Usefulness of the Chang attenuation correction method with use of a CT-based μ map by FBP reconstruction in (201)Tl SPECT-MPI.

Objective: Attenuation correction (AC) on nuclear images of non-uniform domains is generally performed by a change of the computed tomography (CT) values to μ values, which are then inserted as components into the detection probability of iterative reconstruction techniques (OS-EM Iterative AC). We established an AC technique which uses a CT μ map based on the Chang AC. Our purpose in this study was to confirm the appropriateness of the Chang AC with the OS-EM and the FBP method (OS-EM Chang AC and FBP Chang AC) by evaluating the results obtained in a phantom and clinical study for (201)Tl single-photon emission-computed tomography-myocardial perfusion imaging (SPECT-MPI).

Methods: Myocardial phantom study and retrospective clinical study were performed. Evaluations for image quality (uniformity and contrast) and image quantitative values [accurate left ventricular (LV) volume and radioactivity] were performed for both studies.

Results: FBP Chang AC showed good image uniformity and proper contrast in phantom and clinical study. Accurate LV volume and radioactivity in the myocardium were also obtained by the phantom study. On the other hand, the number of iterations influenced the image quality both in OS-EM Iterative AC and OS-EM Chang AC in the phantom study. Different numbers of iterations were necessary for obtaining good contrast ratio in each of the anterior and inferior wall, and accurate LV volume.

Conclusions: The number of iterations influences the image quality and quantitative values on OS-EM Iterative Chang AC and OS-EM Chang AC images. In addition, it is difficult to set an appropriate number of iterations for the iterative reconstruction of these images in phantom and clinical studies. Therefore, FBP Chang AC is considered to be clinically useful.