Heterogeneity of SPECT bull´s-eyes in normal dogs: comparison of attenuation compensation algorithms

Author

DiBella, E.V.R. ; Eisner, R.L. ; Schmarkey, L.S. ; Barclay, A.B. ; Patterson, R.E. ; Nowak, D.J. ; Lalush, D.S. ; Tsui, B.M.W.

Author_Institution

Dept. of Electr. Eng., Georgia Inst. of Technol., Atlanta, GA, USA

Volume

42

Issue

4

fYear

1995

fDate

8/1/1995 12:00:00 AM

Firstpage

1290

Lastpage

1296

Abstract

In normal dogs, SPECT ^99mTc Sestamibi (MIBI) and ²⁰¹Tl myocardial perfusion images reconstructed with filtered backprojection (FBP) show a large decrease of counts in the septal wall (S) compared to the lateral wall (L). We evaluated the iterative method of Chang (1978) at 0 and 1 iterations (Chang0 and Chang1), and the Maximum Likelihood-Expectation Maximization with attenuation compensation (ML-EM-ATN) algorithm on data acquired from 5 normal dogs and from simulated projection data using a homogeneous count-density model of a normal canine myocardium in the attenuation field measured in one dog. Mean counts in the S and L regions were calculated from maximum-count circumferential profile arrays. Our results demonstrate that ML-EM-ATN and Chang1 result in improved uniformity, as measured by the S/L ratio

Keywords

image reconstruction; maximum likelihood detection; single photon emission computed tomography; ²⁰¹Tl myocardial perfusion images; Chang0; Chang1; L regions; ML-EM-ATN; Maximum Likelihood-Expectation Maximization; S regions; S/L ratio; SPECT ^99mTc Sestamibi; SPECT bull´s-eye heterogeneity; attenuation compensation algorithms; dogs; filtered backprojection; homogeneous count-density model; image reconstruction; iterative method; lateral wall; normal canine myocardium; septal wall; Attenuation; Biomedical imaging; Cardiology; Computer simulation; Dogs; Heart; Image reconstruction; Imaging phantoms; Iterative methods; Myocardium;

fLanguage

English

Journal_Title

Nuclear Science, IEEE Transactions on

Publisher

ieee

ISSN

0018-9499

Type

jour

DOI

10.1109/23.467867

Filename

467867