual camera heads fixed at 90°. The phantom filled with ^99mTc water solution was positioned centrally on the FOV (reference data), then shifted horizontally towards the right. Both filtered back projection (FBP) and ordered-subset expectation maximization (OSEM) were used to reconstruct the SPECT images. No attenuation or scatter correction was performed. The SPECT images of shifted phantom were compared with reference SPECT image to determine the effect of truncation artifact. The count profile through trans-axial images of cylinder phantom, and % uptake of 17 segments computed by QPS and count on the myocardium of trans-axial images of the anthropomorphic torso phantom were used to analyze the SPECT images.

Results In the cylinder phantom, the larger truncated area generated a higher artifact only on the edge of the image. It showed that OSEM is more high count than FBP. In the anthropomorphic torso phantom, for both FBP and OSEM, the difference of % uptake among different truncated areas was negligible. However, myocardium count was overestimated for truncated area 21.3% and 34.0%.

Conclusions In case of truncated area larger than 20%, truncation artifact should be greatly considered when counting myocardium such as myocardial blood flow for absolute quantification.