Principal component analysis of multigated cardiac bloodpool studies and correction of count-deficient frames

Abstract

The multigated cardiac bloodpool study is a nuclear medicine procedure that produces a movie loop of the contents of the chambers of the beating heart. The nature of the acquisition of data results in noisy studies that can have underexposed frames late in the cardiac cycle. These darker frames produce a "flashing" effect that disturbs visual interpretation The present means of correcting this problem distort the data in the study. We have used principal component analysis to correct the count deficiency of the last frames of the cardiac movie loop without distorting the temporal information in the study or degrading the visual signal-to-noise ratio. An initial evaluation of twenty cases yielded complete correction of the "flashing" effect and produced visually more pleasing datasets. Application of standard quantitative analysis methods demonstrated no difference in the clinical parameters such as the ejection fraction.