| Abstract | dc.description.abstract | Purpose Myocardial perfusion SPECT is an excellent tool
for the assessment of coronary artery disease (CAD);
however, it is affected by several artifacts, such as patient
motion during acquisition, which increases false-positive
rates. Therefore, the purpose of this work is to analyze
changes in perfusion scores after motion-correction software
application.
Methods The population included 160 99mTc-sestamibi
CAD studies, divided into two groups: with and without
perfusion defects, equally divided into subgroups according
to movement during standard acquisition. A Siemens
ECAM 180 was used for processing without correction
and with automatic and manual e.soft 2.5 modalities. Visual
interpretation as well as QPS software was compared using
Pearson correlation and kappa agreement statistics.
Results Moderate agreement was observed between SPECT
interpretations after motion correction versus the original
report, according to the presence of perfusion defects. Manual
correction using the software obtained the lowest agreements.
Perfusion summed stress scores (SSS) correlation from
different processing modalities versus non-corrected studies
differed significantly independent of the degree of motion.
Mean SSS in 40 patients with no motion was 3.9±3.9 when
no correction was applied; with automatic correction was
8.8±10 (p=0.03) and with manual correction was 3.1±3.5
(p=ns versus non-corrected). Automatic correction was
better when applied to patients with mild to moderate
motion. In those with mild or no motion, software
overestimated or created new perfusion defects.
Conclusion Motion-correction software must be used with
caution when trying to optimize myocardial perfusion
SPECT based on individual analysis. Acquisition should
be always repeated in cases with severe motion and in no or
mild motion it seems preferable to avoid correction. | en_US |
