Angiographic Assessment of Myocardial Perfusion Using Correlation Analysis

Abstract

Although angiography was originally designed as a morphological imaging modality, it is increasingly demanded to retrieve functional information from conventional angiograms. As X-ray angiography provides only 2D projection images, its most significant advantage over volumetric imaging modalities is its high temporal resolution. Thus the angiographic functional imaging relies mainly on the analysis of the temporal variation of single pixels or local regions. In order to assess the myocardial perfusion, the myocardium has to be recognized at first. This can be solved similarly with analysis of the time-intensity curves. Compared to most approaches in which the characteristic of the time-intensity curves are classified, our proposed method based on correlation analysis has the advantage of using all the sampled data points so that it is more robust against noise or outliers. As the correlation coefficient considers only the colinearity of two curves, we add the information about the amplitude and the numerical range of the curves, which further improves the recognition of the myocardium.